This document discusses model-based calibration techniques used to develop an engine calibration to meet emissions standards. It focuses on using design of experiments methods and statistical modeling to optimize engine parameters. Specifically, it describes:
1) Using a two-stage regression approach to separate variables like spark timing into a "local model" and others into a "global model" to better characterize responses.
2) How screening experiments can help select appropriate variables and ranges for the design of experiments to avoid unstable operating points.
3) Techniques for modeling experimental data through "local models" of individual variables and a "global model" that combines local models to reproduce responses for any combination of variables.
This document describes a study that used multi-objective optimization to improve the efficiency of centrifugal pump impellers. The study used computational fluid dynamics simulations and genetic algorithms to optimize impeller blade designs. Two application examples were considered. The first optimized a design with 14 parameters and achieved a 2% efficiency improvement over the baseline design. The second optimized a design with 8 parameters and achieved a 1% efficiency gain and 2m head increase. Overall, the multi-objective optimization approach was able to produce designs with better efficiency and head characteristics than the baseline designs.
Wind farm layout optimization (WFLO) is the process of optimizing the location of turbines in a wind farm site, with the possible objective of maximizing the energy production or minimizing the average cost of energy. Conventional WFLO methods not only limit themselves to prescribing the site boundaries, they are also generally applicable to designing only small-to-medium scale wind farms (<100 turbines). Large-scale wind farms entail greater wake-induced turbine interactions, thereby increasing the computa- tional complexity and expense by orders of magnitude. In this paper, we further advance the Unrestricted WFLO framework by designing the layout of large-scale wind farms with 500 turbines (where energy pro- duction is maximized). First, the high-dimensional layout optimization problem (involving 2N variables for a N turbine wind farm) is reduced to a 6-variable problem through a novel mapping strategy, which allows for both global siting (overall land configuration) and local exploration (turbine micrositing). Sec- ondly, a surrogate model is used to substitute the expensive analytical WF energy production model; the high computational expense of the latter is attributed to the factorial increase in the number of calls to the wake model for evaluating every candidate wind farm layout that involves a large number of turbines. The powerful Concurrent Surrogate Model Selection (COSMOS) framework is applied to identify the best surrogate model to represent the wind farm energy production as a function of the reduced variable vector. To accomplish a reliable optimum solution, the surrogate-based optimization (SBO) is performed by implementing the Adaptive Model Refinement (AMR) technique within Particle Swarm Optimization (PSO). In AMR, both local exploitation and global exploration aspects are considered within a single optimization run of PSO, unlike other SBO methods that often either require multiple (potentially mis- leading) optimizations or are model-dependent. By using the AMR approach in conjunction with PSO and COSMOS, the computational cost of designing very large wind farms is reduced by a remarkable factor of 26, while preserving the reliability of this WFLO within 0.05% of the WFLO performed using the original energy production model.
One of the primary drawbacks plaguing wider acceptance of surrogate models is their low fidelity in general. This issue can be in a large part attributed to the lack of automated model selection techniques, particularly ones that do not make limiting assumptions regarding the choice of model types and kernel types. A novel model selection technique was recently developed to perform optimal model search concurrently at three levels: (i) optimal model type (e.g., RBF), (ii) optimal kernel type (e.g., multiquadric), and (iii) optimal values of hyper-parameters (e.g., shape parameter) that are conventionally kept constant. The error measures to be minimized in this optimal model selection process are determined by the Predictive Estimation of Model Fidelity (PEMF) method, which has been shown to be significantly more accurate than typical cross-validation-based error metrics. In this paper, we make the following important advancements to the PEMF-based model selection framework, now called the Concurrent Surrogate Model Selection or COS- MOS framework: (i) The optimization formulation is modified through binary coding to allow surrogates with differing num- bers of candidate kernels and kernels with differing numbers of hyper-parameters (which was previously not allowed). (ii) A robustness criterion, based on the variance of errors, is added to the existing criteria for model selection. (iii) A larger candidate pool of 16 surrogate-kernel combinations is considered for selection − possibly making COSMOS one of the most comprehensive surrogate model selection framework (in theory and implementation) currently available. The effectiveness of the COSMOS framework is demonstrated by successfully applying it to four benchmark problems (with 2-30 variables) and an airfoil design problem. The optimal model selection results illustrate how diverse models provide important tradeoffs for different problems.
IRJET- Optimum Design of Fan, Queen and Pratt TrussesIRJET Journal
The document discusses the optimum design of various truss configurations including fan, queen, and Pratt trusses. The author formulates the design of these trusses as an optimization problem to minimize weight. The problem considers stress, buckling, and deflection constraints with cross-sectional area as the design variable. A sequential linear programming technique is used to solve the optimization problem. Parametric studies are performed to understand the effect of truss geometry including span, purlin spacing, truss spacing, and height on selecting the best design. The results of optimizing double fan, triple fan, modified queen, double Pratt, and triple Pratt trusses are presented and discussed.
This document describes a visually-informed decision-making platform (VIDMAP) for model-based design of wind farms. It aims to quantify and illustrate the criticality of information exchanged between different models in the wind farm layout optimization process. The platform consists of three main components: (1) uncertainty quantification to quantify variability in inputs and uncertainties introduced by upstream models, (2) sensitivity analysis to analyze sensitivity of downstream models, and (3) information visualization to visualize uncertainties and inter-model sensitivities. Sensitivity analysis is performed to quantify the sensitivity of an energy production model to first-level inputs and errors in upstream models like wind distribution, shear, turbine power response, and wake models.
IRJET- Optimization of Fink and Howe TrussesIRJET Journal
This document describes research on optimizing the weight of different truss configurations, including double fink, triple fink, modified fink, double Howe, and triple Howe trusses. The optimization problem aims to minimize weight by treating cross-sectional areas as design variables, while satisfying stress, buckling, and deflection constraints. An improved sequential linear programming technique is used to solve the optimization problem. The process involves developing a C program for load calculation, using MATLAB for truss analysis, and applying an optimizer based on improved SLP to determine optimized cross-sectional areas. A parametric study is then carried out by varying span, height, and spacing to identify the most economical truss configuration under the given conditions.
Using MEPDG to Develop Rational Pay Factor for Hot Mix Asphalt ConstructionIOSR Journals
Most highway agencies are using subjective measures to rate the quality of construction projects. The main purposes of rating the contractor performance are for qualification, bidding, or payment schedules. The pay factor (PF) can be defined as a multiplication factor that is often used to determine the contractor pay for the unit of work. The objective of this paper is to propose a rational methodology for defining PF based on basic understanding of the effect of different hot mix asphalt parameters on pavement life. Mechanistic-Empirical Pavement Design Guide (MEPDG) software is used to find the effect of different variables on pavement life. Three main factors considered in pay factor equation for their effect on pavement life:air voids, asphalt content, and percent passing sieve # 200 (P200). The pavement lifebased on the fatigue and rutting failure criteriaare used in the PF model to reflect actual contractor performance.The developed PF model was implemented on a case study.
This academic article discusses numerical flow simulation of an elbow draft tube using STAR-CCM+ software. It summarizes the steps taken which include geometric modeling of the draft tube, mesh generation, specification of parameters like material properties and boundary conditions, and setting up the implicit unsteady simulation. Results of the simulation like pressure and velocity variations along the draft tube over different time steps are presented. Key findings are that pressure and velocity distributions are affected by time step size, with smaller time steps showing more uniform distributions. Performance metrics like head recovery and efficiency of the draft tube are computed and seen to decrease initially with time step before leveling off.
This document describes a study that used multi-objective optimization to improve the efficiency of centrifugal pump impellers. The study used computational fluid dynamics simulations and genetic algorithms to optimize impeller blade designs. Two application examples were considered. The first optimized a design with 14 parameters and achieved a 2% efficiency improvement over the baseline design. The second optimized a design with 8 parameters and achieved a 1% efficiency gain and 2m head increase. Overall, the multi-objective optimization approach was able to produce designs with better efficiency and head characteristics than the baseline designs.
Wind farm layout optimization (WFLO) is the process of optimizing the location of turbines in a wind farm site, with the possible objective of maximizing the energy production or minimizing the average cost of energy. Conventional WFLO methods not only limit themselves to prescribing the site boundaries, they are also generally applicable to designing only small-to-medium scale wind farms (<100 turbines). Large-scale wind farms entail greater wake-induced turbine interactions, thereby increasing the computa- tional complexity and expense by orders of magnitude. In this paper, we further advance the Unrestricted WFLO framework by designing the layout of large-scale wind farms with 500 turbines (where energy pro- duction is maximized). First, the high-dimensional layout optimization problem (involving 2N variables for a N turbine wind farm) is reduced to a 6-variable problem through a novel mapping strategy, which allows for both global siting (overall land configuration) and local exploration (turbine micrositing). Sec- ondly, a surrogate model is used to substitute the expensive analytical WF energy production model; the high computational expense of the latter is attributed to the factorial increase in the number of calls to the wake model for evaluating every candidate wind farm layout that involves a large number of turbines. The powerful Concurrent Surrogate Model Selection (COSMOS) framework is applied to identify the best surrogate model to represent the wind farm energy production as a function of the reduced variable vector. To accomplish a reliable optimum solution, the surrogate-based optimization (SBO) is performed by implementing the Adaptive Model Refinement (AMR) technique within Particle Swarm Optimization (PSO). In AMR, both local exploitation and global exploration aspects are considered within a single optimization run of PSO, unlike other SBO methods that often either require multiple (potentially mis- leading) optimizations or are model-dependent. By using the AMR approach in conjunction with PSO and COSMOS, the computational cost of designing very large wind farms is reduced by a remarkable factor of 26, while preserving the reliability of this WFLO within 0.05% of the WFLO performed using the original energy production model.
One of the primary drawbacks plaguing wider acceptance of surrogate models is their low fidelity in general. This issue can be in a large part attributed to the lack of automated model selection techniques, particularly ones that do not make limiting assumptions regarding the choice of model types and kernel types. A novel model selection technique was recently developed to perform optimal model search concurrently at three levels: (i) optimal model type (e.g., RBF), (ii) optimal kernel type (e.g., multiquadric), and (iii) optimal values of hyper-parameters (e.g., shape parameter) that are conventionally kept constant. The error measures to be minimized in this optimal model selection process are determined by the Predictive Estimation of Model Fidelity (PEMF) method, which has been shown to be significantly more accurate than typical cross-validation-based error metrics. In this paper, we make the following important advancements to the PEMF-based model selection framework, now called the Concurrent Surrogate Model Selection or COS- MOS framework: (i) The optimization formulation is modified through binary coding to allow surrogates with differing num- bers of candidate kernels and kernels with differing numbers of hyper-parameters (which was previously not allowed). (ii) A robustness criterion, based on the variance of errors, is added to the existing criteria for model selection. (iii) A larger candidate pool of 16 surrogate-kernel combinations is considered for selection − possibly making COSMOS one of the most comprehensive surrogate model selection framework (in theory and implementation) currently available. The effectiveness of the COSMOS framework is demonstrated by successfully applying it to four benchmark problems (with 2-30 variables) and an airfoil design problem. The optimal model selection results illustrate how diverse models provide important tradeoffs for different problems.
IRJET- Optimum Design of Fan, Queen and Pratt TrussesIRJET Journal
The document discusses the optimum design of various truss configurations including fan, queen, and Pratt trusses. The author formulates the design of these trusses as an optimization problem to minimize weight. The problem considers stress, buckling, and deflection constraints with cross-sectional area as the design variable. A sequential linear programming technique is used to solve the optimization problem. Parametric studies are performed to understand the effect of truss geometry including span, purlin spacing, truss spacing, and height on selecting the best design. The results of optimizing double fan, triple fan, modified queen, double Pratt, and triple Pratt trusses are presented and discussed.
This document describes a visually-informed decision-making platform (VIDMAP) for model-based design of wind farms. It aims to quantify and illustrate the criticality of information exchanged between different models in the wind farm layout optimization process. The platform consists of three main components: (1) uncertainty quantification to quantify variability in inputs and uncertainties introduced by upstream models, (2) sensitivity analysis to analyze sensitivity of downstream models, and (3) information visualization to visualize uncertainties and inter-model sensitivities. Sensitivity analysis is performed to quantify the sensitivity of an energy production model to first-level inputs and errors in upstream models like wind distribution, shear, turbine power response, and wake models.
IRJET- Optimization of Fink and Howe TrussesIRJET Journal
This document describes research on optimizing the weight of different truss configurations, including double fink, triple fink, modified fink, double Howe, and triple Howe trusses. The optimization problem aims to minimize weight by treating cross-sectional areas as design variables, while satisfying stress, buckling, and deflection constraints. An improved sequential linear programming technique is used to solve the optimization problem. The process involves developing a C program for load calculation, using MATLAB for truss analysis, and applying an optimizer based on improved SLP to determine optimized cross-sectional areas. A parametric study is then carried out by varying span, height, and spacing to identify the most economical truss configuration under the given conditions.
Using MEPDG to Develop Rational Pay Factor for Hot Mix Asphalt ConstructionIOSR Journals
Most highway agencies are using subjective measures to rate the quality of construction projects. The main purposes of rating the contractor performance are for qualification, bidding, or payment schedules. The pay factor (PF) can be defined as a multiplication factor that is often used to determine the contractor pay for the unit of work. The objective of this paper is to propose a rational methodology for defining PF based on basic understanding of the effect of different hot mix asphalt parameters on pavement life. Mechanistic-Empirical Pavement Design Guide (MEPDG) software is used to find the effect of different variables on pavement life. Three main factors considered in pay factor equation for their effect on pavement life:air voids, asphalt content, and percent passing sieve # 200 (P200). The pavement lifebased on the fatigue and rutting failure criteriaare used in the PF model to reflect actual contractor performance.The developed PF model was implemented on a case study.
This academic article discusses numerical flow simulation of an elbow draft tube using STAR-CCM+ software. It summarizes the steps taken which include geometric modeling of the draft tube, mesh generation, specification of parameters like material properties and boundary conditions, and setting up the implicit unsteady simulation. Results of the simulation like pressure and velocity variations along the draft tube over different time steps are presented. Key findings are that pressure and velocity distributions are affected by time step size, with smaller time steps showing more uniform distributions. Performance metrics like head recovery and efficiency of the draft tube are computed and seen to decrease initially with time step before leveling off.
This document summarizes the results of lab tests that were conducted to validate a theoretical approach for using Quality Function Deployment (QFD) and Theory of Inventive Problem Solving (TRIZ) for component mathematics modeling in virtual durability simulations. The methodology involved identifying customer needs, evaluating solutions, applying QFD concepts, using TRIZ to solve issues, and analyzing failure modes and effects analysis (FMEA). Simulation results using different element types showed that using RBE3 elements instead of RBE2 elements better represented real-world behavior by allowing for lateral displacement. Simplified models were also developed and tested to validate force equivalence with full vehicle simulations.
IRJET- Optimization of Process Parameter in Injection Moulding using Tagu...IRJET Journal
This document summarizes a study that used Taguchi methods and ANOVA to optimize injection molding process parameters for a polypropylene chair bottom part. The researchers used Moldflow Adviser software to simulate different combinations of mold temperature, melt temperature, injection pressure, and cooling time based on an L9 orthogonal array. The goal was to minimize fill time and volumetric shrinkage. ANOVA results identified melt temperature as the most significant factor influencing both responses. The best optimized parameters found were a low melt temperature, low mold temperature, low injection pressure, and long cooling time.
Invited presentation to 2003 RAeS Aerodynamics research conferencestephen_mcparlin
The document summarizes an assessment process used to evaluate air vehicle concepts from initial design through analysis and testing. The process involves conceptual design using multidisciplinary variable optimization (MVO) followed by detailed configuration design using CAD tools. Performance is then analyzed using computational fluid dynamics (CFD) and wind tunnel testing. Underlying technologies like CAD integration, response surface methods, and CFD methods have improved the accuracy and speed of the assessment process.
Tuning the model predictive control of a crude distillation unitISA Interchange
Tuning the parameters of the Model Predictive Control (MPC) of an industrial Crude Distillation Unit (CDU) is considered here. A realistic scenario is depicted where the inputs of the CDU system have optimizing targets, which are provided by the Real Time Optimization layer of the control structure. It is considered the nominal case, in which both the CDU model and the MPC model are the same. The process outputs are controlled inside zones instead of at fixed set points. Then, the tuning procedure has to define the weights that penalize the output error with respect to the control zone, the weights that penalize the deviation of the inputs from their targets, as well as the weights that penalize the input moves. A tuning approach based on multi-objective optimization is proposed and applied to the MPC of the CDU system. The performance of the controller tuned with the proposed approach is compared through simulation with the results of an existing approach also based on multi-objective optimization. The simulation results are similar, but the proposed approach has a computational load significantly lower than the existing method. The tuning effort is also much lower than in the conventional practical approaches that are usually based on ad-hoc procedures.
The document describes a study that used an orthogonal experiment and numerical simulation to optimize the design of a centrifugal pump impeller. Five impeller geometric parameters were selected as factors to optimize: impeller outlet width, impeller inlet diameter, impeller blade wrapping angle, impeller blade outlet angle, and impeller blade inlet angle. Sixteen impeller designs were generated using an orthogonal table to vary the factors. Computational fluid dynamics was used to simulate the performance of pumps with each impeller design. Variance analysis identified the best parameter combination for improving pump efficiency and head. The optimized design showed significant improvements in efficiency and head compared to the original pump design.
This document summarizes a computational fluid dynamics (CFD) simulation of airflow around a simplified pickup van model. The study used a commercial CFD software to perform a 3D, steady-state simulation using the Reynolds-averaged Navier-Stokes equations and a k-ε turbulence model. The simulation was run at Reynolds numbers of 3x105 and results were validated against experimental wind tunnel data. Key findings included pressure and velocity distributions that matched experimental data well, indicating CFD can be an effective alternative to wind tunnel testing for pickup van aerodynamic analysis.
Compartmental modeling of high purity air separation columnsAbu Hussein
High purity distillation columns are critical unit operations in cryogenic air separation plants. The development of nonlinear control technology is motivated by the need to frequently change production rates in response to time varying utility costs. Detailed column models based on stage-by-stage balance equations are too complex to be incorporated directly into optimization-based strategies such as nonlinear model predictive control. In this paper, we develop reduced order dynamic models for the upper column of a cryogenic air separation plant by applying time scale arguments to a detailed stage-by-stage model that includes mass and energy balances and accounts for non-ideal vapor–liquid equilibrium. The column is divided into compartments according to the locations of liquid distributors and feed and withdrawal streams.
IRJET- Spray Characteristics of Diesel Fuel using Numerical SimulationIRJET Journal
This document discusses a numerical simulation of diesel fuel spray characteristics using computational fluid dynamics (CFD) software. The study aims to improve spray models by simulating atomization of diesel and bio-diesel sprays using dynamic mesh refinement. The simulation predicts spray parameters like Sauter mean diameter, spray cone angle, and spray tip penetration under different injection pressures. Results from the CFD model show good agreement with experimental data. The model can help advance engine design by analyzing spray behavior and mixture formation.
Numerical Simulation of Combustion Behavior of DI Diesel Engine with Conjunct...Khatir NAIMA
Currently, computational fluid dynamics has become an effective supplement to experimentation in the analysis and development of various engineering systems including internal combustion engines. In fact, multi-dimensional modelling of IC engines is less extensive and less time consuming than experimentation. In this aim, CONVERGE code was used to study the combustion behavior in a DI engine with various mesh control techniques including embedding and Adaptive Mesh Refinement (AMR). The simulation covers the compression, spray, combustion and expansion. A single spray plume and 1/6th of the combustion cylinder (60 degrees) is simulated. In light of the simulation results it is extremely recommended to use AMR approach in conjunction with embedding around the nozzle for running engine simulations.
IRJET- Numerical Simulation for Spray Characteristics of Diesel and Biodiesel...IRJET Journal
This document presents a numerical simulation study of the spray characteristics of diesel and biodiesel fuels. Advanced spray models with dynamic mesh refinement were used to simulate the atomization of diesel and biodiesel sprays. Real-time engine simulations using biodiesel fuel injection were conducted. The spray characteristics such as Sauter mean diameter, spray cone angle, spray tip penetration, and fuel-air mixture were studied. The simulation results matched well with experimental data, with error rates below 5% for most measurements. The study found that fuels with higher density exhibited inferior spray and atomization characteristics.
Modeling of assembly line balancing for optimized number of stations and timeIAEME Publication
This document summarizes research on modeling assembly line balancing to optimize the number of stations and cycle time. It begins by classifying assembly line balancing problems based on objectives and problem structure. It then presents the Buxey 29 task precedence diagram and solves the problem to find feasible solutions with minimum cycle time for different numbers of stations (8, 9, 10 stations). The results show the total time remains 324 seconds for the optimal solution as more stations are added, while the cycle time decreases from 41 to 34 seconds. Increasing stations improves flexibility but also costs, so solutions must consider objectives and constraints.
CFD Analysis of ATGM Configurations -- Zeus NumerixAbhishek Jain
Above Research Paper can be downloaded from www.zeusnumerix.com
The research paper aims to study the aerodynamic configuration of anti-tank guided missile (ATGM). Effect of rotation of the ATGM on it own axis is simulated along with the curved fins. Areas of flow separation and higher drag are visualized using the variation of axial force along the length. The paper emphasizes on the contribution of base drag to the total drag. Authors A Venkateshwarlu and Hem Raj (BDL), Kumar Mihir and Sanjay Kumar (Zeus Numerix), Prof KE Prasad JNTU.
Effect of Residual Modes on Dynamically Condensed Spacecraft StructureIRJET Journal
This document discusses the effect of residual modes on the fundamental frequencies of a condensed spacecraft structure. It presents the modeling and dynamic analysis of a spacecraft bus structure using finite element analysis. The structure is condensed using the Craig-Bampton method to reduce the degrees of freedom. Residual modes are then computed and included to recover data lost during condensation. The results show that including residual modes provides frequencies for the condensed structure that closely match those of the original full structure model, demonstrating the effectiveness of using residual modes for data recovery after structural condensation.
The document describes the Model Induced Metropolis-Hastings (MIMH) algorithm for efficiently sampling from high-performance regions of costly objective functions. MIMH performs Metropolis-Hastings random walks on a radial basis function network (RBFN) model of the objective function. After each walk, the endpoint is added to the RBFN training set to improve the model. Experiments show MIMH finds good solutions with significantly fewer objective function evaluations than other algorithms like Niching ES, and the number of evaluations can be reduced further by raising the acceptance probability exponent. MIMH provides an effective way to identify high-performance regions at low cost for initializing more greedy optimization methods.
Aerodynamic Drag Reduction for A Generic Sport Utility Vehicle Using Rear Suc...IJERA Editor
The high demand for new and improved aerodynamic drag reduction devices has led to the invention of flow control mechanisms and continuous suction is a promising strategy that does not have major impact on vehicle geometry. The implementation of this technique on sport utility vehicles (SUV) requires adequate choice of the size and location of the opening as well as the magnitude of the boundary suction velocity. In this paper we introduce a new methodology to identifying these parameters for maximum reduction in aerodynamic drag. The technique combines automatic modeling of the suction slit, computational fluid dynamics (CFD) and a global search method using orthogonal arrays. It is shown that a properly designed suction mechanism can reduce drag by up to 9%.
Parametric optimization of centrifugal magnetic force assisted abrasive flow ...eSAT Journals
Abstract: The traditional Taguchi method is widely used for optimizing the process parameters of a single response problem. Optimization of a single response results the non-optimum values for remaining. But, the performance of a machining process is often evaluated by several quality responses. Under such circumstances, multi-characteristics response optimization may be the solution to optimize multi-responses simultaneously. In the present work, a multi-characteristics response optimization model based on Taguchi and Utility concept is used to optimize process parameters, such as magnetic flux, rotational speed of CFG rod, shape of CFG rod, number of cycles, abrasive-to-iron ratio and abrasive particle size on multiple performance characteristics, namely, surface roughness (Ra) and material removal (MR) during polishing of hollow cylindrical brass work-pieces with Centrifugal-Magnetic Force Assisted Abrasive Flow Machining (CMA2FM) Process. Taguchi’s L27 orthogonal array (OA) is selected for experimental planning. The ANOVA and F-tests are used to analyze the results. It is found that all the input parameters significantly improve the Utility function comprising of two quality characteristics (MR and %ΔRa). Further, the confirmation tests are conducted and the results are found to be within the confidence interval.
Keywords: Abrasive Flow Machining (AFM), Centrifugal Force, Magnetic Force, CFG Rod, CMA2FM, Utility Concept,TaguchiMethod,MultiResponseOptimization
This document discusses improving the process capability of a connecting rod manufacturing process. It begins with defining the critical-to-quality characteristic as the bolt-holes center distance. Next, it measures the process capability indices including Cp, Cpk, and standard deviation. It then analyzes the data using ANOVA to identify sources of variation. Finally, it details improving the process by reducing standard deviation from 0.017 to 0.009 and increasing Cp from 0.97 to 1.77 and Cpk from 0.57 to 1.49 through applying statistical process control methods like process monitoring charts. The goal is to center the process within specifications to produce connecting rods meeting tight dimensional tolerances.
IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology.
Bidyut Kumar Midday is seeking a career in the civil aviation industry with over 20 years of experience maintaining rotary wing aircraft. He has experience servicing and repairing aircraft as well as maintaining tools and equipment. He holds qualifications including a diploma in mechanical engineering and national trade certificate along with licenses and approvals related to specific aircraft. Currently working as an aircraft maintenance engineer, he is seeking a new position as an overhaul, shop, field, service or sales engineer.
SIGMA TEST & RESEARCH CENTRE is one of the fastest growing organization in the field of calibration. We provide calibration services to our customers in accordance to IS/ISO/IEC-17025.
This document summarizes the results of lab tests that were conducted to validate a theoretical approach for using Quality Function Deployment (QFD) and Theory of Inventive Problem Solving (TRIZ) for component mathematics modeling in virtual durability simulations. The methodology involved identifying customer needs, evaluating solutions, applying QFD concepts, using TRIZ to solve issues, and analyzing failure modes and effects analysis (FMEA). Simulation results using different element types showed that using RBE3 elements instead of RBE2 elements better represented real-world behavior by allowing for lateral displacement. Simplified models were also developed and tested to validate force equivalence with full vehicle simulations.
IRJET- Optimization of Process Parameter in Injection Moulding using Tagu...IRJET Journal
This document summarizes a study that used Taguchi methods and ANOVA to optimize injection molding process parameters for a polypropylene chair bottom part. The researchers used Moldflow Adviser software to simulate different combinations of mold temperature, melt temperature, injection pressure, and cooling time based on an L9 orthogonal array. The goal was to minimize fill time and volumetric shrinkage. ANOVA results identified melt temperature as the most significant factor influencing both responses. The best optimized parameters found were a low melt temperature, low mold temperature, low injection pressure, and long cooling time.
Invited presentation to 2003 RAeS Aerodynamics research conferencestephen_mcparlin
The document summarizes an assessment process used to evaluate air vehicle concepts from initial design through analysis and testing. The process involves conceptual design using multidisciplinary variable optimization (MVO) followed by detailed configuration design using CAD tools. Performance is then analyzed using computational fluid dynamics (CFD) and wind tunnel testing. Underlying technologies like CAD integration, response surface methods, and CFD methods have improved the accuracy and speed of the assessment process.
Tuning the model predictive control of a crude distillation unitISA Interchange
Tuning the parameters of the Model Predictive Control (MPC) of an industrial Crude Distillation Unit (CDU) is considered here. A realistic scenario is depicted where the inputs of the CDU system have optimizing targets, which are provided by the Real Time Optimization layer of the control structure. It is considered the nominal case, in which both the CDU model and the MPC model are the same. The process outputs are controlled inside zones instead of at fixed set points. Then, the tuning procedure has to define the weights that penalize the output error with respect to the control zone, the weights that penalize the deviation of the inputs from their targets, as well as the weights that penalize the input moves. A tuning approach based on multi-objective optimization is proposed and applied to the MPC of the CDU system. The performance of the controller tuned with the proposed approach is compared through simulation with the results of an existing approach also based on multi-objective optimization. The simulation results are similar, but the proposed approach has a computational load significantly lower than the existing method. The tuning effort is also much lower than in the conventional practical approaches that are usually based on ad-hoc procedures.
The document describes a study that used an orthogonal experiment and numerical simulation to optimize the design of a centrifugal pump impeller. Five impeller geometric parameters were selected as factors to optimize: impeller outlet width, impeller inlet diameter, impeller blade wrapping angle, impeller blade outlet angle, and impeller blade inlet angle. Sixteen impeller designs were generated using an orthogonal table to vary the factors. Computational fluid dynamics was used to simulate the performance of pumps with each impeller design. Variance analysis identified the best parameter combination for improving pump efficiency and head. The optimized design showed significant improvements in efficiency and head compared to the original pump design.
This document summarizes a computational fluid dynamics (CFD) simulation of airflow around a simplified pickup van model. The study used a commercial CFD software to perform a 3D, steady-state simulation using the Reynolds-averaged Navier-Stokes equations and a k-ε turbulence model. The simulation was run at Reynolds numbers of 3x105 and results were validated against experimental wind tunnel data. Key findings included pressure and velocity distributions that matched experimental data well, indicating CFD can be an effective alternative to wind tunnel testing for pickup van aerodynamic analysis.
Compartmental modeling of high purity air separation columnsAbu Hussein
High purity distillation columns are critical unit operations in cryogenic air separation plants. The development of nonlinear control technology is motivated by the need to frequently change production rates in response to time varying utility costs. Detailed column models based on stage-by-stage balance equations are too complex to be incorporated directly into optimization-based strategies such as nonlinear model predictive control. In this paper, we develop reduced order dynamic models for the upper column of a cryogenic air separation plant by applying time scale arguments to a detailed stage-by-stage model that includes mass and energy balances and accounts for non-ideal vapor–liquid equilibrium. The column is divided into compartments according to the locations of liquid distributors and feed and withdrawal streams.
IRJET- Spray Characteristics of Diesel Fuel using Numerical SimulationIRJET Journal
This document discusses a numerical simulation of diesel fuel spray characteristics using computational fluid dynamics (CFD) software. The study aims to improve spray models by simulating atomization of diesel and bio-diesel sprays using dynamic mesh refinement. The simulation predicts spray parameters like Sauter mean diameter, spray cone angle, and spray tip penetration under different injection pressures. Results from the CFD model show good agreement with experimental data. The model can help advance engine design by analyzing spray behavior and mixture formation.
Numerical Simulation of Combustion Behavior of DI Diesel Engine with Conjunct...Khatir NAIMA
Currently, computational fluid dynamics has become an effective supplement to experimentation in the analysis and development of various engineering systems including internal combustion engines. In fact, multi-dimensional modelling of IC engines is less extensive and less time consuming than experimentation. In this aim, CONVERGE code was used to study the combustion behavior in a DI engine with various mesh control techniques including embedding and Adaptive Mesh Refinement (AMR). The simulation covers the compression, spray, combustion and expansion. A single spray plume and 1/6th of the combustion cylinder (60 degrees) is simulated. In light of the simulation results it is extremely recommended to use AMR approach in conjunction with embedding around the nozzle for running engine simulations.
IRJET- Numerical Simulation for Spray Characteristics of Diesel and Biodiesel...IRJET Journal
This document presents a numerical simulation study of the spray characteristics of diesel and biodiesel fuels. Advanced spray models with dynamic mesh refinement were used to simulate the atomization of diesel and biodiesel sprays. Real-time engine simulations using biodiesel fuel injection were conducted. The spray characteristics such as Sauter mean diameter, spray cone angle, spray tip penetration, and fuel-air mixture were studied. The simulation results matched well with experimental data, with error rates below 5% for most measurements. The study found that fuels with higher density exhibited inferior spray and atomization characteristics.
Modeling of assembly line balancing for optimized number of stations and timeIAEME Publication
This document summarizes research on modeling assembly line balancing to optimize the number of stations and cycle time. It begins by classifying assembly line balancing problems based on objectives and problem structure. It then presents the Buxey 29 task precedence diagram and solves the problem to find feasible solutions with minimum cycle time for different numbers of stations (8, 9, 10 stations). The results show the total time remains 324 seconds for the optimal solution as more stations are added, while the cycle time decreases from 41 to 34 seconds. Increasing stations improves flexibility but also costs, so solutions must consider objectives and constraints.
CFD Analysis of ATGM Configurations -- Zeus NumerixAbhishek Jain
Above Research Paper can be downloaded from www.zeusnumerix.com
The research paper aims to study the aerodynamic configuration of anti-tank guided missile (ATGM). Effect of rotation of the ATGM on it own axis is simulated along with the curved fins. Areas of flow separation and higher drag are visualized using the variation of axial force along the length. The paper emphasizes on the contribution of base drag to the total drag. Authors A Venkateshwarlu and Hem Raj (BDL), Kumar Mihir and Sanjay Kumar (Zeus Numerix), Prof KE Prasad JNTU.
Effect of Residual Modes on Dynamically Condensed Spacecraft StructureIRJET Journal
This document discusses the effect of residual modes on the fundamental frequencies of a condensed spacecraft structure. It presents the modeling and dynamic analysis of a spacecraft bus structure using finite element analysis. The structure is condensed using the Craig-Bampton method to reduce the degrees of freedom. Residual modes are then computed and included to recover data lost during condensation. The results show that including residual modes provides frequencies for the condensed structure that closely match those of the original full structure model, demonstrating the effectiveness of using residual modes for data recovery after structural condensation.
The document describes the Model Induced Metropolis-Hastings (MIMH) algorithm for efficiently sampling from high-performance regions of costly objective functions. MIMH performs Metropolis-Hastings random walks on a radial basis function network (RBFN) model of the objective function. After each walk, the endpoint is added to the RBFN training set to improve the model. Experiments show MIMH finds good solutions with significantly fewer objective function evaluations than other algorithms like Niching ES, and the number of evaluations can be reduced further by raising the acceptance probability exponent. MIMH provides an effective way to identify high-performance regions at low cost for initializing more greedy optimization methods.
Aerodynamic Drag Reduction for A Generic Sport Utility Vehicle Using Rear Suc...IJERA Editor
The high demand for new and improved aerodynamic drag reduction devices has led to the invention of flow control mechanisms and continuous suction is a promising strategy that does not have major impact on vehicle geometry. The implementation of this technique on sport utility vehicles (SUV) requires adequate choice of the size and location of the opening as well as the magnitude of the boundary suction velocity. In this paper we introduce a new methodology to identifying these parameters for maximum reduction in aerodynamic drag. The technique combines automatic modeling of the suction slit, computational fluid dynamics (CFD) and a global search method using orthogonal arrays. It is shown that a properly designed suction mechanism can reduce drag by up to 9%.
Parametric optimization of centrifugal magnetic force assisted abrasive flow ...eSAT Journals
Abstract: The traditional Taguchi method is widely used for optimizing the process parameters of a single response problem. Optimization of a single response results the non-optimum values for remaining. But, the performance of a machining process is often evaluated by several quality responses. Under such circumstances, multi-characteristics response optimization may be the solution to optimize multi-responses simultaneously. In the present work, a multi-characteristics response optimization model based on Taguchi and Utility concept is used to optimize process parameters, such as magnetic flux, rotational speed of CFG rod, shape of CFG rod, number of cycles, abrasive-to-iron ratio and abrasive particle size on multiple performance characteristics, namely, surface roughness (Ra) and material removal (MR) during polishing of hollow cylindrical brass work-pieces with Centrifugal-Magnetic Force Assisted Abrasive Flow Machining (CMA2FM) Process. Taguchi’s L27 orthogonal array (OA) is selected for experimental planning. The ANOVA and F-tests are used to analyze the results. It is found that all the input parameters significantly improve the Utility function comprising of two quality characteristics (MR and %ΔRa). Further, the confirmation tests are conducted and the results are found to be within the confidence interval.
Keywords: Abrasive Flow Machining (AFM), Centrifugal Force, Magnetic Force, CFG Rod, CMA2FM, Utility Concept,TaguchiMethod,MultiResponseOptimization
This document discusses improving the process capability of a connecting rod manufacturing process. It begins with defining the critical-to-quality characteristic as the bolt-holes center distance. Next, it measures the process capability indices including Cp, Cpk, and standard deviation. It then analyzes the data using ANOVA to identify sources of variation. Finally, it details improving the process by reducing standard deviation from 0.017 to 0.009 and increasing Cp from 0.97 to 1.77 and Cpk from 0.57 to 1.49 through applying statistical process control methods like process monitoring charts. The goal is to center the process within specifications to produce connecting rods meeting tight dimensional tolerances.
IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology.
Bidyut Kumar Midday is seeking a career in the civil aviation industry with over 20 years of experience maintaining rotary wing aircraft. He has experience servicing and repairing aircraft as well as maintaining tools and equipment. He holds qualifications including a diploma in mechanical engineering and national trade certificate along with licenses and approvals related to specific aircraft. Currently working as an aircraft maintenance engineer, he is seeking a new position as an overhaul, shop, field, service or sales engineer.
SIGMA TEST & RESEARCH CENTRE is one of the fastest growing organization in the field of calibration. We provide calibration services to our customers in accordance to IS/ISO/IEC-17025.
Neeraj Kumar has over 4 years of experience as a Maintenance Engineer. He currently works for Moser Baer India Ltd, a leading manufacturer of optical storage media. In his role, he is responsible for operation and parameter setting on injection molding machines, handling maintenance and troubleshooting to reduce downtime. He also coordinates with various departments to ensure smooth operations and achieve production goals. Neeraj holds a diploma and bachelor's degree in mechanical engineering and provides details of his educational qualifications, computer skills, expected salary and personal details in his curriculum vitae.
This resume is for Venkatesan Govindharajan, who has over 10 years of experience in powertrain emissions and calibration engineering for diesel and gasoline engines. He is currently a Powertrain Emissions & OBD Calibration Engineer at Ford Motor Company, where he is responsible for vehicle calibration, testing, and validation to meet functional and regulatory requirements. Previously, he held engineering roles at Daimler India, Hyundai Motor India, and the University of Michigan, where he gained experience in engine development, emissions certification, vehicle testing, and research projects related to vehicle thermal management and noise vibration harshness performance.
The document provides an overview of Varun Verma's internship at Honeywell Turbo Technologies in Pune, India from June 11 to July 11, 2012. It includes sections on the company overview, project description of developing an integrated database of inspection equipment data, work summary of the internship, learning experiences, and conclusions. The annexures provide details of the inspection equipment processed and comments on calibration parameters.
This document provides a summary of an individual's professional skills and experience working on Cummins diesel engines. It outlines their career history as a Senior Supervisor and Service Engineer for authorized Cummins dealerships in Oman and Kerala, India from 2010 to present. Their responsibilities included overhauling, troubleshooting, and maintenance of various Cummins engine models and applications. Key competencies highlighted include experience with Cummins software tools like INSITE and INPOWER, operating Cummins controllers, and completing engine training programs in India.
Mahesh Deshmukh is a mechanical engineer with over 3 years of experience in plant maintenance at Denso Haryana Pvt Ltd. He currently holds the position of Senior Engineer and oversees maintenance activities including preventative maintenance planning and breakdown repairs. Prior to this role, he gained experience working on injection molding lines, presses, CNC machines and more. Mahesh holds a Bachelor of Technology degree in Mechanical Engineering and has undergone training in pneumatics, PLCs, injection molding and auditing standards. He is a dynamic professional seeking to contribute his maintenance and problem solving skills.
Nandini Gowda Kodebyle Raju is seeking a position in research and development of fuel efficient and low emission vehicles. She has a Master's in Automotive Engineering from Clemson University and a Bachelor's in Industrial Engineering from Visvesvaraya Technological University in India. Her skills include modeling software like GT Power and languages like English, German, Hindi, and Kannada. She has experience in projects modeling engine systems, vehicle integration, and developing concepts like a coffee harvesting machine and fog detection system. Her professional experience includes executing vehicle launch projects at Toyota Kirloskar Motor in India. She also received a scholarship from Southern Automotive Women's Forum.
K. Sivakumar is seeking a position that utilizes his experience in various mechanical fields. He has over 20 years of experience as a diesel mechanic in India and overseas, working on vehicles and heavy equipment. His experience includes repairing and overhauling engines, transmissions, differentials, and other mechanical components. He is skilled in troubleshooting, maintenance, and repair of engines, transmissions, brakes, and other systems. Sivakumar holds an S.S.L.C. from S.M.H.H.S. School and is licensed for heavy vehicle driving in India.
This document is a resume for John Sonnichsen, a mechanical engineer with over 16 years of experience in engine development, design, calibration, and testing. He currently works as a manager at Hyundai-Kia America Technical Center, where he leads a team that designs and tests new engine technologies. Previously, he held chief engineer and manager roles at Fiat Chrysler Automobiles and Ford Motor Company, managing multimillion dollar engine programs and international teams across various cultures.
Sreejith CS is seeking a job that allows him to improve his skills and knowledge. He has experience working for Hindustan Aeronautics Limited inspecting and testing industrial and marine gas turbines. He has a Bachelor's Degree in Mechanical Engineering and internship experience in engine management and flow meter testing. His skills include CAD/CAM software, C programming, and machine operation.
Rajas Joshi has over 12 years of experience in global quality control and operations management for diesel engine manufacturing. As General Manager of Global Product Quality at Cummins, Inc., he improved product reliability in key markets, reducing warranty returns by 40% and directing teams that achieved $50 million in annual savings through process improvements. Previously he held roles in quality assurance, new product development, and service engineering at Cummins, consistently delivering cost reductions and performance gains.
Waristha Ubolbandit is a 27-year-old project engineer with experience researching engine control systems and developing marine turbine prototypes. She has worked at Advanced Marine Energy since 2016 overseeing quality control, budgeting, and reporting. Previously, she was a mechanical engineer at Toyota developing engine calibration software to improve performance and fuel efficiency. Waristha has a bachelor's degree in mechatronics engineering and speaks English and Japanese.
Saura Jyoti Dhar has over 9 years of experience in embedded systems for the automotive domain. He currently works as a Product Line Architect at Robert Bosch Engineering and Business Solutions developing next-generation multi-core ECU architectures. Previously he has worked on developing engine management systems at Bosch and Honeywell. He has expertise in AUTOSAR standards, ALM tools, and automotive software development processes.
Representative Testing of Emissions and Fuel Consumption of Working Machines ...Reno Filla
Out of necessity, emission and fuel consumption test cycles are a simplified representation of the real-life use of a vehicle or component that is assumed to be most common. In reality, variations are introduced by both the driver and the environment – and to a lesser degree also by the vehicle itself through performance deviations because of tolerances in the components’ characteristics. However, since such simplified test cycles exist and are accepted (or even required by law), OEMs tend to use them also in product development to benchmark their products against the competition, and to make decisions on how to optimize design.
While this approach might give acceptable results for on-road vehicles, it fails to capture reality in the case of versatile working machines. Here, the variety of possible applications cannot be covered by one common application but rather demands a mix of several cycles. This has a large impact on the setup and evaluation of physical and virtual testing of working machines, especially those with alternative power systems like hybrids, which will be discussed in the paper.
Attempts to simplify the complexity of real-life applications are made in some testing standards and standard proposals, but they over-simplify with the result of prescribing a common cycle that is not representative and which therefore, however tempting, cannot be used for any meaningful benchmarking of emissions and fuel consumption.
http://papers.sae.org/2012-01-1946
Efficient and Effective CFD Design Flow for Internal Combustion EnginesReaction Design
1) Traditional IC engine CFD simulations use simplified chemistry models that are insufficient for designing new high-efficiency engine concepts where kinetics effects dominate.
2) Detailed chemical kinetics models have been developed through the Model Fuels Consortium but cannot be directly incorporated into CFD due to high computational cost.
3) Reaction Design's CFD software uses novel parallel chemistry solvers that dramatically reduce simulation times, allowing the use of detailed chemical kinetics models for predictive 3D engine simulations.
This article presents the discrete event simulation developed with Northwestern University to optimize an injector production line. This article was presented at the recent Winter simulation Conference (held in December 2008)
Optimization of time step and cfd study of combustion in di diesel engineeSAT Publishing House
IJRET : International Journal of Research in Engineering and Technology is an international peer reviewed, online journal published by eSAT Publishing House for the enhancement of research in various disciplines of Engineering and Technology. The aim and scope of the journal is to provide an academic medium and an important reference for the advancement and dissemination of research results that support high-level learning, teaching and research in the fields of Engineering and Technology. We bring together Scientists, Academician, Field Engineers, Scholars and Students of related fields of Engineering and Technology
Iaetsd computer simulation of compression ignition engine through matlabIaetsd Iaetsd
This document discusses a computer simulation of a compression ignition engine using MATLAB. The simulation models the engine cycle in two zones - a burned zone and unburned zone. It uses a Wiebe function to determine the mass fraction burned and calculates parameters like pressure, temperature, heat release and emissions over the engine cycle. The simulation is first validated against experimental engine test data. Sensitivity studies are then conducted by varying combustion model constants to better understand their impact on predictions and combustion mechanisms. The simulation calculates performance parameters like brake power, fuel consumption and efficiencies over the engine's operating range. It aims to provide insights into combustion and pollutant formation at different loads and injection timings.
Optimization of Injection Molding Process-literature reviewAlex Larsh
The document discusses optimization of the injection molding process through various techniques including computer simulation, artificial neural networks (ANN), and genetic algorithms (GA). It presents a combining ANN/GA method to model and predict part quality, and optimize process conditions. A novel computer simulation–stochastic data envelopment analysis (CS-SDEA) algorithm is also proposed to optimize layout of an injection molding process. Results show the CS-SDEA algorithm performs comparably to ANN and GA methods in optimizing layout alternatives.
Response Surface Optimization of Chemical Additives and Engine Parameters on ...IRJET Journal
The document discusses using response surface methodology and the Taguchi method to optimize engine performance parameters including hydrogen percentage, nanoparticles, ignition pressure, and ignition timing. The study found that 25% engine load, 20% hydrogen, 50 ppm nanoparticles, 220 bar ignition pressure, and 31 degrees before top dead center ignition timing optimized brake thermal efficiency while minimizing emissions and fuel consumption. Validation tests found the response surface methodology models' predictions to be within acceptable error ranges of experimental results. The methodology can be applied to further optimize additional variables and develop a more holistic engine performance model.
IRJET- Optimization of Plastic Injection MoldingIRJET Journal
The document discusses optimization of plastic injection molding process parameters using Taguchi methods and Moldflow simulation software. It analyzes the effect of melt temperature, cooling time, injection pressure, and mold temperature on the warpage of a honey cup part made of polypropylene. An L9 orthogonal array experiment is designed using the Taguchi approach to study these four parameters at three levels each. The results show that melt temperature has the most significant effect on warpage, followed by mold temperature, injection pressure, and cooling time. The optimal parameter settings found to minimize warpage are a melt temperature of 180°C, cooling time of 15 seconds, injection pressure of 90MPa, and mold temperature of 50°C.
Soot Formation in Diesel Engines By Using CfdIJERA Editor
In order to meet the stringent emission standards significant efforts have been imparted to the research and
development of cleaner IC engines. Diesel combustion and the formation of pollutants are directly influenced by
spatial and temporal distribution of the fuel injected. The development and validation of computational fluid
dynamics (CFD) models for diesel engine combustion and emissions is described. The complexity of diesel
combustion requires simulations with many complex interacting sub models in order to have a success in
improving the performance and to reduce the emissions. In the present work an attempt has been made to
develop a multidimensional axe-symmetric model for CI engine combustion and emissions. Later simulations
have been carried out. Commercial validation tool FLUENT was used for simulation. The tool solves basic
governing equations of fluid flow that is continuity, momentum, species transport and energy equation. Using
finite volume method turbulence was modeled by using RNG K-ɛ model. Injection was modeled using La
Grangian approach and reaction was modeled using non premixed combustion which considers the effects of
turbulence and detailed chemical mechanism into account to model the reaction rates. The specific heats were
approximated using piecewise polynomials. Subsequently the simulated results have been validated with the
existing experimental values
DEA is a non-parametric technique used to measure the relative efficiency of decision making units (DMUs) that use multiple inputs to produce multiple outputs. It works by constructing a production frontier boundary comprised of the most efficient DMUs to evaluate how efficiently other DMUs use inputs to produce outputs. The methodology was originally developed in 1978 and has since been applied in various industries to evaluate organizations, identify best practices, and determine potential efficiency improvements for inefficient units.
This paper describes a generic process for conceptual design and assessment of air vehicles that uses modern computational tools. The process involves conceptual design using multi-variate optimization to meet requirements, followed by detailed 3D design and analysis using computational fluid dynamics and wind tunnel testing to validate performance. Underlying technologies like parametric CAD, design of experiments, and response surface modeling have improved the speed, accuracy, and flexibility of the assessment process.
This document describes the BMW Group's use of computational fluid dynamics (CFD) simulations in the aerodynamic development of passenger cars and motorcycles. It discusses how CFD is used at different stages of the development process to address various aerodynamic questions. It also provides details on BMW's simulation process using a Lattice-Boltzmann code, shows examples of validation cases demonstrating the achievable accuracy, and highlights practical applications of CFD including assessing aerodynamic loads, passenger comfort, and thermal management.
International Journal of Business and Management Invention (IJBMI) is an international journal intended for professionals and researchers in all fields of Business and Management. IJBMI publishes research articles and reviews within the whole field Business and Management, new teaching methods, assessment, validation and the impact of new technologies and it will continue to provide information on the latest trends and developments in this ever-expanding subject. The publications of papers are selected through double peer reviewed to ensure originality, relevance, and readability. The articles published in our journal can be accessed online.
This document contains instructions for a final examination in an introduction to software engineering course. It provides the date, time, location of the exam, as well as instructions that students are to write their name, student ID, and signature on the cover and top of the exam. It also states that the exam is closed book and notes, calculators are permitted, and to circle one answer for multiple choice questions. The exam contains two sections - a multiple choice section and an essay question section where students should write their answers in a separate booklet.
Test Scheduling of Core Based SOC Using Greedy AlgorithmIJERA Editor
Escalating increase in the level of integration has led the design engineers to embed the pre-design and pre-verified logic blocks on chip to make a complete system on chip (SoC) technology. This advancing technology trend has led to new challenges for the design and test engineers. To ensure the testability of the entire system, the test planning needs to be done during design phase. To save the test cost, the test application time needs to be reduced which requires the test to be done concurrently. However the parallel running of test of multiple cores increases the power dissipation. This thereby leads to make test optimization to take care of time and power. This paper presents an approach for the scheduling the cores with the test time, power, test access mechanism and bandwidth constraint based on greedy algorithm. The TAM allotment to the various cores is done dynamically to save the test time and utilize the full bandwidth. Scheduling is done on ITC’02 benchmark circuits. Experiments on these ITC’02 benchmark circuits show that this algorithm offers lower test application time compared to the multiple constraint driven system-on-chip.
IRJET- Enhancing Line Efficiency of Road Machinery Assembly Line at Volvo...IRJET Journal
This document discusses enhancing the line efficiency of an assembly line that produces road machinery at Volvo Construction Equipment in Bengaluru, India. It begins by describing the existing assembly line layout, which includes 7 main assembly stations and 9 feeder stations. The objectives are to determine the current efficiency and improve it using line balancing techniques. Time studies were conducted to collect processing times for tasks. Heuristic techniques like Ranked Positional Weight method, Largest Candidate Rule method, and Kilbridge and Wester Column method were then used to rebalance the line and assign tasks in a way that meets precedence and cycle time constraints while minimizing balance delay. The goal is to deliver products to customers faster while maintaining quality.
This document provides an overview of operational excellence and design of experiments (DOE). It defines key DOE terms and concepts, including factors, levels, interactions, resolution, coding/decoding variables. It discusses the objectives of different DOE designs (screening, modeling, optimizing) and considerations for choosing a design based on factors, levels, and resources. Guidelines are given for planning, executing, and analyzing a DOE. Examples are provided to illustrate DOE concepts like resolution, coding variables, and a full factorial design. The overall purpose is to introduce the reader to the technique of DOE for improving processes.
The increase in design sizes and the complexity of timing checks at 40nm technology nodes
and below is responsible for longer run times, high memory requirements, and the need for a
growing set of gate-level simulation (GLS) applications including design for test (DFT) and lowpower considerations. As a result, in order to complete the verification requirements on time,
it becomes extremely important for GLS to be started as early in the design cycle as possible,
and for the simulator to be run in high-performance mode. This application note describes
new methodologies and simulator use models that increase GLS productivity, focusing on two
techniques for GLS to make the verification process more effective
IRJET- The RTL Model of a Reconfigurable Pipelined MCMIRJET Journal
This document presents two Register Transfer Level (RTL) models for a reconfigurable pipelined multiple constant multiplier (MCM) architecture. The first model uses a ripple carry adder, while the second model replaces it with a carry lookahead adder. Both models are constructed using Verilog and simulated using ModelSim. Simulation results show the second model has a 3.53% higher maximum clock frequency and throughput due to the faster carry lookahead adder. It also has 3.53% lower latency but uses 1.65% more power. The second model provides faster performance with moderate increases in resource usage and power consumption.
KuberTENes Birthday Bash Guadalajara - K8sGPT first impressionsVictor Morales
K8sGPT is a tool that analyzes and diagnoses Kubernetes clusters. This presentation was used to share the requirements and dependencies to deploy K8sGPT in a local environment.
Hierarchical Digital Twin of a Naval Power SystemKerry Sado
A hierarchical digital twin of a Naval DC power system has been developed and experimentally verified. Similar to other state-of-the-art digital twins, this technology creates a digital replica of the physical system executed in real-time or faster, which can modify hardware controls. However, its advantage stems from distributing computational efforts by utilizing a hierarchical structure composed of lower-level digital twin blocks and a higher-level system digital twin. Each digital twin block is associated with a physical subsystem of the hardware and communicates with a singular system digital twin, which creates a system-level response. By extracting information from each level of the hierarchy, power system controls of the hardware were reconfigured autonomously. This hierarchical digital twin development offers several advantages over other digital twins, particularly in the field of naval power systems. The hierarchical structure allows for greater computational efficiency and scalability while the ability to autonomously reconfigure hardware controls offers increased flexibility and responsiveness. The hierarchical decomposition and models utilized were well aligned with the physical twin, as indicated by the maximum deviations between the developed digital twin hierarchy and the hardware.
A review on techniques and modelling methodologies used for checking electrom...nooriasukmaningtyas
The proper function of the integrated circuit (IC) in an inhibiting electromagnetic environment has always been a serious concern throughout the decades of revolution in the world of electronics, from disjunct devices to today’s integrated circuit technology, where billions of transistors are combined on a single chip. The automotive industry and smart vehicles in particular, are confronting design issues such as being prone to electromagnetic interference (EMI). Electronic control devices calculate incorrect outputs because of EMI and sensors give misleading values which can prove fatal in case of automotives. In this paper, the authors have non exhaustively tried to review research work concerned with the investigation of EMI in ICs and prediction of this EMI using various modelling methodologies and measurement setups.
ACEP Magazine edition 4th launched on 05.06.2024Rahul
This document provides information about the third edition of the magazine "Sthapatya" published by the Association of Civil Engineers (Practicing) Aurangabad. It includes messages from current and past presidents of ACEP, memories and photos from past ACEP events, information on life time achievement awards given by ACEP, and a technical article on concrete maintenance, repairs and strengthening. The document highlights activities of ACEP and provides a technical educational article for members.
Understanding Inductive Bias in Machine LearningSUTEJAS
This presentation explores the concept of inductive bias in machine learning. It explains how algorithms come with built-in assumptions and preferences that guide the learning process. You'll learn about the different types of inductive bias and how they can impact the performance and generalizability of machine learning models.
The presentation also covers the positive and negative aspects of inductive bias, along with strategies for mitigating potential drawbacks. We'll explore examples of how bias manifests in algorithms like neural networks and decision trees.
By understanding inductive bias, you can gain valuable insights into how machine learning models work and make informed decisions when building and deploying them.
Harnessing WebAssembly for Real-time Stateless Streaming PipelinesChristina Lin
Traditionally, dealing with real-time data pipelines has involved significant overhead, even for straightforward tasks like data transformation or masking. However, in this talk, we’ll venture into the dynamic realm of WebAssembly (WASM) and discover how it can revolutionize the creation of stateless streaming pipelines within a Kafka (Redpanda) broker. These pipelines are adept at managing low-latency, high-data-volume scenarios.
Harnessing WebAssembly for Real-time Stateless Streaming Pipelines
9 coldengine
1. 1
Cold Engine Emissions Optimization Using Model Based
Calibration
Clive Tindle
General Motors Holden Ltd
ABSTRACT
Emissions calibration development of gasoline engines is becoming increasingly demanding for the calibration engineer,
due to a number of factors: reduced time for development programs, lower exhaust tailpipe emissions, and challenges to
reduce catalyst cost. To meet this demand, the technological complexity of the gasoline engine has increased with the
introduction of continuously variable intake and exhaust camshafts, and more recently the introduction of fuel injection
strategies on direct injection engines, such as double injection. With these technologies, engine data generation and
calibration optimization must be handled using a model based approach and design of experiments (DOE).
This paper focuses on the application of design of experiments methods and optimization of two-stage statistical
response models to develop an engine calibration to current worldwide standards, using MATLAB®
and Model Based
Calibration Toolbox. The model based methods used have been shown to be capable of producing calibration values for
the main actuators. Examples are presented which relate to recent applications during vehicle development. These show
that the use of model based methods is no longer a luxury, but a necessity in engine calibration.
INTRODUCTION
Virtually all the harmful pollutants emitted by modern gasoline powered vehicles are emitted during the initial engine start
and warm-up phases before catalyst light-off. Traditionally, faster catalyst light-off was achieved through retarding of
ignition timing and fuelling strategies which encouraged lower light-off temperatures. In response to more stringent
legislated emissions standards, changes in exhaust system architecture such as catalyst size, precious metal loadings
and placement have had an important role in reducing harmful pollutants. Their benefits can often be limited due to
vehicle packaging and cost restraints. At the same time, new engine technologies have been introduced to reduce engine
out emissions. Variable valve timing allows flexible control of internal exhaust gas recirculation (EGR), which can help
reduce nitrogen oxides (NOx) emissions [1]. In addition, if valve events are timed in such a manner that encourages
exhaust gas backflow into the intake ports, fuel preparation can be improved resulting in lower hydrocarbon emissions
(HC) [2 and 3]. Direct fuel injection introduces more degrees of freedom into the system and if used effectively can be an
emissions-reducing device. Semi-stratified charge combustion is realized through a double injection strategy which injects
some of the fuel during the compression stroke. The benefits are a reduction in cycle to cycle torque fluctuations and the
ability to further retard spark timing compared to homogeneous combustion [4]. The challenge to the engineer is to
optimize these high degree of freedom systems in a manner which produces the most efficient engine operation.
Model based calibration refers to the process of using DOE methods, statistical modeling and optimization to generate an
engine calibration. Figure 1 (courtesy of The MathWorks Inc.) shows a pictorial representation of the subtasks involved in
the process. The experimental plan is devised through application of advanced DOE methods. Statistical modeling uses
data collected from the experimental plan to produce accurate response models. High quality engine calibrations are then
developed through optimization of these models and system and calibration verification.
Traditional calibration methods have focused on optimizing a single variable at a time on the engine test bed which often
negates the interaction between other input variables. With higher degrees of freedom this becomes a very time
consuming and inefficient process. Introduction of the model based calibration approach has made it possible to optimize
all degrees of freedom simultaneously to enable a complete systems approach. This paper demonstrates the application
of model based calibration and discusses the best approach for each subtask based on recent experience at General
Motors Holden Ltd. This process has resulted in a successful emissions calibration that satisfied the programs objectives.
2. 2
Figure 1. Model-Based Calibration process using Mathworks tools
DESIGN OF EXPERIMENTS
In this section, the options available when designing an appropriate experiment are discussed. Experience has shown
that careful selection of appropriate design and of input variable range are the key contributing factors to success in the
model based process [5].
DOE SELECTION
Before selecting a DOE, it is important to consider the type of model that will be fitted to the data. The type of model
selected will influence the number of points required to be tested and will also determine to what extent the curvature of a
given response can best be modeled. For example, if a response feature can be described adequately using a quadratic
model, and we know that three test points are required to generate the coefficients of such a model, then it may seem
pointless to test at more than three levels for each input variable. If the requirement is to improve modeling of the
curvature of the response, then deviation from a quadratic model will be necessary and more than three levels are
required for each factor.
Summarized in Table 1 is the effect of the number of input variables (called factors) on the number of tests for a full
factorial and central composite design (CCD).
Table 1. Summary of tests required for different designs.
# of input variables
Quadratic – Full
Factorial
Quadratic/Cubic
CCD*
3 27 15
8 6561 273
*CCD uses 1 center point only
Depending upon the number of input variables in the experiment it may or may not be practical to use a full factorial DOE.
Considering three input variables, either a full factorial approach or a central composite design could be used. However,
to run a full factorial with eight input variables would involve running 6,561 tests, which in almost all instances would be
unpractical due to time and resource constraints. For the same reasons, it may also not be practical to run a CCD with
273 points. Figure 2 shows the CCD layout for three and eight variables, with Variable 2 plotted against Variable 1.
Other experimental designs offer more flexibility in terms of choosing the number of points. The limitation or overhead with
the CCD is that it is based on a 28
factorial design with the addition of star and center points. The axial distance of the star
points from the center point, known as the ‘alpha’ ratio, increases with the number of input variables. If the star points are
to fall within the variable ranges (known as an inscribed star point design), then the factorial points become concentrated
around the centre of the design space.
Data ModelingExperiment Design
Calibration Implementation
Data Collection
3. 3
Comparison of 3 and 8 variable CCD
-1
-0.8
-0.6
-0.4
-0.2
0
0.2
0.4
0.6
0.8
1
-1 -0.8 -0.6 -0.4 -0.2 0 0.2 0.4 0.6 0.8 1
Variable 1
Variable2
3 Variable CCD
8 Variable CCD
Star
Point
Figure 2 – CCD layout for three and eight variables.
This concentration is undesirable from a modeling standpoint, as large areas without test points exist within the design
space. A more flexible design that takes into consideration the number and type of terms in the model is the optimal
design. This design is computer-based and offers good flexibility especially with constrained design spaces [6]. The
minimum number of points in an optimal design is equal to the number of terms in the model. Table 2 below summarizes
some options when choosing an optimal design compared to a CCD.
Table 2. Comparison of optimal and CCD
DOE Model Minimum # of terms
CCD Cubic 273
Optimal Cubic 165
Clearly, there are savings on the number of test points using comparable models. However, although fewer DOE test
points will save critical test bed testing time, it is worth remembering that the model fit may deteriorate as a result. To
decide how many test points are required, analysis of a statistic known as Prediction Error Variance (PEV) can be used.
Figure 3 shows PEV for a typical design. The aim is to achieve a small PEV; values greater than 1 magnify the predicted
error of the model chosen before any data has been collected [7].
Figure 3. PEV for a typical design.
The dark red areas show PEV > 1. These areas are at the edge of the design space, so deterioration in model accuracy
can be expected. If necessary, extra points can be added through design augmentation, with a space-filling design to
reduce PEV. Alternatively, if the optimum solution ends up in an area where PEV is high, a second DOE can be
conducted with revised input variable ranges.
4. 4
To summarize, a full factorial design with three levels or a CCD is adequate for cases using three or fewer input variables.
As the number of input variables increases, the full factorial design becomes impractical, and other options must be
exploited. A CCD becomes less flexible as the number of input variables increases due to the concentration of points
towards the center of the design space (within inscribed star points). Both optimal and space-filling designs are best
suited for a high number of input variables, given limitations on the number of tests that can be conducted.
SCREENING EXPERIMENTS FOR INPUT VARIABLE SELECTION AND RANGE SETTING
Failure to select appropriate variables and ranges for the DOE can result in tests being run at unstable operating points
and optimum values being found at the edge of the design space. These issues may appear trivial at first, especially if the
DOE is going to be used on an existing application that has history. Even so, it is easy to make the DOE either too
complex, by adding variables which have no effect on the responses, or too simple, by missing one or more key variables.
For a new application that has additional complexity, the same mistakes are easier to make as there is no previous
application knowledge. A general approach used by statisticians is to run a screening experiment with as many relevant
input variables as possible over reasonably large ranges. Analysis of this data will narrow down the relevant input
variables and give a good indication of their ranges. An additional benefit is that relevant data from the screening
experiment can also be used for model fitting, by augmenting the base design or for the purposes of model validation. In
our experience, screening experiments have been used in such cases but have only been necessary for systems with
large degrees of freedom.
RESPONSE SURFACE CHARACTERIZATION
Response surface characterization refers to the process of modeling experimental data that has a number of input
variables. The response is “characterized” or represented by mathematical equations that can be reconstructed to
visualize a surface. This section discusses the techniques used to successfully model the experimental data. Examples of
different responses are used to help explain the approach and provide insight into engine behavior.
The DOE data points were collected on the engine dynamometer, using a standard test automation system. Common
emissions variables and combustion parameters were measured under steady state operating conditions. The purpose of
the exercise was to optimize engine behavior during the warm-up phase. In order to best characterize this, the engine’s
coolant and oil were maintained at 25 ˚C for all test points. This was achieved by using an external refrigeration plant
connected to the engine’s coolant system. All mapping tests were performed at this quasi-steady thermal condition; based
on previous studies [8], this provides more representative data than performing tests with fully warm fluids.
TWO-STAGE REGRESSION APPROACH
The two-stage model approach allows for separation of one or more of the input variables into a ‘first stage’, with the
remainder falling into the ‘second stage.’ It works best when there is prior knowledge of how a response changes with a
given input(s). For gasoline engines, a typical first-stage input would be spark timing, as the majority of responses – brake
torque, emissions, combustion stability, and so forth – can be seen to have approximate, quadratic, or cubic relationships
with spark timing. This type of approach has been shown by others in the engine mapping community to be beneficial [9]
and fits in well with how engine data has been traditionally collected for common engine mapping procedures, such as
torque model calibration.
There are two valuable benefits from the two-stage approach. First, it gives the engineer an opportunity to have a clear
understanding of the factor being varied – in this case spark, at each DOE (second stage) operating condition. This leads
to easier identification of outliers within the test data. Second, the complexity of the DOE can be reduced by one factor,
which will lead to reduced model complexity. Performing sweeps will add to the total number of test points, but the above
benefits far outweigh this disadvantage. Each spark sweep may be considered as a ‘local model’ or ‘first stage,’ and each
DOE test point as the ‘global model’ or ‘second stage.’
The Local Model
The local model is generated from data collected at fixed global settings, and one variable – spark – is varied from a
reference point, for example, maximum brake torque (MBT), towards a retarded spark condition. A mathematical model
that best describes the response’s characteristics – for example, a polynomial – can then be fitted to the data. An
example spark sweep is shown in Figure 4, with a cubic fit to engine-out hydrocarbon emissions (HC). Higher-order
polynomials might better fit the raw data points, but they would more than likely model the noise within the measurement
system and hence not be beneficial. Also shown in Figure 4 is a data outlier; without knowledge of adjacent test points,
identifying this as an outlier would be much more difficult.
5. 5
- 3 0 - 2 0 - 1 0 0 1 0 2 0 3 0
0
0 . 0 0 5
0 . 0 1
0 . 0 1 5
0 . 0 2
0 . 0 2 5
0 . 0 3
0 . 0 3 5
0 . 0 4
Spark Retard
HCemissions
Local Response Model for HC
Outlier
Towards MBT
Figure 4. Spark sweep data for derivation of local model
A cubic model applied to the data shown in Figure 4 takes on the mathematical form:
3
3
2
210 SPARKbSPARKbSPARKbbHC +++=
where, bi are the regression coefficients. The fit of the polynomial at the local stage can be assessed through analysis of,
say, the root mean square error (RMSE). This is performed after all outliers have been removed. At this stage, different
order polynomials can be compared using RMSE values. It is also important from the calibration engineer’s point of view
to scrutinize this model and perhaps think about whether this type (shape) of behavior is expected and follows the laws of
physics. This model also makes available the limitations and expectations of the engine.
The Global Model
The purpose of the global model is to combine the collection of local models (i.e. spark sweeps) to allow reproduction for
any combination of global input variables. Taking HC emissions as an example, the DOE data collection produced N
spark sweeps, or, in other words, N local cubic models of the form depicted in Figure 4. This provides N sets of values for
local model coefficients bi, which will be used to formulate the global model. The global model will encompass all global
inputs. The type of model used to describe it will depend on how the shape of the cubic used for the local model changes
over the range of each global input variable. Figure 5 shows the characteristics of the HC response model as a function of
two global inputs: air flow and intake camshaft angle. These are the characteristic that the global model is required to
replicate by fitting an appropriate model to the coefficients of all the local models. Various options for model type are
discussed later in this section. To help explain the approach, a cubic model will be used here, due to its simplicity.
2 5
3 0
3 5
4 0
4 5
5 0
5 5
6 0
6 5
7 0
- 2 5
- 2 0
- 1 5
- 1 0
- 5
0
5
1 0
- 2
- 1
0
1
2
3
4
5
6
7
8
x 1 0
- 3
Air Flow
Effect of Global Inputs on HC Emissions
Intake Cam Advance
HCEmissions
6. 6
Figure 5. Effect of global inputs on engine-out HC emissions.
Considering the two global inputs - air flow and intake camshaft angle - the resulting global model equation for b0 is as
follows.
3
10
2
9
2
8
3
7
3
65
2
4
2
32100
**
*
INCAMcAIRFLOWINCAMcINCAMAIRFLOWcINCAMc
AIRFLOWcINCAMAIRFLOWcINCAMcAIRFLOWcINCAMcAIRFLOWccb
++++
++++++=
where cn are the coefficients of the cubic model. Similar equations can be written for b1, b2, and b3. When combined, the
response characteristics at the local stage can be reproduced for any set of global conditions within the ranges set in the
DOE.
Choosing and fitting the global model requires greater skill compared to the local level. The danger is that the global
model can easily be ‘over-fitted,’ or skewed in an undesirable way, through selection of an overly complex model. The
skewing can happen due to errors within the data set, or perhaps due to discrete changes in the response. For example,
if some of the local sweeps contained data indicating that one or more of the cylinders were misfiring, then the engine-out
HC emissions for those sweeps would be unexpectedly higher than others. This random data inclusion would have to be
removed through proper screening to ensure a good data foundation for the model-building process.
It is often the case that the cubic model is not sufficient to model the bi coefficients, and a more complex model is required
to achieve a more accurate model fit. Analysis of the statistic RMSE and other statistics such as predicted sum of squares
(PRESS RMSE) [7] were used successfully to judge model fit, after the removal of data outliers. Chasing the lowest value
for RMSE may well ‘over fit’ the model, so it is sensible to also consider PRESS RMSE, which checks the model’s ability
to predict a value after removal of a data point. A more complex model, which offers more flexibility and is capable of
modeling non-linear responses, is a radial basis function (RBF) [10]. This was the chosen model applied to the data in
Figure 5, which shows only a slice of the whole model. The complete model had eight input variables in the DOE, and its
shape changes as other inputs change. Other models to describe combustion stability, brake torque, and exhaust energy
flow rate were compiled, using the same principles described above.
CALIBRATION OPTIMIZATION
An optimized calibration refers to the data set that resides in the engine control unit (ECU) in the form of look-up tables or
constants. Combined with control strategy, the optimized calibration determines the actuator positions for a given set of
engine operating conditions. It must be capable not only of controlling the engine to meet or exceed the desired emissions
targets, but also of doing so in a manner that will not compromise other vehicle attributes, such as drivability or noise and
vibration.
This is not a simple task. A common mistake would be to optimize for a single response and ignore all other outputs. For
example, the objective of the optimization might be to seek the lowest possible HC emissions, but frequently, engine
operation at the minimum results in poor combustion stability. Hence it is best to make full use of the model-based
approach study all relevant responses, and introduce constraints during such an optimization.
There is also a certain amount of trial and error in any optimization. Relating modeled engine test bed behavior to engine-
in-vehicle behavior can require several iterations, especially when combustion stability is considered. It may therefore be
necessary to carry out optimizations with different combustion stability constraints. Without a model, derivation of these
optimums would require a re-visit to the engine test bed or ad-hoc revisions based on subjective vehicle behavior.
Optimization can either be performed manually through visual inspection of all response models or automatically through
use of the optimization function available within MATLAB (called fmincon). Experience has shown that the best approach
is to use a combined approach. The response models can be used as an engine test bed simulator. This is a very
powerful tool for the engineer, as engine behavior can be examined by changing one input variable at a time. Hence, a
comprehensive understanding can be built up of how the input variables affect different responses, and an expected
outcome from the optimization may be theorized.
To perform an automated optimization, the knowledge gained from visualizing the model responses manually are used to
select an objective(s). For emissions calibration, the target is to achieve catalyst light-off and minimize engine-out
emissions at the same time. For a given exhaust system architecture, the catalyst light-off is governed primarily by the
amount of heat energy generated by the engine.
7. 7
This can be described by the following simple equation:
TCmsJrgyFlowExhaustEne pΔ=
•
]/[
In most cases, minimization of engine-out HC emissions offers the biggest challenge, as this is often close to the
emissions limit: on the other hand, strategies that minimize HC emissions also aid reduction in other gaseous emissions,
such as carbon monoxide (CO).
To determine the operating requirements of the engine (i.e., speed and torque) the vehicle emissions drive cycle must be
analyzed. The engine is required to produce torque, to power the vehicle at target speed. Apart from basic vehicle design
parameters such as mass, overall gearing, and driveline losses, other factors, such as selected gear and transmission
torque converter behavior, can alter the power requirements for a given cycle. If these are considered to be optimized,
then the engine will operate at a given speed and torque to achieve the target vehicle speed. To simplify this further, the
amount of time spent at each speed and load point can be calculated and a list of speed and load points (mini map
points) that account for > 95 percent of the catalyst heating operation period can be generated.
Taking one of these mini map points as an example, and recalling that engine speed and torque are fixed, to achieve
maximum exhaust energy flow rate, mass flow rate out of the engine and exhaust gas temperature need to be maximized.
This can be achieved through increasing intake manifold pressure (opening of the throttle blade) and retarding spark
timing. Figure 6 shows the response model for exhaust energy flow rate and HC emissions as a function of exhaust mass
flow rate and spark advance. Engine torque is not constant for each surface plot in Figure 6: for clarity, an approximate
constant torque vector has been sketched on each surface plot.
The solution to the optimization problem lies on these surfaces along the vector of constant torque. To find the best
solution along this vector, the main options available for optimization depend upon whether a single or multiple objectives
are used. A single objective would simply find the maximum or minimum of a particular response, whereas two or more
objectives would produce a Pareto trade-off [11].
- 3 0
- 2 0
- 1 0
0
1 0
2 0
3 0
2 5
3 0
3 5
4 0
4 5
5 0
5 5
6 0
6 5
7 0
0
0 . 5
1
1 . 5
2
2 . 5
3
3 . 5
x 1 0
4
Exhaust Energy and HC Emissions fn(mass flow rate, spark advance)
Spark Advance
Mass Flow Rate
ExhaustEnergy
- 3 0
- 2 0
- 1 0
0
1 0
2 0
3 0
2 5
3 0
3 5
4 0
4 5
5 0
5 5
6 0
6 5
7 0
- 0 . 0 0 5
0
0 . 0 0 5
0 . 0 1
0 . 0 1 5
0 . 0 2
0 . 0 2 5
0 . 0 3
0 . 0 3 5
Mass Flow Rate
Hydrocarbons
Approximate constant torque vector
Approximate constant torque vector
Spark Advance
Figure 6. Exhaust energy and HC emissions response models.
Single-Objective Optimization
As the models in Figure 6 would suggest, a single objective optimization is adequate to achieve the goal of obtaining
maximum exhaust gas energy flow rate while minimizing HC emissions. In fact, either one of these responses could have
been the objective for the optimization, as long as the other response was constrained not to exceed a target value. Other
constraints were applied to the optimization, such as:
8. 8
• Combustion stability <= desired target.
• Manifold vacuum => desired target.
• Model boundary constraint; the optimum needs to lie in an area where the model is valid.
These constraints were necessary because without them during the process of minimizing HC emissions, combustion
stability would have deteriorated, and manifold vacuum decreased to levels that offered insufficient brake booster
assistance.
The above optimization was run at different mini map points to generate a preliminary calibration before the process of
map smoothing which is discussed later.
Multi-Objective Optimization
Using more than one objective can be very useful because in the process of optimizing one objective, another response
deteriorates. When maximizing exhaust energy flow rate, it is often the case that engine-out HC emissions start to
increase as the combustion limit is approached. To understand the trade-off between these two responses – in other
words, the best compromise between minimum HC emissions and maximum exhaust energy flow rate – a multi-objective
optimization was used, with similar constraints to those used for the single objective optimization.
The normal boundary intersection (NBI) algorithm is used for a multi-objective optimization. This algorithm considers the
optimum for each individual objective and then calculates spacing between these optima for a desired number of trade-off
points. Each trade-off point is then moved in a direction normal to the line connecting the individual optima. More
information can be found in published literature [11].
Figure 7 shows an example of the NBI algorithm applied to minimize HC emissions and maximize exhaust energy flow
rate. Two Pareto curves are shown where different combustion stability limits were chosen. Accepting higher combustion
instability results in similar HC emissions and higher exhaust energy flow rates.
Derivation of an optimum calibration from a Pareto is more time-consuming, as there is more than one possible solution.
More often than not, the calibration values are similar for adjacent Pareto points, as the engine will be operating close to
an optimum condition. As a rule of thumb, the knee point (usually the middle point on the trade-off) is used as the starting
point for the vehicle calibration. The NBI algorithm was applied to different mini map points to generate a preliminary
calibration before map smoothing.
Pareto: HC Emissions and Exhaust Gas Energy
2 .4
2 .5
2 .6
2 .7
2 .8
2 .9
3 .0
3 .1
3 .2
4 2 4 4 4 6 4 8 5 0 5 2 5 4 5 6
Exhaust Energy
EngineOutHCEmissions
Stability <= 100 % Stability <= 120 %
Pareto extended by
increasing Stability limit
Figure 7. NBI Optimization showing effect of stability limits.
CALIBRATION MAP GENERATION
Regardless of whether a single or multi objective optimization is used, the model-optimized values of the actuator settings
generally require modification for one reason or another. Most frequently, the settings are modified to generate a
reasonably smooth map for the calibration to avoid steep gradients and odd transient behavior. Generating a smooth map
9. 9
can be quite involved and becomes increasingly difficult as the number of input variables increase, due to the number of
resulting maps.
Using the output of an optimization as the starting point a manual trade-off – where all responses can be viewed
simultaneously – was found to be adequate. Small adjustments on one actuator always need to be counteracted by
adjustments on other actuators, so that the optimum is maintained whenever possible and constraints are not exceeded.
VERIFICATION / VALIDATION
The verification part of the process should not be neglected, as it is important to understand how accurate the models are
at predicting engine responses. Any model is a simplification of a complex real life situation, so the process of model
verification needs to be treated subjectively. Hence, during the verification and validation process, test data repeatability
must always be considered when a new data set is compared with a given model. Quantitatively the data may not be the
same but subjectively it may be acceptable.
Calibration Verification On FTP Drive Cycle
0
10
20
30
40
50
60
70
80
90
100
0 25 50 75 100 125 150
Time [s]
NormalisedNMHC[%]
0
10
20
30
40
50
60
70
VehicleSpeed[mph]
Vehicle Speed
HC Baseline before Optimisation
HC After Optimisation
Modal Data
Figure 8 Verification using FTP Drive Cycle
Two methods have been used here for model verification: 1) presentation of new/unseen data to the model and analysis
of statistics such as PRESS RMSE: 2) vehicle emissions tests. The first method is very good for determining whether any
over-fitting has occurred. The second method is very useful for summarizing the outcome of the model-based process, as
this result is a clear statement of whether or not the process has worked. Figure 8 compares cumulative tailpipe NMHC
(non-methane hydrocarbons) emissions for two different engine calibrations, using an FTP75 drive cycle. The only
significant difference between the calibrations were that Test 1 used non-optimized intake and exhaust camshaft settings,
and Test 2 used settings produced by the approach described in this paper. The most significant part of this comparison
is the first 60 seconds. Towards the end of this period, catalyst light-off is achieved, but prior to that, most if not all
pollutants pass directly through the catalyst and are not oxidized/reduced. The optimized calibration has outperformed the
baseline calibration reducing total NMHC by approximately 15 percent.
CONCLUSION
Model-Based-Calibration development, using DOE and statistical modeling, is a key tool in developing high-technology
engines with demanding emissions and driveability constraints. Optimization of systems with large degrees of freedom
using traditional methods is impractical. Key to the success of such methods is careful execution of each subtask
10. 10
involved. Experiments need to be designed using the most appropriate design with the correct number of points, input
variables and ranges. Engine test data need to be continuously scrutinized to ensure development of accurate models.
Calibration generation from optimization results must consider multiple output variables and constraints, and trade-offs
need to be carefully selected. A combination of visual inspection, manual manipulation, and calibration map smoothing
will result in the best practical and easy-to-implement solution in the ECU. Experience is a key factor in successful
application of this process.
ACKNOWLEDGMENTS
The author would like to acknowledge the assistance of the following people at General Motors Holden Ltd. Greg Horn,
Martin Jansz, Richard Hurley, Kevin Yardley, Joshua Wood, and Julian Banfield.
Thanks also to David Sampson at The MathWorks for his continuing support of Model-Based-Calibration Toolbox.
REFERENCES
1. U. Kramer and P. Phillips. “Phasing Strategy for an Engine with Twin Variable Cam Timing.” SAE International 2002.
SAE Paper No. 2002-01-1101.
2. T. Kidokoro et al. “Development of PZEV Exhaust Emission Control System.” SAE International 2003. SAE Paper No.
2003-01-0817.
3. J.J. Batteh et al. “Transient Fuel Modeling and Control for Cold Start Intake Cam Phasing.” SAE International 2006.
SAE Paper No. 2006-01-1049.
4. K. Morita et al. “Emissions Reduction of a Stoichiometric Gasoline Direct Injection Engine.” SAE International 2005.
SAE Paper No. 2005-01-3687.
5. J. Seabrook. Practical Implementation of Design of Experiments in Engine Development. Statistics for Engine
Optimization, Edited by Edwards S.P., Grove D.M. & Wynn H.P. Professional Engineering Publishing 2000. ISBN-1-
86058-201-X
6. Douglas C. Montgomery, Design and Analysis of Experiments, 5th
Edition, Wiley 2000. ISBN 0-471-31649-0
7. Model-Based Calibration Toolbox User’s Guide. The MathWorks, Inc., 2004
8. C. R. Tindle, “Comparison of Engine-Out Emissions at Different Operating Temperatures.” Internal GM publication,
August 2005.
9. S. Flint and P. Cawsey, “Use of Experimental Design and Two-Stage Modeling in Calibration Generation for Variable
Camshaft Timing Engines.” 2nd
Conference Design of Experiments in Engine Development. Expert Verlag June 2003
ISBN 3-8169-2271-6.
10. T. Morton and S. Knott, “Radial Basis Functions for Engine Mapping.” Institution of Mechanical Engineers, IMechE
Paper No. C606/022/2002 from International Conference on Statistics and Analytical Methods in Automotive
Engineering. Professional Engineering Publishing 2002. ISBN 1-86058-387-3.
11. O. Roudenko et al., “Application of a Pareto-based Evolutionary Algorithm to Fuel Injection Optimization.” Institution of
Mechanical Engineers, IMechE Paper No. C606/003/2002 from International Conference on Statistics and Analytical
Methods in Automotive Engineering. Professional Engineering Publishing 2002. ISBN 1-86058-387-3.
12. Das Indraneel, The Normal-Boundary Intersection Home Page;
http://www.caam.rice.edu/%7Eindra/NBIhomepage.html
CONTACT
Dr. Clive Tindle
Powertrain Engineering - MP302
Holden Proving Ground
Bass Highway
Lang Lang 3984
VICTORIA - AUSTRALIA
Email: clive.tindle@gm.com
Tel: +61 (0) 3 594 58382