An engine model parameter dynamic calibration optimization method

Abstract

The application discloses an engine model parameter dynamic calibration optimization method, which comprises the following steps: obtaining a target parameter to be optimized, and determining a corresponding dynamic optimization problem according to the target parameter to be optimized and a working scene; calling an algorithm portfolio according to the dynamic optimization problem, and updating optimal solutions of each algorithm in the algorithm portfolio through optimization information interaction and sharing strategies among algorithms; determining optimization promotion corresponding to each algorithm according to the optimal solutions obtained by each algorithm, and distributing corresponding calculation resources to each algorithm according to the determined optimization promotion degree; optimizing the target parameter to be optimized through the distributed algorithm portfolio, and calibrating the optimized target parameter as a parameter meeting a performance index according to an actual output value and an expected output value of an aero-engine simulation model. The application applies the algorithm portfolio to the dynamic calibration problem of the engine parameter, and improves the robustness and stability of the parameter calibration.

Description

Technical Field

[0001] This invention relates to the field of engine model calibration technology, and in particular to a method for dynamic calibration and optimization of engine model parameters. Background Technology

[0002] Intelligent optimization methods are widely used in the calibration of aero-engine models. Because the data in the working scenario of aero-engine model calibration is dynamically changing, the calibration of engine model parameters becomes a dynamic optimization problem. Typically, the dynamic change of the objective function in a dynamic optimization problem means that algorithms that performed well before the dynamic change show significantly reduced performance after the change.

[0003] Therefore, in the process of dynamic calibration of engine model parameters, the use of a single algorithm may be affected by dynamic changes, and the robustness of parameter calibration is difficult to guarantee.

[0004] Therefore, existing technologies still need improvement. Summary of the Invention

[0005] The technical problem to be solved by the present invention is to provide a dynamic calibration and optimization method for engine model parameters, in order to address the shortcomings of the existing technology and solve the technical problem of low robustness in the dynamic calibration process of existing engine model parameters.

[0006] The technical solution adopted by this invention to solve the technical problem is as follows:

[0007] In a first aspect, the present invention provides a method for dynamic calibration and optimization of engine model parameters, comprising:

[0008] Obtain the target parameters to be optimized, and determine the corresponding dynamic optimization problem based on the target parameters and the working scenario;

[0009] The algorithm portfolio is invoked according to the dynamic optimization problem, and the optimal solutions of each algorithm in the algorithm portfolio are updated through optimization information interaction and sharing strategies between algorithms; wherein, the optimization information interaction and sharing strategies between algorithms are to set up an archive set according to the optimization population corresponding to each algorithm, and use the archive set to share and update the optimal solutions of each algorithm;

[0010] The optimization improvement of each algorithm is determined based on the optimal solution obtained by each algorithm, and the corresponding computing resources are allocated to each algorithm according to the determined optimization improvement.

[0011] The target parameters to be optimized are optimized by the allocated algorithm portfolio, and the optimized target parameters are calibrated as parameters that meet the performance indicators based on the actual output value and expected output value of the aero-engine simulation model.

[0012] In one implementation, obtaining the target parameter to be optimized and determining the corresponding dynamic optimization problem based on the target parameter and the working scenario includes:

[0013] Obtain several aero-engine model parameters to be adjusted, and then obtain the target parameters to be optimized;

[0014] Determine the operating scenario of the aero-engine model; wherein the operating scenario is any one or a combination of the following: cruise scenario, takeoff scenario, and landing scenario;

[0015] Based on the target parameters to be optimized and the working scenario, a model is created to obtain the corresponding dynamic optimization problem.

[0016] In one implementation, the step of invoking an algorithm portfolio based on the dynamic optimization problem and updating the optimal solutions of each algorithm in the algorithm portfolio through optimization information interaction and sharing strategies between algorithms includes:

[0017] Multiple different algorithms are invoked based on the dynamic optimization problem, and an algorithm portfolio is constructed based on the invoked algorithms.

[0018] Determine the optimization population corresponding to each algorithm and set the archive set corresponding to each algorithm;

[0019] The optimal solution found by each algorithm in each iteration is stored in the corresponding archive set to update the optimal solution of each algorithm.

[0020] In one implementation, storing the optimal solution found by each algorithm in each iteration into the corresponding archive set includes:

[0021] Determine the worst solution in the optimization population corresponding to each algorithm, and replace the worst solution with the best solution in the corresponding archive set.

[0022] In one implementation, the step of determining the optimization improvement corresponding to each algorithm based on the optimal solution obtained by each algorithm, and allocating corresponding computing resources to each algorithm according to the determined optimization improvement, includes:

[0023] Determine the lift of the optimal solution obtained in the current iteration of each algorithm compared to the optimal solution obtained in the previous iteration;

[0024] Computational resources are allocated to each algorithm based on the confidence upper bound algorithm and the lift.

[0025] In one implementation, allocating corresponding computing resources to each algorithm based on the confidence upper bound algorithm and the lift includes:

[0026] The reward value of the corresponding algorithm in the current iteration is determined based on the lift.

[0027] Substitute the reward value into the confidence upper bound algorithm and assign the corresponding algorithm a confidence value in the current iteration;

[0028] The reliability values ​​of each algorithm are updated according to the iteration order, and the corresponding computing resources are allocated to each algorithm based on the updated reliability values.

[0029] In one implementation, the confidence upper bound algorithm is as follows:

[0030]

[0031]

[0032] Where s = T j (n) represents the total number of times algorithm j is selected in the first n iterations;

[0033] Let be the average reward value of the j-th algorithm after n iterations;

[0034] μ j,t Let be the reward of algorithm j when it is selected t times.

[0035] In one implementation, the optimization of the target parameters to be optimized using the allocated algorithmic portfolio, and the calibration of the optimized target parameters as parameters that meet performance indicators based on the actual and expected output values ​​of the aero-engine simulation model, includes:

[0036] The target parameters to be optimized are obtained by optimizing the allocated algorithmic portfolio.

[0037] The optimized target parameters are input into the aero-engine simulation model to obtain the corresponding actual output values;

[0038] Calculate the difference between the actual output value and the expected output value, and calibrate the optimized target parameters as parameters that meet the performance indicators of the working scenario according to the optimization objective function.

[0039] In a second aspect, the present invention also provides a terminal, comprising: a processor and a memory, wherein the memory stores an engine model parameter dynamic calibration and optimization program, and the engine model parameter dynamic calibration and optimization program, when executed by the processor, is used to implement the engine model parameter dynamic calibration and optimization method as described in the first aspect.

[0040] Thirdly, the present invention also provides a storage medium, which is a computer-readable storage medium, storing an engine model parameter dynamic calibration and optimization program, which, when executed by a processor, is used to implement the engine model parameter dynamic calibration and optimization method as described in the first aspect.

[0041] The present invention, by employing the above technical solution, has the following effects:

[0042] This invention obtains the target parameters to be optimized and determines the corresponding dynamic optimization problem based on the target parameters and the working scenario. It can invoke an algorithm portfolio according to the dynamic optimization problem, and update the optimal solutions of each algorithm in the portfolio through optimization information interaction and sharing strategies. Furthermore, it can determine the optimization improvement corresponding to each algorithm based on the optimal solutions obtained by each algorithm, and allocate corresponding computing resources to each algorithm according to the determined optimization improvement. Finally, it optimizes the target parameters to be optimized through the allocated algorithm portfolio, and calibrates the optimized target parameters to meet performance indicators based on the actual and expected output values ​​of the aero-engine simulation model. This invention proposes applying the concept of an algorithm portfolio to the dynamic calibration problem of engine parameters, proposing an algorithm portfolio framework that satisfies the allocation of computing resources in the dynamic calibration of engine model parameters. This ensures that the optimal algorithm for the engine model parameter calibration problem in different working scenarios can be allocated the maximum computing resources, thereby improving the optimization efficiency of the dynamic calibration problem at each stage and enhancing the robustness and stability of parameter calibration. Attached Figure Description

[0043] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the structures shown in these drawings without creative effort.

[0044] Figure 1 This is a flowchart of a dynamic calibration and optimization method for engine model parameters in one implementation of the present invention.

[0045] Figure 2 This is a schematic diagram of parameter adjustment of an aero-engine simulation model in one implementation of the present invention.

[0046] Figure 3 This is an iterative diagram of the intelligent optimization algorithm in one implementation of the present invention.

[0047] Figure 4 This is a schematic diagram illustrating the optimal solution update of the population optimization in one implementation of the present invention.

[0048] Figure 5 This is a functional schematic diagram of the terminal in one implementation of the present invention.

[0049] The realization of the objective, functional features and advantages of the present invention will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation

[0050] To make the objectives, technical solutions, and advantages of this invention clearer and more explicit, the 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 merely illustrative of the invention and are not intended to limit the invention.

[0051] Exemplary methods

[0052] Aero-engine model calibration is a crucial issue in aero-engine design. The calibration of engine model parameters under different operating scenarios can be modeled as a dynamic optimization problem. Existing methods for dynamic optimization problems almost universally employ a single algorithm supplemented by dynamic strategies. This means using the same algorithm before and after the problem's dynamic changes, while employing strategies such as increasing diversity to improve the efficiency of solving dynamic problems. However, dynamic changes cause changes in the objective function. Algorithms that were effective before the dynamic changes may become significantly less efficient afterward. In this situation, a single intelligent algorithm struggles to meet the demands of changing objective functions due to variations in the given operating scenarios in engine parameter calibration.

[0053] To address the issue that a single algorithm can potentially lead to a decrease in optimization efficiency in the dynamic calibration of engine parameters, this embodiment applies the concept of an algorithm portfolio to the dynamic calibration of engine parameters, proposing an algorithm portfolio framework that satisfies the allocation of computational resources in the dynamic calibration of engine model parameters. This ensures that the optimal algorithm for the engine model parameter calibration problem under different working scenarios can be allocated the maximum amount of computational resources, thereby improving the optimization efficiency of the dynamic calibration problem at each stage and enhancing the robustness and stability of parameter calibration.

[0054] like Figure 1 As shown, this embodiment of the invention provides a method for dynamic calibration and optimization of engine model parameters, including the following steps:

[0055] Step S100: Obtain the target parameters to be optimized, and determine the corresponding dynamic optimization problem based on the target parameters to be optimized and the working scenario.

[0056] In this embodiment, the dynamic calibration and optimization method for engine model parameters is applied in a terminal, which includes, but is not limited to, devices such as computers and mobile terminals.

[0057] Aero-engine model calibration is a crucial issue in aero-engine design. The primary objective is to adjust a set of target parameters to ensure the engine meets performance targets across various scenarios within its operational envelope, thereby achieving optimal output.

[0058] In this embodiment, among a series of aero-engine operating scenario data (i.e., the input target parameters and the expected values ​​of the output values), it is necessary to use an optimization algorithm to calibrate the input target parameters and finally obtain a set of adjusted parameters. The adjusted parameters can make the simulation output of the engine model in a given operating scenario consistent with the actual experimental output data (i.e., the expected values ​​of the output values) of these operating scenarios, that is, to minimize the error between the simulation output and the actual experimental output data.

[0059] Specifically, in one implementation of this embodiment, step S100 includes the following steps:

[0060] Step S101: Obtain several aero-engine model parameters to be adjusted, and obtain the target parameters to be optimized;

[0061] Step S102: Determine the working scenario of the aero-engine model;

[0062] Step S103: Model the target parameters to be optimized and the working scenario to obtain the corresponding dynamic optimization problem.

[0063] like Figure 2 As shown, during the calibration of the aero-engine model, the set engine operating scenarios are any one or a combination of the following: cruise mode, takeoff mode, and landing mode. By selecting any engine operating scenario, several aero-engine model parameters to be adjusted and several fixed, non-adjustable aero-engine model parameters are obtained and input into the aero-engine simulation model. Simulation tests are then performed based on the input engine parameters, and the input adjustable parameters are calibrated based on the measured output values.

[0064] In one implementation of this embodiment, during the simulation test of the aero-engine model, the parameters that need to be input to the aero-engine simulation model include: 4 fixed and non-adjustable aero-engine model parameters and 28 adjustable aero-engine model parameters (e.g., nozzle thrust coefficient, nozzle flow coefficient, and internal and external bypass heat transfer coefficients, etc.); among which, the 28 adjustable aero-engine model parameters are the target parameters to be optimized.

[0065] For each working scenario, the parameter values ​​(measured output values) obtained from the output measurement of the aero-engine model include: 15 measured output values ​​(e.g., fan outlet pressure, compressor outlet pressure, etc.); among which, the 15 measured output values ​​can be used as engine performance indicators for each working scenario.

[0066] In this embodiment, by calibrating the input adjustable aero-engine model parameters, the simulation output of the engine model under a given working scenario is made consistent with the actual experimental output data of these working scenarios, that is, the error between the simulation output and the actual experimental output data is minimized. Specifically, after inputting a series of adjustable aero-engine model parameters, the corresponding engine output measurement values ​​are obtained; then, the objective function for optimization is determined based on the engine output measurement values, and the input adjustable parameters are optimized through an intelligent optimization algorithm.

[0067] It's worth noting that in actual engine parameter calibration, data for different operating scenarios are obtained through experiments. However, the order of these experiments results in a varying number of operating scenarios being processed at different points in time during the calibration process. This means that the optimization objective faced by the intelligent optimization algorithm changes as the given operating scenario evolves during engine parameter calibration. Therefore, engine model parameter calibration can be modeled as a dynamic optimization problem, meaning it can be modeled based on adjustable input parameters and operating scenarios to obtain the corresponding dynamic optimization problem.

[0068] like Figure 1 As shown, in one implementation of this invention, the dynamic calibration and optimization method for engine model parameters further includes the following steps:

[0069] Step S200: Invoke the algorithm portfolio according to the dynamic optimization problem, and update the optimal solution of each algorithm in the algorithm portfolio through optimization information interaction and sharing strategy between algorithms.

[0070] In this embodiment, the concept of algorithm portfolio can be applied to the dynamic calibration problem of engine parameters, and an algorithm portfolio framework is proposed to meet the computational resource allocation in the dynamic calibration of engine model parameters. This allows the optimal algorithm for the engine model parameter calibration problem in different working scenarios to be allocated the maximum computational resources, thereby improving the optimization efficiency of the dynamic calibration problem at each stage.

[0071] Specifically, in one implementation of this embodiment, step S200 includes the following steps:

[0072] Step S201: Invoke multiple different algorithms according to the dynamic optimization problem, and construct an algorithm portfolio based on the invoked algorithms;

[0073] Step S202: Determine the optimization population corresponding to each algorithm and set the archive set corresponding to each algorithm;

[0074] Step S203: Store the optimal solution found by each algorithm in each iteration into the corresponding archive set to update the optimal solution of each algorithm;

[0075] Step S204: Determine the worst solution in the optimization population corresponding to each algorithm, and replace the worst solution with the best solution in the corresponding archive set.

[0076] In this embodiment, the allocation of computational resources in the dynamic calibration of engine model parameters is mainly achieved through an algorithm portfolio framework. The algorithm portfolio in the framework is a parallel algorithm execution strategy that maximizes the efficiency of various algorithms by allocating different computational resources (e.g., computation time) to different algorithms.

[0077] In one implementation of this embodiment, assuming that there are a total of m algorithms in the algorithm pool of the algorithm portfolio framework, the algorithm portfolio mainly includes three strategies: interaction and sharing of optimization information between algorithms, dynamic allocation of computing resources, and dynamic inheritance and utilization of problem knowledge.

[0078] The optimization information interaction and sharing strategy among the algorithms involves setting up an archive set based on the optimization population corresponding to each algorithm. This archive set is used to share and update the optimal solutions of each algorithm. In this strategy of interaction and sharing optimization information among algorithms, due to the differences between intelligent optimization algorithms, independent operation of the algorithms ensures the maximization of their optimization capabilities. Therefore, in the algorithm portfolio framework of this embodiment, the algorithms remain relatively independent, and each algorithm maintains its own optimization population (solution set). During the optimization process of each algorithm, the optimal solution found by each algorithm in each generation is stored in an archive set; that is, the archive set contains the optimal solutions found by all algorithms in each iteration. Then, for each algorithm, the worst solution in its solution set is replaced with the best solution from the archive set.

[0079] like Figure 3As shown, in intelligent optimization algorithms, the optimal solution to the target problem is found through the iteration of an individual population. Each algorithm maintains a population (e.g., particle swarm optimization). A population consists of a group of individuals, and each individual represents a candidate solution to the problem. That is, the individual population of each algorithm includes multiple sets of solutions for solving the target problem. Assuming that during the nth generation of population iteration, the population selects the worst solution through crossover and mutation, replacing it with the optimal solution, thus obtaining a new population. Furthermore, based on the nth generation of population iteration, the (n+1)th generation of population iteration continues until the optimized population is sufficient to make the result of the target problem satisfy the requirements of the objective function, that is, the output value corresponding to the calibrated engine model parameters satisfies the expected value.

[0080] like Figure 4 As shown, in one implementation of this embodiment, in order to achieve the sharing of optimization information among algorithms, during each population iteration, a corresponding archive set is set for each algorithm (for example, population 1 corresponds to archive set 1), and the optimal solution found by each population is stored in its corresponding archive set (for example, the optimal solution found by population 1 is stored in the corresponding archive set 1). In this way, the optimal solution is stored in each archive set. By merging all the archive sets, a total archive set is obtained. In this total archive set, the optimal individual (i.e., the optimal solution of the total archive set) can be determined, and the optimal individual replaces the worst individual in the population corresponding to each algorithm (for example, the worst solution in population 1), thereby achieving the sharing of optimization information among algorithms.

[0081] like Figure 1 As shown, in one implementation of this invention, the dynamic calibration and optimization method for engine model parameters further includes the following steps:

[0082] Step S300: Determine the optimization improvement corresponding to each algorithm based on the optimal solution obtained by each algorithm, and allocate corresponding computing resources to each algorithm according to the determined optimization improvement.

[0083] In this embodiment, due to the differences between algorithms (i.e., the different times required for algorithm computation), more computational resources need to be allocated to the algorithm with relatively higher optimization efficiency. Therefore, this embodiment designs a dynamic allocation of computational resources within the algorithm portfolio framework to address the dynamic changes in the optimization problem. In the dynamic allocation strategy of computational resources, in the initial stage of optimization, due to the lack of knowledge about the algorithms and the relationship between them and the black-box problem, computational resources can be evenly allocated to the m algorithms within the algorithm portfolio framework.

[0084] Specifically, in one implementation of this embodiment, step S300 includes the following steps:

[0085] Step S301: Determine the lift of the optimal solution obtained in the current iteration of each algorithm compared to the optimal solution obtained in the previous iteration;

[0086] Step S302: Allocate corresponding computing resources to each algorithm according to the confidence upper bound algorithm and the lift.

[0087] In this embodiment, in the dynamic allocation strategy of computing resources, during the population optimization phase of each algorithm, each algorithm will find some solutions to the optimization problem. The algorithm portfolio framework will use this obtained information (i.e. the solution set in the optimization population) to allocate limited computing resources to different algorithms.

[0088] Specifically, in this embodiment, the upper confidence bound algorithm (UCB) from the multi-armed slot machine problem can be used for information processing and resource allocation. For each algorithm, an optimal solution can be found in each iteration. The optimal solution found in each generation can be compared with the optimal solution found in the previous generation to obtain the lift between the two generations of optimal solutions (i.e., the lift is obtained based on the difference between the target value of the optimal solution in each generation and the target value of the optimal solution in the previous generation). Then, this lift is set as the reward of the algorithm in this generation. Substituting this reward into the upper confidence bound algorithm, the corresponding confidence can be obtained.

[0089] It is worth mentioning that if an algorithm no longer converges, its lift is 0, which means that the algorithm does not need to be allocated corresponding computing resources.

[0090] Specifically, in one implementation of this embodiment, step S302 includes the following steps:

[0091] Step S302a: Determine the reward value of the corresponding algorithm in the current iteration based on the lift.

[0092] Step S302b: Substitute the reward value into the confidence upper bound algorithm and assign the corresponding algorithm a confidence value in the current iteration;

[0093] Step S302c: Update the confidence value of each algorithm according to the iteration order, and allocate corresponding computing resources to each algorithm according to the updated confidence value.

[0094] In this embodiment, the confidence upper bound algorithm is as follows:

[0095]

[0096]

[0097] Where s = T j(n) represents the total number of times algorithm j is selected in the first n iterations;

[0098] Let be the average reward value of the j-th algorithm after n iterations;

[0099] μ j,t Let be the reward of algorithm j when it is selected t times.

[0100] In this embodiment, by substituting the reward value obtained above into the confidence upper bound algorithm (i.e., the UCB formula), a confidence score can be assigned to each algorithm. The higher the confidence score of each algorithm, the more optimal it is for the current optimization problem, and therefore, more computing resources should be allocated to it. Thus, after each iteration, the confidence score of each optimization algorithm is calculated using the UCB algorithm. For the optimization algorithm with the highest confidence score, the iteration count of that algorithm is increased by one, and the confidence score is also updated, thereby ensuring that algorithms with higher confidence scores receive more computing resources; that is, for algorithms with high confidence scores, the more iterations there are, the higher the confidence score becomes, and the more computing resources are ultimately obtained.

[0101] like Figure 1 As shown, in one implementation of this invention, the dynamic calibration and optimization method for engine model parameters further includes the following steps:

[0102] Step S400: The target parameters to be optimized are optimized by the allocated algorithm portfolio, and the optimized target parameters are calibrated as parameters that meet the performance indicators based on the actual output value and expected output value of the aero-engine simulation model.

[0103] In this embodiment, during the initial optimization, due to a lack of knowledge about the correlation between algorithms and the problem, computing resources are evenly distributed among the algorithms. Then, when the problem dynamically changes (i.e., when new working scenario data is obtained), since the algorithms have accumulated some knowledge about the problem before the dynamic change, this accumulated knowledge can be used to initialize the allocation of computing resources, instead of using an even distribution. That is, the computing resources are inherited through a dynamic problem knowledge inheritance strategy.

[0104] Specifically, in the dynamic problem knowledge inheritance and utilization strategy, when the problem undergoes dynamic changes, the solutions saved in the optimization process are first re-evaluated (i.e., evaluated using an engine model simulator). These saved solutions correspond to their respective algorithms. Based on the merits of these solutions on the new problem, algorithms with potentially better optimization effects on the current problem are determined (i.e., selecting algorithms that match the new problem based on the saved solution set, and selecting the initial allocation of computational resources for the new problem based on the saved solution set). In this embodiment, the difference between the optimal solution for the current problem and the worst solution in the archive set for each algorithm can be used as the algorithm's reward; the optimal and worst solutions for the current problem can be obtained through evaluation using an engine model simulator (i.e., evaluating each solution, finding the optimal and worst solutions, and calculating the difference in the objective functions of the two solutions); after obtaining the algorithm's reward, each algorithm is given an initial confidence level, and then the selection of algorithms begins, thereby obtaining the optimal algorithm that matches the new problem.

[0105] Specifically, in one implementation of this embodiment, step S400 includes the following steps:

[0106] Step S401: Optimize the target parameters to be optimized using the allocated algorithm portfolio to obtain the optimized target parameters;

[0107] Step S402: Input the optimized target parameters into the aero-engine simulation model to obtain the corresponding actual output values;

[0108] Step S403: Calculate the difference between the actual output value and the expected output value, and calibrate the optimized target parameter as a parameter that meets the performance index of the working scenario according to the optimization objective function.

[0109] In this embodiment, after allocating corresponding computing resources to each algorithm through an algorithm portfolio, the target parameters to be optimized can be optimized using the allocated algorithm portfolio to obtain optimized target parameters. Then, the optimized target parameters are input into the aero-engine simulation model, and the actual output value is measured to determine whether the optimized parameters meet the performance indicators of the working scenario. That is, the difference between the actual output value and the expected output value is calculated, and the optimization objective function is used as the criterion. If the difference between the two satisfies the conditions of the optimization objective function, the optimized parameters (i.e., the input adjustable parameters) can be calibrated as parameters that meet the performance indicators of the working scenario.

[0110] This embodiment achieves the following technical effects through the above technical solution:

[0111] This embodiment proposes applying the concept of algorithm portfolio to the dynamic calibration problem of engine parameters, and puts forward an algorithm portfolio framework to meet the computational resource allocation in the dynamic calibration of engine model parameters. This ensures that the optimal algorithm for the engine model parameter calibration problem in different working scenarios can be allocated to the maximum computational resources, thereby improving the optimization efficiency of the dynamic calibration problem at each stage and improving the robustness and stability of parameter calibration.

[0112] Exemplary device

[0113] Based on the above embodiments, the present invention also provides a terminal, comprising: a processor, a memory, an interface, a display screen, and a communication module connected via a system bus; wherein, the processor is used to provide computing and control capabilities; the memory includes a storage medium and internal memory; the storage medium stores an operating system and computer programs; the internal memory provides an environment for the operation of the operating system and computer programs in the storage medium; the interface is used to connect to external devices, such as mobile terminals and computers; the display screen is used to display corresponding information; and the communication module is used to communicate with a cloud server or a mobile terminal.

[0114] When the computer program is executed by the processor, it is used to implement an operation of a dynamic calibration and optimization method for engine model parameters.

[0115] It will be understood by those skilled in the art that Figure 5 The schematic diagram shown is merely a partial structural diagram related to the present invention and does not constitute a limitation on the terminal to which the present invention is applied. A specific terminal may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0116] In one embodiment, a terminal is provided, comprising: a processor and a memory, the memory storing an engine model parameter dynamic calibration and optimization program, which, when executed by the processor, is used to implement the engine model parameter dynamic calibration and optimization method as described above.

[0117] In one embodiment, a storage medium is provided, wherein the storage medium stores an engine model parameter dynamic calibration and optimization program, which, when executed by the processor, is used to implement the engine model parameter dynamic calibration and optimization method as described above.

[0118] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile storage medium, and when executed, it can include the processes of the embodiments of the methods described above. Any references to memory, storage, databases, or other media used in the embodiments provided by this invention can include non-volatile and / or volatile memory.

[0119] In summary, this invention provides a dynamic calibration and optimization method for engine model parameters. The method includes: obtaining the target parameters to be optimized and determining the corresponding dynamic optimization problem based on the target parameters and the working scenario; calling an algorithm portfolio according to the dynamic optimization problem and updating the optimal solutions of each algorithm in the portfolio through optimization information interaction and sharing strategies; determining the optimization improvement corresponding to each algorithm based on the optimal solutions obtained by each algorithm, and allocating corresponding computing resources to each algorithm according to the determined optimization improvement; optimizing the target parameters to be optimized using the allocated algorithm portfolio, and calibrating the optimized target parameters to meet performance indicators based on the actual and expected output values ​​of the aero-engine simulation model. This invention applies an algorithm portfolio to the dynamic calibration problem of engine parameters, improving the robustness and stability of parameter calibration.

[0120] It should be understood that the application of the present invention is not limited to the examples above. Those skilled in the art can make improvements or modifications based on the above description, and all such improvements and modifications should fall within the protection scope of the appended claims.

Claims

1. An engine model parameter dynamic calibration optimization method, characterized in that, The engine model parameter dynamic calibration optimization method comprises: acquiring target parameters to be optimized, and determining a corresponding dynamic optimization problem according to the target parameters to be optimized and a working scenario; calling an algorithm portfolio according to the dynamic optimization problem, and updating optimal solutions of algorithms in the algorithm portfolio through optimization information interaction and sharing strategies among the algorithms; wherein the optimization information interaction and sharing strategies among the algorithms are to set an archive set according to optimization populations corresponding to the algorithms, and share and update the optimal solutions of the algorithms by using the archive set; determining optimization promotion degrees corresponding to the algorithms according to the optimal solutions obtained by the algorithms, and allocating corresponding calculation resources to the algorithms according to the determined optimization promotion degrees; optimizing the target parameters to be optimized through the algorithm portfolio after the allocation, and calibrating the optimized target parameters as parameters meeting performance indexes according to actual output values and expected output values of an aero-engine simulation model; the calling of the algorithm portfolio according to the dynamic optimization problem, and the updating of the optimal solutions of the algorithms in the algorithm portfolio through the optimization information interaction and sharing strategies among the algorithms, comprises: calling multiple different algorithms according to the dynamic optimization problem, and organizing the algorithm portfolio according to the called algorithms; determining optimization populations corresponding to the algorithms, and setting archive sets corresponding to the algorithms; storing optimal solutions found by the algorithms in each iteration into the corresponding archive sets, so as to update the optimal solutions of the algorithms; the determination of the optimization promotion degrees corresponding to the algorithms according to the optimal solutions obtained by the algorithms, and the allocation of the corresponding calculation resources to the algorithms according to the determined optimization promotion degrees, comprises: determining optimization promotion degrees of optimal solutions obtained in a current iteration and optimal solutions obtained in a last iteration of each algorithm; allocating corresponding calculation resources to the algorithms according to a confidence upper bound algorithm and the optimization promotion degrees; the allocation of the corresponding calculation resources to the algorithms according to the confidence upper bound algorithm and the optimization promotion degrees, comprises: determining reward values of corresponding algorithms in a current iteration according to the optimization promotion degrees; substituting the reward values into the confidence upper bound algorithm, and giving the corresponding algorithms confidence in the current iteration; updating confidences corresponding to the algorithms according to iteration sequences, and allocating corresponding calculation resources to the algorithms according to the confidences after the update.

2. The method of claim 1, wherein, the acquisition of the target parameters to be optimized, and the determination of the corresponding dynamic optimization problem according to the target parameters to be optimized and a working scenario, comprises: acquiring aero-engine model parameters to be adjusted, to obtain the target parameters to be optimized; determining a working scenario of the aero-engine model; wherein the working scenario is any one or a combination of the following: a cruising scenario, a take-off scenario, and a landing scenario; modeling according to the target parameters to be optimized and the working scenario, to obtain the corresponding dynamic optimization problem.

3. The method of claim 1, wherein, the storing of optimal solutions found by the algorithms in each iteration into the corresponding archive sets, and then comprises: determining worst solutions in optimization populations corresponding to the algorithms, and replacing the worst solutions with optimal solutions in the corresponding archive sets.

4. The method of claim 1, wherein, the confidence upper bound algorithm is: ； ； wherein representing algorithm in front number of times the user is selected in the iteration process; The average return value of the first algorithm after the first iteration; and The average return value of the second algorithm after the first iteration. The average return value of the first algorithm after the second iteration; and To algorithm In The payoff when the second is chosen.

5. The method of claim 1, wherein, The target parameter to be optimized is optimized by the algorithm portfolio after the allocation, and the optimized target parameter is calibrated as a parameter satisfying the performance index according to the actual output value and the expected output value of the aero-engine simulation model, and the method comprises the steps of: The target parameter to be optimized is optimized by the algorithm portfolio after the allocation, and the optimized target parameter is obtained; The optimized target parameter is input into the aero-engine simulation model to obtain a corresponding actual output value; The difference between the actual output value and the expected output value is calculated, and the optimized target parameter is calibrated as a parameter satisfying the performance index of the working scene according to the optimization objective function.

6. A terminal, characterized by comprising: Comprise: A processor and a memory, the memory stores an engine model parameter dynamic calibration optimization program, and the engine model parameter dynamic calibration optimization program is used to implement the operation of the engine model parameter dynamic calibration optimization method in any one of claims 1-5 when executed by the processor.

7. A storage medium, characterized by The storage medium is a computer readable storage medium, and the storage medium stores an engine model parameter dynamic calibration optimization program, and the engine model parameter dynamic calibration optimization program is used to implement the operation of the engine model parameter dynamic calibration optimization method in any one of claims 1-5 when executed by the processor.

Images