Electromechanical system simulation parameter single working condition calibration method

Abstract

The application provides a single-working-condition calibration method for electromechanical system simulation parameters. The method comprises the following steps: constructing a target function with the highest consistency between simulation data output by an electromechanical system simulation model under a target working condition and measured data output by an actual electromechanical system as the target; performing single-target optimization on the target function and adding the single-target optimization result to an experience knowledge base; dividing each solution in the experience knowledge base into multiple subspaces to perform primary Bayesian optimization in each subspace; adding the primary Bayesian optimization result to the experience knowledge base, performing multi-modal detection on each solution to perform secondary Bayesian optimization in each mode; adding the secondary Bayesian optimization result to the experience knowledge base, and re-executing the cycle based on the secondary Bayesian optimization result to obtain the single-target optimization result when the termination condition is met as the single-working-condition calibration result of the simulation parameters of the electromechanical system simulation model. The application can more accurately obtain the real optimal solution of the simulation parameters of the electromechanical system.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of simulation model calibration, in particular to a method for calibrating simulation parameters of an electromechanical system under a single working condition. BACKGROUND

[0002] Electromechanical systems are widely used in manufacturing industry, automotive industry, aerospace, medical equipment, consumer electronics, energy systems and other fields, and are the cross integration of mechanical engineering, electronic engineering, control engineering, computer science and information technology, etc., and have the characteristics of high integration. Due to the high integration of electromechanical systems, it is often necessary to use simulation models to test and verify their functions and performance, so it is necessary to establish accurate simulation models of electromechanical systems.

[0003] The essence of the simulation model of the electromechanical system is an approximation of the actual electromechanical system. In order to establish an accurate simulation model of the electromechanical system, the consistency of the simulation data output by the simulation model of the electromechanical system and the measured data output by the actual electromechanical system is usually evaluated. When the consistency is poor, the simulation parameters of the simulation model of the electromechanical system are adjusted to obtain simulation data again, and then the consistency is evaluated again until the simulation data and the measured data reach a high consistency. This process is the calibration of the simulation parameters of the simulation model of the electromechanical system.

[0004] In the process of calibrating the simulation parameters of the simulation model of the electromechanical system, the inventors found that: currently, multi-agent model assisted evolutionary algorithm is usually used to calibrate the simulation parameters of the simulation model of the electromechanical system to overcome the problem that the target function calculation cost of the simulation model of the electromechanical system is high and the simulation model behaves as a complex black box function, and to efficiently search for the global optimal solution of the simulation model of the electromechanical system under limited computing resources. However, even if the simulation parameters of the electromechanical system are calibrated under a single working condition, since the consistency of the simulation data and the measured data is obtained by comprehensively considering multiple indexes of the electromechanical system, the calibration of the simulation parameters of the electromechanical system under a single working condition may have multiple global optimal solutions or multiple global or local optimal solutions, and the decision maker often needs to find the real optimal solution among the multiple global optimal solutions or the multiple global or local optimal solutions. SUMMARY

[0005] The embodiments of the present application provide a method for calibrating simulation parameters of an electromechanical system under a single working condition to solve the problem that the current method for calibrating simulation parameters of an electromechanical system under a single working condition often cannot accurately determine the real optimal solution.

[0006] In a first aspect, the embodiments of the present application provide a method for calibrating simulation parameters of an electromechanical system under a single working condition, comprising:

[0007] A target function is constructed with the highest consistency of the simulation data output by the simulation model of the electromechanical system under the target working condition and the measured data output by the actual electromechanical system as the target.

[0008] taking each simulation parameter of the electromechanical system simulation model as a decision vector, performing single-objective optimization on the target function, and adding the single-objective optimization result to the experience knowledge base;

[0009] dividing each solution about each simulation parameter of the electromechanical system simulation model in the experience knowledge base into a plurality of subspaces, and constructing a proxy model in each subspace to perform primary Bayesian optimization;

[0010] adding the primary Bayesian optimization result to the experience knowledge base, performing multi-modal detection on each solution in the experience knowledge base, constructing a proxy model in each detected mode to perform secondary Bayesian optimization;

[0011] adding the secondary Bayesian optimization result to the experience knowledge base, and re-executing the step of performing single-objective optimization on the target function and subsequent steps based on the secondary Bayesian optimization result, until a termination condition is met, and obtaining the simulation parameter single-working-condition calibration result of the electromechanical system simulation model according to the single-objective optimization result.

[0012] In a possible implementation, dividing each solution about each simulation parameter of the electromechanical system simulation model in the experience knowledge base into a plurality of subspaces comprises:

[0013] obtaining an optimal solution in each solution about each simulation parameter of the electromechanical system simulation model in the experience knowledge base;

[0014] determining (n-1) solutions most adjacent to the optimal solution as a center to form a current subspace with n solutions;

[0015] deleting the current subspace from the experience knowledge base to obtain a new experience knowledge base, re-executing the step of obtaining an optimal solution in each solution about each simulation parameter of the electromechanical system simulation model in the experience knowledge base and subsequent steps, until the number of solutions in the new experience knowledge base is less than n, and obtaining a plurality of subspaces.

[0016] In a possible implementation, the determination process of the number of subspaces is:

[0017] determining the number of subspaces according to

[0018] wherein m (N t (t) is the number of subspaces at iteration time t, N t is the number of solutions in the experience knowledge base at iteration time t, n is the number of solutions in the subspace, θ is a subspace number control parameter, θ ∈ [0.5, 1], and N max is the maximum target function calculation number of the algorithm.​

[0019] In a possible implementation, after obtaining the plurality of subspaces, the method further includes:

[0020] each of the subspaces is recorded as a first subspace, each edge solution in each first subspace is taken as a center, (n-1) solutions closest to each edge solution are determined, and a plurality of new subspaces with a solution number of n are constructed, which are recorded as second subspaces;

[0021] the constructing, in each subspace, of the surrogate model for the preliminary Bayesian optimization includes:

[0022] each target subspace is determined according to each first subspace and each second subspace;

[0023] the constructing, in each target subspace, of the surrogate model for the preliminary Bayesian optimization.

[0024] In a possible implementation, the determining of each target subspace according to each first subspace and each second subspace includes:

[0025] a union set of each second subspace corresponding to each first subspace is calculated, which is recorded as a second target subspace corresponding to each first subspace;

[0026] each first subspace and each second target subspace are respectively determined as a target subspace.

[0027] In a possible implementation, the multi-modal detection of each solution in the experience knowledge base includes:

[0028] an optimal solution is selected from each solution in the experience knowledge base as a seed;

[0029] a distance between each solution in the experience knowledge base except each seed and each seed is respectively calculated;

[0030] the distance between each solution in the experience knowledge base except each seed and each seed is respectively compared with a set distance threshold;

[0031] if the distance between a solution in the experience knowledge base and each seed is greater than the set distance threshold, the solution is taken as a new seed;

[0032] the calculating of the distance and the subsequent steps are re-executed based on the new seed until all solutions in the experience knowledge base are traversed, and a plurality of seeds in the experience knowledge base are obtained, where each seed represents a mode.

[0033] In a possible implementation, the determining of the distance threshold comprises:

[0034] calculating a distance between the optimal solution in each subspace and other solutions in the subspace, denoted as a first distance;

[0035] determining a representative distance corresponding to each subspace according to each first distance in the subspace;

[0036] determining the distance threshold according to the representative distance corresponding to each subspace.

[0037] In a second aspect, an embodiment of the present application provides a device for single-working-condition calibration of simulation parameters of a mechatronic system, comprising:

[0038] a target function construction module, configured to construct a target function with the highest consistency between simulation data output by a simulation model of the mechatronic system under a target working condition and measured data output by the actual mechatronic system as a target;

[0039] a first optimization module, configured to perform single-target optimization on the target function with each simulation parameter of the simulation model of the mechatronic system as a decision vector, and add a single-target optimization result to an experience knowledge base;

[0040] a second optimization module, configured to divide each solution about each simulation parameter of the simulation model of the mechatronic system in the experience knowledge base into a plurality of subspaces, and construct a proxy model in each subspace to perform primary Bayesian optimization;

[0041] a third optimization module, configured to add a primary Bayesian optimization result to the experience knowledge base, perform multi-modal detection on each solution in the experience knowledge base, construct a proxy model in each detected mode to perform secondary Bayesian optimization;

[0042] a cyclic calibration module, configured to add a secondary Bayesian optimization result to the experience knowledge base, and re-execute the step of performing single-target optimization on the target function and subsequent steps based on the secondary Bayesian optimization result, until a termination condition is met, and obtain a single-working-condition calibration result of simulation parameters of the simulation model of the mechatronic system according to a single-target optimization result.

[0043] In a third aspect, an embodiment of the present application provides a terminal, comprising a memory and a processor, the memory is configured to store a computer program, and the processor is configured to call and run the computer program stored in the memory to execute steps of the method in the first aspect or any possible implementation manner of the first aspect.

[0044] In a fourth aspect, an embodiment of the present application provides a computer readable storage medium storing a computer program, the computer program, when executed by a processor, implements the steps of the method in the first aspect or any possible implementation manner of the first aspect.

[0045] The embodiment of the present application provides a single-working-condition calibration method for simulation parameters of an electromechanical system. The method comprises the following steps: constructing a target function by taking the consistency between simulation data output by a simulation model of the electromechanical system under a target working condition and measured data output by the actual electromechanical system as the highest goal; taking each simulation parameter of the simulation model of the electromechanical system as a decision vector, performing single-objective optimization on the target function, and adding the single-objective optimization result to an experience knowledge base; further dividing each solution of each simulation parameter of the simulation model of the electromechanical system in the experience knowledge base into multiple subspaces, constructing a proxy model in each subspace to perform primary Bayesian optimization; then adding the primary Bayesian optimization result to the experience knowledge base, performing multi-modal detection on each solution in the experience knowledge base, constructing a proxy model in each mode detected to perform secondary Bayesian optimization; finally adding the secondary Bayesian optimization result to the experience knowledge base, and re-executing the cycle of "performing single-objective optimization on the target function" based on the secondary Bayesian optimization result, and then obtaining the single-working-condition calibration result of the simulation parameters of the simulation model of the electromechanical system according to the single-objective optimization result when the termination condition is met. On the one hand, the global optimization of the simulation parameters of the simulation model of the electromechanical system can be realized through single-objective optimization, and on the other hand, the local optimization can be more targeted based on the global optimization and the primary Bayesian optimization. Moreover, after obtaining the local optimization of the primary Bayesian optimization, the local optimization of each mode (i.e., each global optimal solution or local optimal solution) in the result of the local optimization can be further performed through the secondary Bayesian optimization, so that the real optimal solution of the simulation parameters of the electromechanical system can be more accurately obtained, thereby facilitating the improvement of the accuracy of the simulation model of the electromechanical system under the single working condition. BRIEF DESCRIPTION OF DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following will briefly introduce the drawings needed to be used in the embodiments or prior art description. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without any creative labor.

[0047] Figure 1 is the implementation flowchart of the single-working-condition calibration method for simulation parameters of an electromechanical system provided by the embodiment of the present application;

[0048] Figure 2 is the implementation block diagram of the single-working-condition calibration method for simulation parameters of an electromechanical system provided by the embodiment of the present application;

[0049] Figure 3 is a structural schematic diagram of the electromechanical system simulation parameter single working condition calibration device provided by the embodiment of the present application.

[0050] Figure 4 is a schematic diagram of the terminal provided by the embodiment of the present application. DETAILED DESCRIPTION

[0051] In the following description, for the purpose of explanation and not limitation, specific details are set forth, such as particular system configurations, techniques, etc., in order to provide a thorough understanding of the embodiments of the present application. However, it will be apparent to those skilled in the art that the present application can be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, devices, circuits, and methods are omitted so as not to obscure the description of the present application with unnecessary detail.

[0052] In order to make the objects, technical solutions and advantages of the present application clearer, the following will be described by specific embodiments in conjunction with the accompanying drawings.

[0053] Referring to Figure 1 , it shows an implementation flowchart of the electromechanical system simulation parameter single working condition calibration method provided by the embodiment of the present application, which is described in detail as follows:

[0054] In step 101, a target function is constructed with the highest consistency between the simulation data output by the electromechanical system simulation model under the target working condition and the measured data output by the actual electromechanical system as the target.

[0055] For example, the target function can be to maximize i wherein ω i is the weight coefficient of the i th index under the target working condition, s is the consistency evaluation score of the i th index under the target working condition.

[0056] The target function is constructed with the highest consistency between the simulation data output by the electromechanical system simulation model under the target working condition and the measured data output by the actual electromechanical system as the target, which can convert the minimum error of different indexes into the highest consistency of each index, so as to convert the multi-objective optimization problem into a single-objective optimization problem, thereby facilitating the optimization and calibration of the simulation parameters of the electromechanical system simulation model.

[0057] In step 102, each simulation parameter of the electromechanical system simulation model is taken as a decision vector, the target function is single-objective optimized, and the single-objective optimization result is added to the experience knowledge base.

[0058] For example, in combination with Figure 2As shown, a particle swarm optimization (PSO) algorithm can be used to perform single-objective optimization on the objective function. The PSO algorithm simulates the social behavior of a bird swarm searching for food, in which each "particle" represents a potential solution in the solution space. The particles update their positions based on individual and group experience, effectively exploring the solution space. This algorithm is simple, intuitive and easy to implement, and is particularly suitable for complex optimization scenarios where the analytical model is not clear or there is no gradient information.

[0059] In the single-objective optimization based on PSO, first, the population size, the speed and position of each particle are initialized.

[0060] Then, the fitness value of the initial population, i.e., the value of the objective function corresponding to the particles, is calculated, and the historical optimal solution of the particle and the global optimal solution are recorded.

[0061] Then, based on the speed and position of the initialized particles, as well as the historical optimal solution of the particle and the global optimal solution, the speed and position of the particles are updated by combining the speed update formula and the position update formula.

[0062] Then, the fitness value corresponding to the updated particles is calculated, and based on the fitness value corresponding to the updated particles, the historical optimal solution of the particle and the global optimal solution can be updated.

[0063] In this embodiment, the PSO algorithm simulates the social behavior of a bird swarm to achieve effective global search, and the collective wisdom helps to avoid falling into local optimum and improves the probability of finding the global optimal solution. In addition, the historical information recorded by the PSO algorithm (i.e., the historical global optimal solution) can be added to the experience knowledge base to provide dynamic prior knowledge for subsequent Bayesian optimization through the experience knowledge base, further enhancing the breadth and depth of the search.

[0064] In step 103, each solution of each simulation parameter of the electromechanical system simulation model in the experience knowledge base is divided into multiple subspaces, and a surrogate model is constructed in each subspace to perform initial Bayesian optimization.

[0065] In the Bayesian optimization (BO), a Gaussian process is used as a surrogate model, and global search and local optimization are combined to handle unknown objective functions. By integrating prior knowledge and real-time data updates, the algorithm predicts the behavior of the objective function with uncertainty. BO is particularly suitable for complex optimization problems that need to consider the influence of measurement noise.

[0066] In this embodiment, considering that the BO has high computational cost and long update period when processing large-scale data sets or high-dimensional problems, each solution in the experience knowledge base is divided into multiple subspaces, and a proxy model is constructed in each subspace to perform the first Bayesian optimization. By constructing a high-precision proxy model through Bayesian optimization of multiple subspaces, fast and accurate local optimization can be achieved under limited computing resources.

[0067] In step 104, the first Bayesian optimization result is added to the experience knowledge base, and each solution in the experience knowledge base is subjected to multi-modal detection, and a proxy model is constructed in each detected mode to perform the second Bayesian optimization.

[0068] In combination Figure 2 , after adding the first Bayesian optimization result to the experience knowledge base, considering that the solutions in the experience knowledge base may contain multiple global optimal solutions or multiple global optimal solutions and multiple local optimal solutions, in order to more accurately determine the real optimal solution, each solution in the experience knowledge base is subjected to multi-modal detection, that is, to detect how many global optimal solutions and / or local optimal solutions are contained in the experience knowledge base, so as to perform optimization in each global optimal solution and / or local optimal solution again.

[0069] In step 105, the second Bayesian optimization result is added to the experience knowledge base, and the step of performing single-objective optimization on the target function and subsequent steps are re-executed based on the second Bayesian optimization result, until the termination condition is met, and the simulation parameter single-working-condition calibration result of the electromechanical system simulation model is obtained according to the single-objective optimization result.

[0070] As Figure 2 shown, in this embodiment, the optimal solution of PSO is fed back to the first Bayesian optimization at each iteration, the optimal solution of the first Bayesian optimization is fed back to the second Bayesian optimization, and the optimal solution of the second Bayesian optimization is fed back to PSO, thereby realizing dynamic sharing of information in the optimization process, which not only improves the global search capability, but also significantly improves the overall optimization performance.

[0071] The embodiment of the application aims to build a target function by taking the consistency between the simulation data output by the electromechanical system simulation model under the target working condition and the measured data output by the actual electromechanical system as the highest target; then, each simulation parameter of the electromechanical system simulation model is taken as a decision vector to perform single-target optimization on the target function, and the single-target optimization result is added to the experience knowledge base; then, each solution of each simulation parameter of the electromechanical system simulation model in the experience knowledge base is divided into multiple subspaces, and a proxy model is built in each subspace to perform primary Bayesian optimization; then, the primary Bayesian optimization result is added to the experience knowledge base, and each solution in the experience knowledge base is detected for multiple modes, and a proxy model is built in each mode detected to perform secondary Bayesian optimization; finally, the secondary Bayesian optimization result is added to the experience knowledge base, and the cycle of "single-target optimization on the target function" is re-executed based on the secondary Bayesian optimization result, and then the single-target optimization result when the termination condition is met is taken as the simulation parameter single-working-condition calibration result of the electromechanical system simulation model. On the one hand, the global optimization of the simulation parameter of the electromechanical system simulation model can be realized through single-target optimization, and on the other hand, the local optimization can be more targeted based on the global optimization and the primary Bayesian optimization. Moreover, after the local optimization of the primary Bayesian optimization is obtained, the secondary Bayesian optimization is performed on each mode (i.e., each global optimal solution or local optimal solution) in the local optimization result to perform local optimization again, so that the real optimal solution of the electromechanical system simulation parameter can be more accurately obtained, thereby facilitating the improvement of the accuracy of the electromechanical system simulation model under the single working condition.

[0072] In an embodiment, dividing each solution of each simulation parameter of the electromechanical system simulation model in the experience knowledge base into multiple subspaces can include:

[0073] Obtaining the optimal solution in each solution of each simulation parameter of the electromechanical system simulation model in the experience knowledge base.

[0074] Taking the optimal solution as the center, determining (n-1) solutions closest to the optimal solution to form a current subspace with n solutions.

[0075] Deleting the current subspace from the experience knowledge base to obtain a new experience knowledge base, and re-executing the step of "obtaining the optimal solution in each solution of each simulation parameter of the electromechanical system simulation model in the experience knowledge base" and subsequent steps until the number of solutions in the new experience knowledge base is less than n, to obtain multiple subspaces.

[0076] For example, assuming that the experience knowledge base D is known, and the number of samples (i.e., the number of solutions) in each subspace is n, first, the optimal sample point (i.e., the optimal solution) can be selected as the starting point from the experience knowledge base D, which can be obtained by α1=argminf(x),x∈K1, wherein K1=D is the current experience knowledge base.

[0077] In the process of constructing the subspaces, the n-1 nearest sample points to a1 are found to form a set of subspaces M1. Specifically, the n-1 nearest sample points to a1 can be found by a distance metric such as Euclidean distance.

[0078] Then, the construction of the subspaces is repeated, i.e., the samples contained in M1 are removed from K1 or D, i.e., K2 = K1 \ M1, and the above process of finding the optimal point and constructing the subspaces is repeated in K2 until all samples are assigned to a certain subspace, and finally a set of subspaces is obtained, each of which can focus on the optimization of its specific region to enhance the global search ability.

[0079] In this embodiment, by dividing a large data set into multiple smaller subsets and constructing proxy models on these subsets in parallel, the overall computational burden can be significantly reduced and the optimization process can be accelerated, and in addition, the existing sample information can be more accurately and efficiently utilized to provide the possibility of dynamic adjustment for the optimization algorithm.

[0080] For example, the process of determining the number of subspaces can be:

[0081] According to the number of subspaces is determined.

[0082] where m(N t ) is the number of subspaces at iteration time t, N t is the number of solutions in the experience knowledge base at iteration time t, n is the number of solutions in the subspace, θ is a subspace number control parameter, θ ∈ [0.5, 1], N max is the maximum number of objective function calculations of the algorithm.

[0083] This embodiment considers that as the amount of samples in the experience knowledge base increases, the number of subspaces also increases, in order to provide appropriate exploration and convergence speed at different stages of the algorithm, the number of subspaces m(N t ) is dynamically adjusted, in the early stage of algorithm iteration, a larger number of subspaces is provided to provide extensive spatial exploration to avoid falling into local optimum. In the later stage of algorithm iteration, a smaller number of subspaces is used to speed up the convergence speed and focus on subspaces with higher confidence. Moreover, the calculation formula for determining the number of subspaces is simple, which can avoid introducing complex calculation process.

[0084] In an embodiment, after obtaining multiple subspaces, the following can also be included:

[0085] Each of the subspaces is denoted as a first subspace, and each edge solution in each of the first subspaces is taken as a center to determine (n-1) solutions closest to each of the edge solutions, to form a plurality of new subspaces with a solution number of n, denoted as second subspaces.

[0086] Correspondingly, constructing a proxy model in each of the subspaces to perform the preliminary Bayesian optimization can include:

[0087] According to each of the first subspaces and each of the second subspaces, each of the target subspaces is determined.

[0088] A proxy model is constructed in each of the target subspaces to perform the preliminary Bayesian optimization.

[0089] In this embodiment, after the experience knowledge base is divided into a plurality of subspaces, in order to enhance the exploration of the edges of each of the subspaces, each of the subspaces is taken as a center to construct a plurality of new subspaces, so as to achieve sufficient exploration of the local solutions and the edge solutions by dividing the subspaces and the new subspaces of the experience knowledge base, thereby helping to optimize the overall optimization performance.

[0090] For example, according to each of the first subspaces and each of the second subspaces, each of the target subspaces can be determined by:

[0091] The union of each of the second subspaces corresponding to each of the first subspaces is calculated, denoted as a second target subspace corresponding to each of the first subspaces.

[0092] Each of the first subspaces and each of the second target subspaces is determined as a target subspace, respectively.

[0093] In this embodiment, in order to sufficiently explore the edges of each of the subspaces and avoid increasing too many subspaces, after constructing each of the second subspaces corresponding to each of the first subspaces, the union of each of the second subspaces is calculated as a second target subspace, so as to sufficiently explore the edges of each of the subspaces by using the second target subspace.

[0094] In an embodiment, the multi-modal detection can be performed on each of the solutions in the experience knowledge base, which can include:

[0095] An optimal solution is selected from each of the solutions in the experience knowledge base as a seed.

[0096] The distance between each of the solutions in the experience knowledge base except for each of the seeds and each of the seeds is calculated, respectively.

[0097] The distance between each of the solutions in the experience knowledge base except for each of the seeds and each of the seeds is compared with a set distance threshold, respectively.

[0098] If the distance between each seed and a solution in the experience knowledge base is greater than the set distance threshold, the solution is taken as a new seed.

[0099] The steps of calculating the distance between each seed and each solution in the experience knowledge base and the subsequent steps are re-executed based on the new seed until all solutions in the experience knowledge base are traversed, and multiple seeds in the experience knowledge base are obtained, wherein each seed represents a modality.

[0100] For example, the determination process of the set distance threshold can be as follows:

[0101] The distance between the optimal solution in each subspace and other solutions in the subspace is calculated, denoted as the first distance.

[0102] According to each first distance in each subspace, the representative distance corresponding to each subspace is determined.

[0103] According to the representative distance corresponding to each subspace, the set distance threshold is determined.

[0104] In this embodiment, when each solution in the experience knowledge base is detected for multi-modalities, in order to more reasonably identify different modalities, the set distance threshold can be determined by the distance between the optimal solution in each subspace and other solutions in the subspace when the subspace is divided before the initial Bayesian optimization. For example, the representative distance of each subspace is determined by the mean of each first distance in each subspace, and the set distance threshold is determined by the mean of the representative distance of each subspace. Or the representative distance of each subspace is determined by the median and maximum of each first distance in each subspace, and the set distance threshold is determined by the mean, median and maximum of the representative distance of each subspace.

[0105] It should be understood that the size of the serial number of each step in the above embodiment does not mean the order of execution, and the execution order of each process should be determined according to its function and internal logic, and should not constitute any limitation on the implementation process of the embodiment of the present application.

[0106] The following is a device embodiment of the present application, and for details not described in detail, reference can be made to the corresponding method embodiments described above.

[0107] Figure 3 The structural schematic diagram of the electromechanical system simulation parameter single working condition calibration device provided by the embodiment of the present application is shown, only the parts related to the embodiment of the present application are shown for the convenience of description, and the details are described as follows:

[0108] As shown in Figure 3 The electromechanical system simulation parameter single working condition calibration device includes a target function construction module 31, a first optimization module 32, a second optimization module 33, a third optimization module 34 and a cyclic calibration module 35.

[0109] a target function construction module 31, configured to construct a target function with the highest consistency between simulation data output by the electromechanical system simulation model under the target working condition and measured data actually output by the electromechanical system as a target;

[0110] a first optimization module 32, configured to perform single-target optimization on the target function with each simulation parameter of the electromechanical system simulation model as a decision vector, and add a single-target optimization result to an experience knowledge base;

[0111] a second optimization module 33, configured to divide each solution about each simulation parameter of the electromechanical system simulation model in the experience knowledge base into multiple subspaces, and construct a proxy model in each subspace to perform primary Bayesian optimization;

[0112] a third optimization module 34, configured to add a primary Bayesian optimization result to the experience knowledge base, perform multi-modal detection on each solution in the experience knowledge base, construct a proxy model in each detected mode to perform secondary Bayesian optimization;

[0113] a cycle calibration module 35, configured to add a secondary Bayesian optimization result to the experience knowledge base, and re-perform the step of performing single-target optimization on the target function and subsequent steps based on the secondary Bayesian optimization result, until a termination condition is met, and obtain a single-working-condition calibration result of the simulation parameter of the electromechanical system simulation model according to a single-target optimization result.

[0114] This invention constructs an objective function with the goal of maximizing the consistency between the simulation data output by the electromechanical system simulation model and the measured data output by the actual electromechanical system under the target operating condition. Then, using each simulation parameter of the electromechanical system simulation model as a decision vector, single-objective optimization is performed on the objective function, and the single-objective optimization result is added to an empirical knowledge base. Next, the solutions for each simulation parameter of the electromechanical system simulation model in the empirical knowledge base are divided into multiple subspaces, and a surrogate model is constructed in each subspace for initial Bayesian optimization. The initial Bayesian optimization result is then added to the empirical knowledge base, and multimodal detection is performed on each solution in the empirical knowledge base. A surrogate model is constructed in each detected modality for secondary Bayesian optimization. Finally, the secondary Bayesian optimization result is added to the empirical knowledge base, and the loop of "single-objective optimization of the objective function" is re-executed based on the secondary Bayesian optimization result. Finally, the single-operating-condition calibration result of the simulation parameters of the electromechanical system simulation model is obtained based on the single-objective optimization result when the termination condition is met. On the one hand, global optimization of simulation parameters of electromechanical system simulation models can be achieved through single-objective optimization. On the other hand, performing initial Bayesian optimization based on global optimization allows for more targeted local optimization. Moreover, after obtaining the local optimization results from the initial Bayesian optimization, performing secondary Bayesian optimization on each mode (i.e., each global optimal solution or local optimal solution) in the local optimization results can more accurately obtain the true optimal solution of the electromechanical system simulation parameters, thereby improving the accuracy of the electromechanical system simulation model under single operating conditions.

[0115] In one possible implementation, the second optimization module 33 can be used to obtain the optimal solution among the solutions for each simulation parameter of the electromechanical system simulation model in the experience knowledge base; take the optimal solution as the center, determine the (n-1) solutions closest to the optimal solution to form a current subspace with n solutions; delete the current subspace from the experience knowledge base to obtain a new experience knowledge base, and re-execute the step of "obtaining the optimal solution among the solutions for each simulation parameter of the electromechanical system simulation model in the experience knowledge base" and subsequent steps until the number of solutions in the new experience knowledge base is less than n, thus obtaining multiple subspaces.

[0116] In one possible implementation, the second optimization module 33 can be used to... Determine the number of the subspaces;

[0117] Where, m(N) t N represents the number of subspaces at iteration time t. t Let be the number of solutions in the empirical knowledge base at iteration time t, n be the number of solutions in the subspace, and θ be the subspace number control parameter, θ∈[0.5,1], N maxThe number of times the algorithm's maximum objective function is computed.

[0118] In one possible implementation, the second optimization module 33 can also be used to denote each subspace as a first subspace, and, taking each edge solution in each first subspace as the center, determine the (n-1) nearest neighbors to each edge solution, forming multiple new subspaces with n solutions, denoted as second subspaces; the step of constructing a surrogate model in each subspace for initial Bayesian optimization includes: determining each target subspace based on each first subspace and each second subspace; and constructing a surrogate model in each target subspace for initial Bayesian optimization.

[0119] In one possible implementation, the second optimization module 33 can be used to calculate the union of each second subspace corresponding to each first subspace, denoted as the second target subspace corresponding to each first subspace; and to determine each first subspace and each second target subspace as a target subspace.

[0120] In one possible implementation, the third optimization module 34 can be used to select the optimal solution from the solutions in the empirical knowledge base as a seed; calculate the distance between each solution in the empirical knowledge base other than the seeds and each seed; compare the distance between each solution in the empirical knowledge base other than the seeds and each seed with a set distance threshold; if the distance between a certain solution in the empirical knowledge base and each seed is greater than the set distance threshold, then the solution is used as a new seed; based on the new seed, the steps of "calculating the distance between each solution in the empirical knowledge base other than the seeds and each seed" and subsequent steps are repeated until all solutions in the empirical knowledge base are traversed, resulting in multiple seeds in the empirical knowledge base, where each seed represents a mode.

[0121] In one possible implementation, the third optimization module 34 can be used to calculate the distance between the optimal solution in each subspace and other solutions in that subspace, denoted as the first distance; determine the representative distance corresponding to each subspace based on each of the first distances in each subspace; and determine the set distance threshold based on the representative distance corresponding to each subspace.

[0122] Figure 4 This is a schematic diagram of a terminal provided in an embodiment of the present invention. Figure 4 As shown, the terminal 4 in this embodiment includes a processor 40, a memory 41, and a computer program 42 stored in the memory 41 and executable on the processor 40. When the processor 40 executes the computer program 42, it implements the steps in the above embodiments of the single-condition calibration method for electromechanical system simulation parameters, for example... Figure 1The steps 101-105 are shown. Alternatively, the processor 40 implements the functions of the modules / units in the above-mentioned device embodiments when executing the computer program 42, for example Figure 3 The functions of the modules / units 31-35 are shown.

[0123] The computer program 42 can be segmented into one or more modules / units for example, which are stored in the memory 41 and executed by the processor 40 to complete the present application. One or more modules / units can be a series of computer program instruction segments capable of completing a specific function, which are used to describe the execution process of the computer program 42 in the terminal 4. For example, the computer program 42 can be segmented into Figure 3 The modules / units 31-35 are shown.

[0124] The terminal 4 can be a desktop computer, a notebook, a palm computer, a cloud server and other computing devices. The terminal 4 can include, but is not limited to, the processor 40 and the memory 41. Those skilled in the art can understand that Figure 4 The terminal 4 is only an example and does not constitute a limitation on the terminal 4, which can include more or fewer components than shown, or combine certain components, or different components, for example, the terminal can also include an input / output device, a network access device, a bus, etc.

[0125] The processor 40 can be a central processing unit (CPU), and can also be other general-purpose processors, digital signal processors (DSP), application specific integrated circuits (ASIC), field programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor can be a microprocessor or the processor can also be any conventional processor.

[0126] The memory 41 can be an internal storage unit of the terminal 4, such as a hard disk or a memory of the terminal 4. The memory 41 can also be an external storage device of the terminal 4, such as a plug-in hard disk, a smart media card (SMC), a secure digital (SD) card, a flash card, or the like equipped on the terminal 4. Further, the memory 41 can also include both the internal storage unit and the external storage device of the terminal 4. The memory 41 is used to store computer programs and other programs and data required by the terminal. The memory 41 can also be used to temporarily store data that has been output or will be output.

[0127] It should be clearly understood by those skilled in the art that, for the convenience and brevity of description, only the above-mentioned division of each functional unit and module is exemplified, and in actual application, the above-mentioned functions can be completed by different functional units and modules according to needs, that is, the internal structure of the device is divided into different functional units or modules to complete all or part of the functions described above. Each functional unit and module in the embodiment can be integrated in one processing unit, or each unit can exist physically, or two or more units can be integrated in one unit. The integrated unit can be realized in the form of hardware or in the form of software functional unit. In addition, the specific names of each functional unit and module are only for the convenience of mutual distinction, and do not limit the protection scope of the present application. The specific working process of the unit and module in the above system can refer to the corresponding process in the foregoing method embodiments, which will not be described here.

[0128] In the above embodiments, the description of each embodiment has its own emphasis, and the parts not described or recorded in detail in a certain embodiment can be referred to the relevant description of other embodiments.

[0129] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be realized in electronic hardware or in a combination of computer software and electronic hardware. Whether the functions are executed in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the present application.

[0130] In the embodiments of the present application, it should be understood that the disclosed apparatus / terminal and method can be implemented in other manners. For example, the apparatus / terminal embodiments described above are merely schematic, and the division of the modules or units is merely logical function division, and there can be another division manner in actual implementation. For example, a plurality of units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the displayed or discussed mutual couplings or direct couplings or communication connections can be indirect couplings or communication connections through some interfaces, devices or units, and can be electrical, mechanical or in other forms.

[0131] The units described as separated components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e., can be located in one place or can be distributed on a plurality of network units. Some or all of the units can be selected according to actual needs to achieve the purposes of the embodiments of the present application. In addition, the functional units in each embodiment of the present application can be integrated in a processing unit, or each unit can be a physically independent unit, or two or more units can be integrated in a unit. The integrated unit can be implemented in the form of hardware or in the form of a software functional unit.

[0132] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer readable storage medium. Based on such understanding, all or part of the flow of the above-mentioned embodiment method can be implemented by a computer program instructing related hardware to complete, and the computer program can be stored in a computer readable storage medium. When the processor executes the computer program, the steps of each of the above-mentioned mechanical system simulation parameter single working condition calibration method embodiments can be implemented. The computer program includes computer program code, which can be in the form of source code, object code, executable file or some intermediate form. The computer readable medium can include any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, magnetic disk, optical disk, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signal, telecommunication signal and software distribution medium, etc.

[0133] The above examples are only used to illustrate the technical solutions of the present application, and are not intended to limit the present application; although the present application has been described in detail with reference to the foregoing examples, those skilled in the art should understand that the technical solutions recorded in the foregoing examples can be modified, or some technical features can be replaced by equivalent features; and these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application, and should be included in the protection scope of the present application.

Claims

1. A method for single-condition calibration of simulation parameters of an electromechanical system, characterized in that, include: The objective function is constructed with the goal of maximizing the consistency between the simulation data output by the electromechanical system simulation model and the measured data output by the actual electromechanical system under the target operating condition. Each simulation parameter of the electromechanical system simulation model is used as a decision vector to perform single-objective optimization on the objective function, and the single-objective optimization results are added to the empirical knowledge base. The solutions for each simulation parameter of the electromechanical system simulation model in the experience knowledge base are divided into multiple subspaces, and a surrogate model is constructed in each subspace for initial Bayesian optimization. The initial Bayesian optimization results are added to the empirical knowledge base. Multimodal detection is performed on each solution in the empirical knowledge base. A surrogate model is constructed within each detected modality for secondary Bayesian optimization. The results of the quadratic Bayesian optimization are added to the empirical knowledge base, and the steps of "single-objective optimization of the objective function" and subsequent steps are re-executed based on the results of the quadratic Bayesian optimization until the termination condition is met. Then, the single-condition calibration results of the simulation parameters of the electromechanical system simulation model are obtained based on the single-objective optimization results.

2. The single-condition calibration method for electromechanical system simulation parameters according to claim 1, characterized in that, The solutions for each simulation parameter of the electromechanical system simulation model in the aforementioned knowledge base are divided into multiple subspaces, including: Obtain the optimal solution from the solutions for each simulation parameter of the electromechanical system simulation model in the knowledge base. Centered on the optimal solution, determine the (n-1) nearest neighbors of the optimal solution to form a current subspace with n solutions; The current subspace is deleted from the empirical knowledge base to obtain a new empirical knowledge base. The steps of "obtaining the optimal solution among the solutions of each simulation parameter of the electromechanical system simulation model in the empirical knowledge base" and subsequent steps are repeated until the number of solutions in the new empirical knowledge base is less than n, thus obtaining multiple subspaces.

3. The single-condition calibration method for electromechanical system simulation parameters according to claim 2, characterized in that, The process for determining the number of subspaces is as follows: according to Determine the number of the subspaces; Where, m(N) t N represents the number of subspaces at iteration time t. t Let be the number of solutions in the empirical knowledge base at iteration time t, n be the number of solutions in the subspace, and θ be the subspace number control parameter, θ∈[0.5,1], N max The number of times the algorithm's maximum objective function is computed.

4. The single-condition calibration method for electromechanical system simulation parameters according to claim 2, characterized in that, After obtaining multiple subspaces, it also includes: Each subspace is designated as the first subspace. Taking each edge solution in each first subspace as the center, the (n-1) nearest solutions to each edge solution are determined to form multiple new subspaces with a solution count of n, which are designated as the second subspace. The initial Bayesian optimization, which involves constructing a proxy model in each subspace, includes: Each target subspace is determined based on each of the first subspaces and each of the second subspaces; In each target subspace, a proxy model is constructed for initial Bayesian optimization.

5. The single-condition calibration method for electromechanical system simulation parameters according to claim 4, characterized in that, Each target subspace is determined based on each of the first subspaces and each of the second subspaces, including: Calculate the union of each second subspace corresponding to each first subspace, and denote it as the second target subspace corresponding to each first subspace; Each of the first subspaces and each of the second target subspaces is defined as a target subspace.

6. The single-condition calibration method for electromechanical system simulation parameters according to claim 1, characterized in that, Multimodal detection is performed on each solution in the experience knowledge base, including: Select the optimal solution from the solutions in this knowledge base as a seed; Calculate the distance between each solution in the empirical knowledge base (excluding each of the seeds) and each of the seeds; The distance between each solution in the experience knowledge base (excluding each of the seeds) and each of the seeds is compared with a set distance threshold. If the distance between a solution in the experience knowledge base and each of the seeds is greater than the set distance threshold, then the solution is regarded as a new seed; Based on the new seeds, the steps of "calculating the distance between each solution in the empirical knowledge base other than each of the seeds and each of the seeds" and subsequent steps are repeated until all solutions in the empirical knowledge base have been traversed, resulting in multiple seeds in the empirical knowledge base, where each seed represents a mode.

7. The single-condition calibration method for electromechanical system simulation parameters according to claim 6, characterized in that, The process for determining the set distance threshold is as follows: Calculate the distance between the optimal solution in each subspace and other solutions in that subspace, and denote it as the first distance; Based on each of the first distances in each subspace, determine the representative distance corresponding to each subspace; The set distance threshold is determined based on the representative distance corresponding to each subspace.

8. A single-condition calibration device for simulation parameters of an electromechanical system, characterized in that, include: The objective function construction module is used to construct an objective function with the goal of maximizing the consistency between the simulation data output by the electromechanical system simulation model and the measured data output by the actual electromechanical system under the target operating conditions. The first optimization module is used to perform single-objective optimization on the objective function by taking each simulation parameter of the electromechanical system simulation model as a decision vector, and add the single-objective optimization result to the experience knowledge base. The second optimization module is used to divide the solutions of each simulation parameter of the electromechanical system simulation model in the experience knowledge base into multiple subspaces, and to build a surrogate model in each subspace for initial Bayesian optimization. The third optimization module is used to add the initial Bayesian optimization result to the empirical knowledge base, perform multimodal detection on each solution in the empirical knowledge base, and construct a surrogate model for secondary Bayesian optimization in each detected modality. The cyclic calibration module is used to add the quadratic Bayesian optimization results to the empirical knowledge base, and re-execute the step of "single-objective optimization of the objective function" and subsequent steps based on the quadratic Bayesian optimization results until the termination condition is met. Then, the single-condition calibration results of the simulation parameters of the electromechanical system simulation model are obtained based on the single-objective optimization results.

9. A terminal, characterized in that, It includes a memory and a processor, the memory being used to store a computer program, and the processor being used to call and run the computer program stored in the memory to perform the method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method as described in any one of claims 1 to 7 above.

Images