An automatic optimization method for calibrating diesel engine injection strategies

Abstract

The present invention is applicable to the field of diesel engine injection strategy calibration, and provides an automatic optimization method for diesel engine injection strategy calibration. The method includes: collecting diesel engine operation and emission data through a virtual diesel engine platform or numerical simulation software; processing and analyzing the obtained data, and selecting influencing factor data and optimization target data; using the NSGA-III multi-objective genetic algorithm to generate non-dominated solutions; generating an automatic optimization scheme for diesel engine injection strategy calibration. The present invention can consider the influence of operating conditions and injection parameters on diesel engine emissions in multiple aspects, and can improve the economic performance and emission performance of diesel engines.

Description

Technical Field

[0001] The present invention relates to the field of diesel engine injection strategy calibration, and specifically relates to an automatic optimization method for diesel engine injection strategy calibration based on the NSGA-III multi-objective genetic algorithm. Background Art

[0002] Diesel engines are a common type of internal combustion engine and are widely used in fields such as automobiles, ships, and generator sets. During the operation of a diesel engine, fuel injection control has an important impact on combustion efficiency and emission performance. Optimizing the injection control strategy of a diesel engine is the key to improving its performance and economy.

[0003] With the development of optimization algorithms, optimization algorithm technology has been widely applied to the field of diesel engine performance evaluation. In order to meet regulatory standards, diesel engines have adopted various new optimization technologies, and the control systems of diesel engines have become increasingly complex. Their control parameters form a huge system, and the calibration methods of diesel engines also need to be continuously innovated and optimized. Optimizing the calibration of diesel engine control parameters is a key technology that affects the actual fuel consumption and emission performance of diesel engines. However, in the face of a complex control parameter system, traditional manual calibration methods require a large amount of time and economic costs, and the optimization results are often not the optimal parameters.

[0004] In summary, in order to further improve the economy of diesel engines and control diesel engine exhaust emissions, an automatic optimization method for diesel engine injection strategy calibration based on the NSGA-III multi-objective genetic algorithm is proposed. Summary of the Invention

[0005] In order to solve the above problems, the present invention proposes an automatic optimization method for diesel engine injection strategy calibration, which is characterized by including the following steps:

[0006] S100: Collect diesel engine operation and emission data through a virtual diesel engine platform or numerical simulation software;

[0007] S200: Process and analyze the obtained data, and select influencing factor data and optimization target data;

[0008] S300: Use the NSGA-III multi-objective genetic algorithm to generate non-dominated solutions;

[0009] S400: Generate an automatic optimization scheme for diesel engine injection strategy calibration.

[0010] Furthermore, in the process of S200, the processing and analysis of the obtained data, and the selection of influencing factor data and required optimization target data include:

[0011] S201: Data preprocessing;

[0012] S202: Select the operating data points evenly distributed within a certain operating condition range of the diesel engine and the corresponding emission data;

[0013] S203: Select the influencing factor data and the required optimization target data;

[0014] During the process of S201, the data preprocessing includes: filtering the S100 data using an adaptive filter to obtain the denoised operating and emission data;

[0015] During the process of S202, the selection of the operating data points evenly distributed within a certain operating condition range of the diesel engine and the corresponding emission data includes: selecting the test data points evenly distributed within the operating condition range of 1400 r / min to 3000 r / min and 30% to 100% load of the diesel engine after the S201 preprocessing and the corresponding emission data;

[0016] During the process of S203, the selection of the influencing factor data and the required optimization target data includes: selecting five influencing factors, namely speed, torque, main injection timing, pre-injection timing, and pre-injection fuel quantity, to optimize the emissions of NOx and CO of the diesel engine and the brake specific fuel consumption (BSFC) of the diesel engine.

[0017] Furthermore, during the process of S300, the use of the NSGA-III multi-objective genetic algorithm to generate non-dominated solutions includes:

[0018] S301: Create an optimization calibration objective function for the diesel engine;

[0019] S302: Analyze the conflict of the diesel engine optimization objectives;

[0020] S303: Analyze the dominance relationship of the optimal solutions for the multi-objective optimization of the diesel engine;

[0021] S304: Generate a population for the multi-objective optimization of the diesel engine;

[0022] During the process of S301, the creation of the optimization calibration objective function for the diesel engine includes: O is the optimization objective function; O BSFC 、O NOx 、O CO are respectively the optimization objective functions for BSFC, NOx emissions, and CO emissions; X T is the decision vector for multi-objective optimization, which is composed of speed, torque, main injection timing, pre-injection timing, and pre-injection fuel quantity; gj is the inequality constraint, that is, the boundary constraint of the diesel engine; each optimization objective sets a weight ω BSFC ,ω NOx ,ω CO；

[0023] Decision vector for multi-objective optimization is obtained based on the following formula:

[0024]

[0025] The optimization objective function is is obtained based on the following formula:

[0026]

[0027] The inequality constraints are obtained based on the following formula:

[0028]

[0029] The multi-objective optimization objective function of the diesel engine is:

[0030]

[0031] In the process of S302, the contradiction analysis of the diesel engine optimization objective includes: according to the actual calibration requirements, constraints can be given in the genetic algorithm, and through setting weights, the optimization result can be made to focus more on a certain calibration target emission, so as to achieve flexible calibration under different working conditions and calibration targets;

[0032] In the process of S303, the analysis of the dominance relationship of the optimal solution of the multi-objective optimization of the diesel engine includes: introducing the concept of dominance to evaluate the quality of each solution, comparing each solution with other solutions in the population in turn, determining whether the solution is a dominant front solution, realizing the effective solution of the multi-objective optimization under constraints, outputting the corresponding Pareto solution, and all the optimal solutions form the Pareto front;

[0033] In the process of S304, the generation of the multi-objective optimization population of the diesel engine includes: establishing the connection between the Pareto solution in S303 and the reference point, generating non-dominated solutions through the reference point, and the reference point is the reference for the uniform distribution of the population individuals in the solution space; generating genetic individuals uniformly in the optimization objective space to ensure the integrity of the constraint region; the reference point is defined by a structured method and adjusted according to the diesel engine constraint conditions, and through the boundary crossing construction weight method, the reference point is placed on the standardized decision boundary;

[0034] Furthermore, in the process of S400, the generation of the automatic optimization scheme for the diesel engine fuel injection strategy calibration includes:

[0035] S401: Solve all solutions that meet the optimization objectives according to the Pareto optimization method;

[0036] S402: Select the solutions in the appropriate Pareto solution set under different working conditions as the calibration parameters;

[0037] The beneficial effects of adopting the present invention:

[0038] Currently, relying on traditional manual calibration methods, the injection strategies of diesel engines are usually not fully optimized. To solve this problem, the present invention proposes an automatic optimization method for calibrating the injection strategy of a diesel engine. This method collects the operation and emission data of the diesel engine through a virtual diesel engine platform or numerical simulation software, extracts the target operating condition data and emission data of the diesel engine, and then uses the NSGA-III multi-objective genetic algorithm to optimize the injection calibration strategy scheme of the diesel engine, establishing a calibration method that can multi-objectively and intelligently optimize the injection strategy of the diesel engine, achieving multi-objective optimization calibration of the diesel engine, realizing the comprehensive improvement of optimization accuracy and test cost, solving the comprehensive optimization problem of complex control parameters and multi-objective characteristics requirements of the diesel engine, and solving the problems of low efficiency, high cost, subjectivity, and inability to achieve the global optimal solution in manual off-line calibration, thus meeting the requirements of the development of diesel engine technology. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] Figure 1 is a schematic flow chart of an automatic optimization method for calibrating the injection strategy of a diesel engine according to the present invention.

[0040] Figure 2 is a schematic flow chart of the data processing and analysis of the diesel engine according to the present invention.

[0041] Figure 3 is a schematic flow chart of the NSGA-III multi-objective genetic algorithm according to the present invention.

[0042] Figure 4 is a schematic diagram of the generation of an automatic optimization scheme for calibrating the injection strategy of a diesel engine according to the present invention.

[0043] Figure 5 is a schematic diagram of the generation of a multi-objective optimization population of a diesel engine according to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0044] In order to make the objectives, technical solutions and advantages of the present invention clearer and more understandable, the present invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0045] Refer to Figure 1 , the present invention provides an automatic optimization method for calibrating the injection strategy of a diesel engine, and the steps are as follows:

[0046] S100: Collect the operation and emission data of the diesel engine through a virtual diesel engine platform or numerical simulation software;

[0047] S200: Process and analyze the obtained data, and select the influencing factor data and optimization target data;

[0048] S300: Use the NSGA-III multi-objective genetic algorithm to generate non-dominated solutions;

[0049] S400: Generate an automatic optimization scheme for diesel engine injection strategy calibration;

[0050] Further, refer to Figure 2 , in the process of S200, process and analyze the obtained data, and select the influencing factor data and optimization target data including:

[0051] S201: Data preprocessing;

[0052] S202: Select the operating data points evenly distributed within a certain operating condition range of the diesel engine and the corresponding emission data;

[0053] S203: Select the influencing factor data and the required optimization target data;

[0054] In the process of S201, the data preprocessing includes: using an adaptive filter to filter the data of S100 to obtain the denoised operating and emission data;

[0055] In the process of S202, selecting the operating data points evenly distributed within a certain operating condition range of the diesel engine and the corresponding emission data includes: selecting the test data points evenly distributed within the operating condition range of 1400 r / min to 3000 r / min and 30% to 100% load of the diesel engine after S201 preprocessing and the corresponding emission data;

[0056] In the process of S203, selecting the influencing factor data and the required optimization target data includes: selecting five influencing factors of speed, torque, main injection timing, pre-injection timing, and pre-injection fuel quantity to optimize the emissions of NOx and CO of the diesel engine and the brake specific fuel consumption (BSFC) of the diesel engine.

[0057] Even further, refer to Figure 3 , in the process of S300, using the NSGA-III multi-objective genetic algorithm to generate non-dominated solutions includes:

[0058] S301: Create a diesel engine optimization calibration objective function;

[0059] S302: Analyze the conflict of diesel engine optimization objectives;

[0060] S303: Analyze the dominance relationship of the optimal solutions of diesel engine multi-objective optimization;

[0061] S304: Generate a population for diesel engine multi-objective optimization;

[0062] In the process of S301, creating a diesel engine optimization calibration objective function includes: O is the optimization objective function; O BSFC 、ONOx and O CO are the optimization objective functions for BSFC, NOx emissions, and CO emissions respectively; X T is the decision vector for multi-objective optimization, which consists of engine speed, torque, main injection timing, pilot injection timing, and pilot injection quantity; gj is the inequality constraint, i.e., the boundary constraint of the diesel engine; the weights ω BSFC , ω NOx , ω CO；

[0063] Decision vector for multi-objective optimization is obtained based on the following formula:

[0064]

[0065] The optimization objective function is obtained based on the following formula:

[0066]

[0067] The inequality constraint is obtained based on the following formula:

[0068]

[0069] The multi-objective optimization objective function of the diesel engine is:

[0070]

[0071] Furthermore, since the optimization objectives are all to minimize values, it can be directly transformed into a standard minimization problem:

[0072]

[0073] br>In the process of S302, the analysis of the conflict of the diesel engine optimization objectives includes: according to the actual calibration requirements, constraints can be given in the genetic algorithm, and by setting weights, the optimization results can be made to focus more on a certain calibration target emission to achieve flexible calibration under different working conditions and calibration targets;

[0074] In the process of S303, the analysis of the dominance relationship of the optimal solutions for multi-objective optimization of the diesel engine includes: introducing the concept of dominance to evaluate the quality of each solution, comparing each solution with other solutions in the population in turn to determine whether the solution is a dominant front solution, achieving an effective solution for multi-objective optimization under constraints, outputting the corresponding Pareto solutions, and all the optimal solutions form the Pareto front;

[0075] During the process of S304, the generation of the multi-objective optimization population for the diesel engine includes: establishing the connection between the Pareto solutions in S303 and the reference points, generating non-dominated solutions through the reference points, where the reference points are the references for the uniform distribution of population individuals in the solution space; uniformly generating genetic individuals in the optimization objective space to ensure the integrity of the constraint region; the reference points are defined by a structured method and adjusted according to the constraint conditions of the diesel engine, and a weight method is constructed through boundary crossover to place the reference points on the standardized decision boundary;

[0076] See Figure 5 , in the optimization objective space of the diesel engine, construct a standard simplex plane P with a dimension of 2 2 ; P 2 has the same slope for the BSFC, NOx, and CO target axes. Divide into p regions along the target axis direction, the number of objectives is M, and calculate the number of reference points: ; The decision boundary P 2 has the same slope relative to the BSFC, NOx, and CO axes, and is divided into p parts along each optimization objective direction; for a 3D optimization problem, intermediate points are generated by adding a second-layer decision boundary to uniformly generate samples.

[0077] Furthermore, see Figure 4 , during the process of S400, the generation of the automatic optimization scheme for the diesel engine fuel injection strategy calibration includes:

[0078] S401: Solve all solutions that meet the optimization objectives according to the Pareto optimization method;

[0079] S402: Select the solutions in the appropriate Pareto solution sets under different working conditions as the calibration parameters;

[0080] The present invention proposes an automatic optimization method for diesel engine fuel injection strategy calibration. This method collects the operation and emission data of the diesel engine through a virtual diesel engine platform or numerical simulation software, extracts the target operation condition data and emission data of the diesel engine, and then uses the NSGA-III multi-objective genetic algorithm to optimize the diesel engine fuel injection calibration strategy scheme, establishing a calibration method that can multi-objectively and intelligently optimize the diesel engine fuel injection strategy, achieving multi-objective optimization calibration of the diesel engine, realizing the comprehensive improvement of optimization accuracy and test cost, solving the comprehensive optimization problem of complex control parameters and multi-objective characteristics requirements of the diesel engine, and solving the problems of low efficiency, high cost, subjectivity, and inability to achieve the global optimal solution existing in manual off-line calibration, thus meeting the requirements of the development of diesel engine technology.

[0081] The above are only the preferred embodiments of the present invention and are not used to limit this method. Any modifications, equivalent replacements, and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. An automatic optimization method for calibrating a diesel engine injection strategy, characterized in that, Including: S100: Collect diesel engine operation and emission data through a virtual diesel engine platform or numerical simulation software; S200: Process and analyze the obtained data, and select influencing factor data and optimization target data; S300: Use the NSGA-III multi-objective genetic algorithm to generate non-dominated solutions; S400: Generate an automatic optimization scheme for calibrating the diesel engine injection strategy; In the process of S300, the use of the NSGA-III multi-objective genetic algorithm to generate non-dominated solutions includes: S301: Create an objective function for optimizing the calibration of the diesel engine; S302: Analyze the conflict of the diesel engine optimization objectives; S303: Analyze the dominance relationship of the optimal solutions for multi-objective optimization of the diesel engine; S304: Generate a population for multi-objective optimization of the diesel engine; In the process of S301, the creation of the diesel engine optimization calibration objective function includes: O is the optimization objective function; O BSFC , O NOx , O CO are the optimization objective functions of BSFC, NOx emissions, and CO emissions respectively; X T is the decision vector for multi-objective optimization, which consists of engine speed, torque, main injection timing, pre-injection timing, and pre-injection fuel quantity; gj is the inequality constraint, that is, the boundary constraint of the diesel engine; each optimization objective sets a weight ω BSFC , ω NOx , ω CO ; The decision vector X for multi-objective optimization T is obtained based on the following formula: X T = [X Speed , X Torque , X Main , X Pilot , X PilotMass ​ The optimization objective function O(X) is obtained based on the following formula: O(X T ) = [O BSFC (X T ), O NOx (X T ), O CO (X T )] T The inequality constraint is obtained based on the following formula: s.t.g j (X T ) ≥ 0, j = i, n, …, J The objective function for multi-objective optimization of the diesel engine is: O(X T ) = ω BSFC O BSFC (X T ) + ω NOx O NOx (X T ) + ω CO O CO (X T ) In the process of S302, the analysis of the conflict of the diesel engine optimization objectives includes: According to the actual calibration requirements, constraints can be given in the genetic algorithm, and the optimization result can be made to focus more on a certain calibration target emission through weight setting, so as to achieve flexible calibration under different working conditions and calibration targets; In the process of S303, the analysis of the dominance relationship of the optimal solutions for multi-objective optimization of the diesel engine includes: Introduce the concept of dominance to evaluate the quality of each solution, compare each solution with other solutions in the population in turn, determine whether the solution is a dominance front solution, realize the effective solution of multi-objective optimization under constraints, output the corresponding Pareto solutions, and all the optimal solutions constitute the Pareto front; In the process of S304, the generation of a population for multi-objective optimization of the diesel engine includes: Establish the connection between the Pareto solutions in S303 and the reference points, generate non-dominated solutions through the reference points, and the reference points are references for the uniform distribution of population individuals in the solution space; Generate genetic individuals uniformly in the optimization objective space to ensure the integrity of the constraint region; The reference points are defined by a structured method and adjusted according to the diesel engine constraint conditions, and the reference points are placed on the standardized decision boundary through the boundary crossing construction weight method.

2. The automatic optimization method for calibrating a diesel engine fuel injection strategy according to claim 1, characterized in that In the process of S200, the processing and analysis of the obtained data, and the selection of influencing factor data and required optimization target data include: S201: Data preprocessing; S202: Select the operation data points and corresponding emission data evenly distributed within a certain working condition range of the diesel engine; S203: Select influencing factor data and required optimization target data; In the process of S201, the data preprocessing includes: Use an adaptive filter to filter the data in S100 to obtain the denoised operation and emission data; In the process of S202, the selection of the operation data points and corresponding emission data evenly distributed within a certain working condition range of the diesel engine includes: Select the test data points and corresponding emission data evenly distributed within the working condition range of 1400 r / min to 3000 r / min and 30% to 100% load of the diesel engine after the preprocessing in S201; In the process of S203, the selection of influencing factor data and required optimization target data includes: selecting five influencing factors, namely rotational speed, torque, main injection timing, pilot injection timing, and pilot injection fuel quantity, to optimize the emissions of NOx and CO of the diesel engine and the brake specific fuel consumption (BSFC) of the diesel engine.

3. The automatic optimization method for calibrating a diesel engine fuel injection strategy according to claim 1, characterized in that, In the process of S400, the generation of the automatic optimization scheme for the calibration of the diesel engine injection strategy includes: S401: Solve all solutions that meet the optimization target according to the Pareto optimization method; S402: Select the solutions in the appropriate Pareto solution set under different working conditions as the calibration parameters.

Images