CN118690666B - A motor intelligent drive optimization method and system - Google Patents

Abstract

The invention provides a motor intelligent driving optimization method and a system, wherein the method comprises the following steps: determining an optimization target of motor driving, and quantifying the optimization target; constructing a motor performance evaluation model based on the optimization target, defining an adaptability function closely related to the optimization target according to the output of the motor performance evaluation model, and evaluating the advantages and disadvantages of each driving scheme; and the driving scheme is optimized through an evolutionary optimizing algorithm and an annealing optimizing algorithm by combining the evaluation of the fitness of the driving scheme so as to output an optimal motor driving scheme. According to the invention, through the combination of the motor performance evaluation model, the fitness function and the two optimizing algorithms, the multi-objective optimization problem can be processed, the driving scheme for achieving the best balance among a plurality of targets is found by searching the global optimal solution in a wide driving scheme space, and the optimal motor driving scheme with excellent performance indexes is finally output.

Description

A motor intelligent drive optimization method and system

Technical Field

The present invention relates to the technical field of motor drive optimization, and in particular to a motor intelligent drive optimization method and system.

Background Art

With the rapid development of industrial automation and the widespread application of intelligent manufacturing, motors, as core power sources, play a vital role in various equipment and systems. The performance and drive solutions of motors directly affect production efficiency, energy consumption, equipment life and other aspects. Therefore, how to optimize motor drive has become an urgent problem to be solved in the current industrial automation field.

At present, motor drive optimization mainly relies on traditional trial and error methods and empirical adjustments. This traditional method is not only inefficient but also has high R&D costs, because each adjustment requires actual testing to verify the effect. Moreover, traditional methods are incapable of dealing with multi-objective optimization problems. It is difficult to find the best balance between multiple performance indicators such as efficiency, stability, and response speed. This often leads to the resulting drive solution having outstanding performance in some aspects, but serious deficiencies in other aspects, making it difficult to fully meet the comprehensive needs of modern complex engineering applications.

Summary of the invention

Based on this, the purpose of the present invention is to propose a motor intelligent drive optimization method and system to solve the above-mentioned problems.

According to a motor intelligent drive optimization method proposed in the present invention, the method comprises:

Determining an optimization target for the motor drive and quantifying the optimization target;

Building a motor performance evaluation model based on the optimization target, and defining a fitness function closely related to the optimization target according to the output of the motor performance evaluation model to evaluate the pros and cons of each driving scheme;

Combined with the evaluation of the fitness of the drive scheme, the drive scheme is optimized through evolutionary optimization algorithm and annealing optimization algorithm to output the optimal motor drive scheme.

Furthermore, the step of constructing a motor performance evaluation model based on the optimization target includes:

Collecting the operation data and motor performance data of the motor under different driving schemes, wherein the operation data includes speed, current, voltage and temperature;

Extracting features directly related to the optimization objective from the operating data;

The extracted features and the corresponding motor performance data are used as a training set to train the deep learning model to obtain a motor performance evaluation model, which is used to predict the motor performance evaluation results under different driving schemes based on the input motor operation data.

Furthermore, in the step of defining a fitness function closely related to the optimization objective according to the output of the motor performance evaluation model:

The fitness function is used to weight and calculate various performance indicators of the motor performance evaluation result.

Furthermore, the step of combining the evaluation of the fitness of the driving scheme and optimizing the driving scheme by the evolutionary optimization algorithm and the annealing optimization algorithm to output the optimal motor driving scheme includes:

Combined with the evaluation of the fitness of the driving scheme, and through the selection, crossover and mutation operations of the evolutionary optimization algorithm, a global search is performed in the driving scheme space to find an initial optimal solution set that matches the optimization goal;

Based on the initial optimal solution set, combined with the evaluation of the fitness of the solution, and through the annealing optimization algorithm, a local search is performed in the neighborhood of the initial optimal solution to approximate the optimal solution and obtain the global optimal solution;

The global optimal solution is taken as the optimal motor driving solution.

Furthermore, the step of combining the evaluation of the fitness of the driving scheme and performing a global search in the driving scheme space through the selection, crossover and mutation operations of the evolutionary optimization algorithm to find an initial optimal solution set matching the optimization goal includes:

Randomly generate a number of initial driving schemes, and encode the driving schemes into gene sequences as individuals of the initial population;

Evaluating the fitness value of each individual in the population by combining the motor performance evaluation model and the fitness function;

According to a preset screening ratio, excellent individuals with high fitness values in the population are screened out as the next generation of the population;

Randomly select two individuals from the population after the selection operation, perform a crossover operation, generate a crossover individual, and add the crossover individual to the population;

Performing a mutation operation on the individuals in the population after the crossover operation to generate mutant individuals, and adding the mutant individuals to the population;

Sort by fitness value and select individuals with high fitness value to form a new population;

Repeat the iteration until the termination condition of the evolutionary optimization algorithm is reached, then stop the iteration, and use the current population as the initial optimal solution set, and the individuals in the current population as the initial optimal solutions.

Furthermore, the step of randomly selecting two individuals from the population after the selection operation, performing a crossover operation, generating crossover individuals, and adding the crossover individuals to the population includes:

Randomly select two individuals from the population after the selection operation as individuals to be crossed;

At the intersection point of the two individuals to be crossed, exchanging genes to create two crossover individuals, wherein the intersection point is one or more positions randomly selected in the gene sequence of the individuals to be crossed;

Determine whether the new crossover individual meets the constraints;

If satisfied, the crossover individual is added to the population.

Furthermore, the step of performing a mutation operation on the individuals in the population after the crossover operation to generate mutant individuals and adding the mutant individuals to the population includes:

Using the individuals in the population after the crossover operation as individuals to be mutated;

Randomly selecting one or more mutation points in the gene sequence of the individual to be mutated;

Randomly adjusting the gene at the mutation point to generate a mutant individual;

Determine whether the variant individual satisfies the constraint condition;

If satisfied, the mutant individual is added to the population.

Furthermore, the step of taking the initial optimal solution set as a basis, combining the evaluation of the fitness of the solution, and performing a local search in the neighborhood of the initial optimal solution through the annealing optimization algorithm to approximate the optimal solution to obtain the global optimal solution includes:

Using the initial optimal solution in the initial optimal solution set as the initial solution of the annealing optimization algorithm;

Setting an initial temperature value T ₀ higher than a preset threshold;

Randomly search for new solutions in the neighborhood of the current solution;

Decide whether to accept the new solution based on the Metropolis criteria;

The temperature is gradually lowered according to the preset cooling strategy, and the neighborhood search and acceptance criterion judgment process are repeated until the termination condition is met, and the current solution is used as the candidate set of the global optimal solution.

Furthermore, the step of judging whether to accept the new solution according to the Metropolis criterion includes:

The fitness values of the current solution and the new solution are evaluated respectively in combination with the motor performance evaluation model and the fitness function;

Calculate the difference Δf between the fitness value of the new solution and the current solution, the formula is: , where f(x′) is the fitness value of the new solution x′, and f(x) is the fitness value of the current solution x;

like , then accept the new solution and update the current solution so that and ;

like , then the preset acceptance probability is used to decide whether to accept the new solution, and the preset acceptance probability is equal to , where T is the current temperature.

The present invention also provides a motor intelligent drive optimization system, comprising:

Optimization target module: used to determine the optimization target of the motor drive and quantify the optimization target;

Performance evaluation module: used to construct a motor performance evaluation model based on the optimization target, and define a fitness function closely related to the optimization target according to the output of the motor performance evaluation model to evaluate the pros and cons of each driving scheme;

Drive scheme optimization module: used to combine the evaluation of the fitness of the drive scheme, and optimize the drive scheme through evolutionary optimization algorithm and annealing optimization algorithm to output the optimal motor drive scheme.

In summary, according to the above-mentioned motor intelligent drive optimization method, by clarifying and quantifying the optimization target of motor drive, a clear direction and evaluation criteria are provided for the optimization process; further based on the optimization target, a motor performance evaluation model is constructed so that the performance of each drive scheme can be predicted, and a fitness function closely related to the optimization target is defined according to the output of the motor performance evaluation model to accurately evaluate the pros and cons of the comprehensive performance of each drive scheme, providing an effective selection basis for the optimization algorithm; combined with the evolutionary optimization algorithm and the annealing optimization algorithm, it is not only possible to search for the optimal solution in the global range, but also to fine-tune the optimal solution initially found through the local search capability of the annealing optimization algorithm, thereby increasing the probability of finding the global optimal solution. The method of the present invention can handle multi-objective optimization problems through the combination of motor performance evaluation model, fitness function and two optimization algorithms, and can find the global optimal solution in a wide range of drive scheme space to find the drive scheme that achieves the best balance between multiple targets, and finally output the optimal motor drive scheme that performs well in multiple performance indicators, which is expected to provide better performance in practical applications and meet specific engineering needs.

Additional aspects and advantages of the present invention will be given in part in the following description and in part will be obvious from the following description or will be learned through embodiments of the present invention.

BRIEF DESCRIPTION OF THE DRAWINGS

The above and/or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

FIG1 is a flow chart of a motor intelligent drive optimization method according to a first embodiment of the present invention;

FIG. 2 is a system block diagram of a motor intelligent drive optimization system according to a second embodiment of the present invention.

DETAILED DESCRIPTION

In order to facilitate the understanding of the present invention, the present invention will be described more fully below with reference to the relevant drawings. Several embodiments of the present invention are given in the drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. On the contrary, the purpose of providing these embodiments is to make the disclosure of the present invention more thorough and comprehensive.

It should be noted that when an element is referred to as being "fixed to" another element, it may be directly on the other element or there may be a central element. When an element is considered to be "connected to" another element, it may be directly connected to the other element or there may be a central element at the same time. The terms "vertical", "horizontal", "left", "right" and similar expressions used herein are for illustrative purposes only.

Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which the present invention belongs. The terms used herein in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and/or" used herein includes any and all combinations of one or more of the related listed items.

Embodiment 1: Referring to FIG. 1 , the present invention proposes a motor intelligent drive optimization method, the method comprising steps S101 to S103:

S101, determining an optimization target of a motor drive, and quantifying the optimization target.

It should be noted that determining the optimization goals of motor drive and quantifying these goals can provide a clear direction for the subsequent optimization process and provide specific criteria for evaluating the pros and cons of different drive schemes.

First, determine the optimization goals of the motor drive. These goals revolve around the main performance indicators of the motor, which can be one or more of high efficiency, low energy consumption, fast response, smooth operation, long life, etc.

The goal of quantitative optimization is to convert it into a measurable and comparable indicator. For example, high efficiency can be measured by the ratio of the motor's output power to its input power, and a percentage of efficiency improvement can be set as a goal; low energy consumption can be quantified by measuring the motor's energy consumption per unit time, and a percentage of energy consumption reduction can be set as a goal; fast response can be quantified by measuring the time required for the motor to reach a predetermined state (such as a specific speed), and a specific goal of shortening the response time can be set as a goal; smooth operation can be quantified by vibration sensors and noise measurement equipment to quantify the vibration amplitude and noise level, and a percentage of reduction can be set as a goal; long life can be estimated by simulation or actual testing to estimate the motor's service life, and a life extension can be set as a goal.

Quantitative targets not only make the optimization process clearer and more specific, but also provide clear evaluation criteria for subsequent optimization algorithms, thereby guiding the optimization algorithm to continuously approach the optimal solution during the search process, achieving a comprehensive improvement in motor performance, and making the comparison of the advantages and disadvantages of different drive solutions more objective and accurate.

S102, constructing a motor performance evaluation model based on the optimization target, and defining a fitness function closely related to the optimization target according to an output of the motor performance evaluation model to evaluate the pros and cons of each driving scheme.

It should be noted that based on the quantitative optimization goal, a motor performance evaluation model is constructed. This model can comprehensively consider multiple performance indicators and output a comprehensive evaluation result. Model construction can include: collecting the motor's operating data under various drive schemes, including speed, current, voltage, temperature, etc.; extracting features related to the optimization goal from the collected operating data; and then using machine learning algorithms or deep learning algorithms to train the model so that it can predict the motor's performance based on the operating data under different drive schemes.

The fitness function is the key to evaluating the pros and cons of each drive solution. It is a function based on the output of the motor performance evaluation model and is closely related to the optimization goal. The fitness function can be in the form of a weighted sum, where the evaluation result of each performance indicator is multiplied by a weight coefficient and then summed, where the weight coefficient reflects the importance of each indicator.

If some optimization objectives have constraints (such as energy consumption cannot exceed a certain value), the fitness function can include processing of these constraints. For solutions that violate the constraints, a penalty term can be added to the fitness function to reduce its fitness value. For example, if the energy consumption exceeds the limit, a penalty can be imposed according to the amount of excess.

S103, combining the evaluation of the fitness of the driving scheme, and optimizing the driving scheme through an evolutionary optimization algorithm and an annealing optimization algorithm, so as to output an optimal motor driving scheme.

It should be noted that, by combining the evaluation of the fitness of the drive scheme and optimizing the drive scheme through evolutionary optimization algorithm and annealing optimization algorithm, the optimal motor drive scheme is finally output, multi-objective optimization is achieved, and the global optimal solution is found in a wide range of drive scheme space, and finally the optimal motor drive scheme with excellent performance in multiple performance indicators is output.

Among them, in the process of optimizing the drive scheme through the evolutionary optimization algorithm and the annealing optimization algorithm, the fitness of the drive scheme is evaluated. For each drive scheme, the above-mentioned trained motor performance evaluation model can be used to predict the motor performance indicators, and then the weighted sum of the performance indicators is calculated through the fitness function to obtain a comprehensive evaluation result, that is, the fitness of the drive scheme.

Since motor design usually involves the trade-off of multiple performance indicators, the evolutionary optimization algorithm and the annealing optimization algorithm can handle this multi-objective optimization problem and find the drive solution that achieves the best balance between multiple objectives by searching the solution space. Specifically, the optimal solution can be searched globally by the evolutionary optimization algorithm, and then the local search capability of the annealing optimization algorithm can be used to fine-tune the initial optimal solution to increase the probability of finding the global optimal solution. Moreover, these optimization algorithms do not have high requirements on the specific form of the problem and can handle complex and nonlinear optimization problems. Therefore, even if the relationship between motor performance and drive solutions is very complex, these algorithms can effectively find the optimal solution and improve the optimization efficiency, which is especially important for motor design processes that require rapid iteration and optimization. After algorithm optimization, the final output motor drive solution is the optimal solution obtained after comprehensive consideration of multiple performance indicators. This solution is expected to provide better performance in practical applications and meet specific engineering needs.

Based on step S101 to step S103, by clarifying and quantifying the optimization target of motor drive, a clear direction and evaluation criteria are provided for the optimization process; further based on the optimization target, a motor performance evaluation model is constructed so that the performance of each drive scheme can be predicted, and a fitness function closely related to the optimization target is defined according to the output of the motor performance evaluation model to accurately evaluate the pros and cons of the comprehensive performance of each drive scheme, providing an effective selection basis for the optimization algorithm; combined with the evolutionary optimization algorithm and the annealing optimization algorithm, not only can the optimal solution be searched in a global range, but also the local search capability of the annealing optimization algorithm can be used to fine-tune the optimal solution initially found, thereby increasing the probability of finding the global optimal solution. The method of the present invention can handle multi-objective optimization problems through the combination of motor performance evaluation model, fitness function and two optimization algorithms, and can find the global optimal solution in a wide range of drive scheme spaces to find the drive scheme that achieves the best balance between multiple targets, and finally output the optimal motor drive scheme that performs well in multiple performance indicators, which is expected to provide better performance in practical applications and meet specific engineering needs.

The following is a further introduction to a motor intelligent drive optimization method according to an embodiment of the present invention:

Further optionally, in step S102, the step of constructing a motor performance evaluation model based on the optimization target includes:

Collecting the operation data and motor performance data of the motor under different driving schemes, wherein the operation data includes speed, current, voltage and temperature;

Extracting features directly related to the optimization objective from the operating data;

The extracted features and the corresponding motor performance data are used as a training set to train the deep learning model to obtain a motor performance evaluation model, which is used to predict the motor performance evaluation results under different driving schemes based on the input motor operation data.

Understandably, the operating data and motor performance data of the motor under different drive schemes are collected. The operating data include but are not limited to speed, current, voltage, temperature and other important operating state parameters of the motor to ensure that the collected data is extensive and representative and can cover the performance of the motor under various possible working conditions; from the operating data, extract the features directly related to the optimization target. For example, for efficiency optimization, you can focus on the ratio of current to voltage, power factor, etc., and select a suitable deep learning algorithm to train the model, such as a neural network; use the extracted features and the corresponding motor performance data as a training set, and train the model by adjusting the model parameters and learning algorithm so that it can accurately predict the performance of the motor. Using the trained model, input the motor operating data under the new drive scheme, and the model will output the predicted values of one or more performance indicators. Based on these predicted values, a comprehensive evaluation result can be calculated, which comprehensively considers multiple performance indicators, such as efficiency, energy consumption, response speed, etc. The comprehensive evaluation result can be defined as fitness.

In the process of optimizing the drive scheme through the evolutionary optimization algorithm and the annealing optimization algorithm, the fitness of the drive scheme is evaluated. For each drive scheme, the trained motor performance evaluation model can be used to predict the motor performance indicators, and then the weighted sum of the performance indicators is calculated through the fitness function to obtain a comprehensive evaluation result, that is, the fitness of the drive scheme.

Further optionally, in step S103, the step of combining the evaluation of the fitness of the driving scheme and optimizing the driving scheme by the evolutionary optimization algorithm and the annealing optimization algorithm to output the optimal motor driving scheme includes:

Combined with the evaluation of the fitness of the driving scheme, and through the selection, crossover and mutation operations of the evolutionary optimization algorithm, a global search is performed in the driving scheme space to find an initial optimal solution set that matches the optimization goal;

Based on the initial optimal solution set, combined with the evaluation of the fitness of the solution, and through the annealing optimization algorithm, a local search is performed in the neighborhood of the initial optimal solution to approximate the optimal solution and obtain the global optimal solution;

The global optimal solution is taken as the optimal motor driving solution.

Understandably, combined with the fitness evaluation of the drive scheme, the global search strategy of selecting, crossover and mutation operations through the evolutionary optimization algorithm is to widely explore the possible solution space and find the initial optimal solution set that matches the optimization goal. This characteristic of the evolutionary optimization algorithm enables it to jump out of the local optimum and find a wider range of optimization possibilities. Subsequently, starting from the initial optimal solution set, the annealing optimization algorithm is used to perform a local fine search in the neighborhood of the initial optimal solution. The annealing optimization algorithm simulates the physical annealing process, gradually lowers the "temperature", reduces the randomness in the search process, makes the search more focused on high-quality solutions, and gradually approaches the optimal solution, thereby obtaining the global optimal solution. Finally, this global optimal solution is determined as the optimal motor drive solution. The entire process combines the advantages of global search and local fine search, ensuring the comprehensiveness and depth of the optimization process. The final output of the motor drive solution achieves the best in comprehensive performance and meets the core requirements of motor design.

In the process of optimizing the motor drive scheme, the evaluation of the drive scheme fitness is an important means of performance evaluation, and it runs through the entire process of evolutionary optimization and annealing optimization. In the evolutionary optimization stage, for each individual corresponding to the drive scheme, the motor performance index is predicted using the trained motor performance evaluation model, and then these performance indexes are weighted and summed through the fitness function to obtain a comprehensive evaluation result, which represents the fitness of the individual. Similarly, in the annealing optimization stage, the solution corresponding to each drive scheme is processed similarly: first, the performance index is predicted using the trained motor performance evaluation model, and then the weighted sum is calculated through the fitness function to obtain the comprehensive evaluation result of the solution, that is, the fitness of the solution. Through the fitness evaluation, the advantages and disadvantages of each drive scheme can be comprehensively and objectively measured.

Further optionally, the step of combining the evaluation of the fitness of the driving scheme and performing a global search in the driving scheme space through the selection, crossover and mutation operations of the evolutionary optimization algorithm to find an initial optimal solution set matching the optimization objective includes:

Randomly generate a number of initial driving schemes, and encode the driving schemes into gene sequences as individuals of the initial population;

Evaluating the fitness value of each individual in the population by combining the motor performance evaluation model and the fitness function;

According to a preset screening ratio, excellent individuals with high fitness values in the population are screened out as the next generation of the population;

Randomly select two individuals from the population after the selection operation, perform a crossover operation, generate a crossover individual, and add the crossover individual to the population;

Performing a mutation operation on the individuals in the population after the crossover operation to generate mutant individuals, and adding the mutant individuals to the population;

Sort by fitness value and select individuals with high fitness value to form a new population;

Repeat the iteration until the termination condition of the evolutionary optimization algorithm is reached, then stop the iteration, and use the current population as the initial optimal solution set, and the individuals in the current population as the initial optimal solutions.

Understandably, in the process of finding the initial optimal solution set that matches the optimization target, first, several initial drive schemes are randomly generated, and these schemes are encoded as gene sequences as individuals of the initial population. Then, the fitness value of each individual is evaluated in combination with the motor performance evaluation model and the fitness function to measure the pros and cons of the drive scheme. According to the preset screening ratio, excellent individuals with high fitness values are selected from the population to ensure the quality of the next generation of populations. Subsequently, through the crossover operation, two individuals are randomly selected to cross the gene sequence, and new crossover individuals are generated and added to the population to increase the diversity of the population. At the same time, the individuals in the population are mutated, new genetic mutations are introduced, and mutant individuals are generated, and they are also added to the population, which helps to explore a wider search space and avoid falling into the local optimum. Then, the individuals in the population are sorted according to the fitness value, and the individuals with high fitness values are screened out to form a new population, so that the search direction continues to approach the global optimal solution. By repeated iterations, until the termination condition of the evolutionary optimization algorithm is reached, the iteration is stopped, and the current population is used as the initial optimal solution set, and the individuals therein are the initial optimal solutions. This method combines global search and step-by-step optimization strategies to find the driving scheme that best matches the optimization goal.

The fitness value of each individual in the population is evaluated in combination with the motor performance evaluation model and the fitness function. Specifically, for the individual corresponding to each drive scheme, the trained motor performance evaluation model is used to predict its motor performance index. Subsequently, these performance indexes are weighted and summed through the fitness function to obtain a comprehensive evaluation result, which represents the fitness of the individual.

In the process of searching for the optimal solution, when dealing with constrained optimization problems, some strategies can also be used to ensure that the generated solution meets the given constraints (such as energy consumption cannot exceed a certain value). For example, when initializing the population, in order to ensure that the generated initial solution meets the constraints, the initial solution can be generated by some specific heuristic methods or by combining random generation with verification. For individuals that do not meet the constraints, their fitness values can be reduced by adding penalty items to the fitness function. The size of the penalty item can be determined according to the degree of constraint violation. A constraint satisfaction index can also be defined to measure the degree to which an individual meets the constraints. This index can be used as an additional fitness criterion. If the individual does not meet the constraints after crossover or mutation, a repair strategy can be used to adjust its genes so that it meets the constraints.

Further optionally, the step of randomly selecting two individuals from the population after the selection operation, performing a crossover operation, generating crossover individuals, and adding the crossover individuals to the population includes:

Randomly select two individuals from the population after the selection operation as individuals to be crossed;

At the intersection point of the two individuals to be crossed, exchanging genes to create two crossover individuals, wherein the intersection point is one or more positions randomly selected in the gene sequence of the individuals to be crossed;

Determine whether the new crossover individual meets the constraints;

If satisfied, the crossover individual is added to the population.

Understandably, two individuals are randomly selected from the population that has undergone the selection operation as individuals to be crossed. Then, one or more positions are randomly selected in the gene sequence of the individuals to be crossed as crossover points, and some genes are exchanged at these crossover points to create two new crossover individuals. Then, the constraints of the new crossover individuals (such as energy consumption cannot exceed a certain value) are checked to ensure the effectiveness and feasibility of the new solution. If the constraints are not met, a repair strategy can be used to adjust its genes to meet the constraints. Once the crossover individual meets all the constraints, it will be added to the population for further optimization and selection in the subsequent evolution process. In this way, the solution space can be effectively explored and the algorithm can be promoted to evolve towards a better solution.

Further optionally, the step of performing a mutation operation on the individuals in the population after the crossover operation to generate mutant individuals and adding the mutant individuals to the population includes:

Using the individuals in the population after the crossover operation as individuals to be mutated;

Randomly selecting one or more mutation points in the gene sequence of the individual to be mutated;

Randomly adjusting the gene at the mutation point to generate a mutant individual;

Determine whether the variant individual satisfies the constraint condition;

If satisfied, the variant individual is added to the population.

Understandably, in the process of evolutionary optimization, mutation operations can help the algorithm jump out of the local optimum and explore a wider solution space. Specifically, the individuals in the population after the crossover operation are regarded as individuals to be mutated, and one or more points are randomly selected as mutation points in the gene sequences of these individuals. At these mutation points, the genes are randomly adjusted to generate new mutant individuals. This adjustment can be to add a small random perturbation to the original parameters, thereby introducing new genetic mutations. Then, the newly generated mutant individuals are strictly checked for constraints to ensure that they meet all the requirements of the problem. If the mutant individual does not meet the constraints, specific repair measures can be taken to make it meet the requirements. Only when the mutant individual fully meets the constraints is it added to the population. In this way, the mutation operation not only increases the diversity of the population, but also provides more possible search directions for the algorithm, promoting the search for the global optimal solution.

Further optionally, the step of taking the initial optimal solution set as a basis, combining the evaluation of the fitness of the solution, and performing a local search in the neighborhood of the initial optimal solution through the annealing optimization algorithm to approximate the optimal solution to obtain the global optimal solution includes:

Using the initial optimal solution in the initial optimal solution set as the initial solution of the annealing optimization algorithm;

Setting an initial temperature value T ₀ higher than a preset threshold;

Randomly search for new solutions in the neighborhood of the current solution;

Decide whether to accept the new solution based on the Metropolis criteria;

The temperature is gradually lowered according to the preset cooling strategy, and the neighborhood search and acceptance criterion judgment process are repeated until the termination condition is met, and the current solution is used as the candidate set of the global optimal solution.

It can be understood that in the optimization process, the initial optimal solution set is used as the starting point, and the annealing optimization algorithm is used for further fine search. Specifically, the initial optimal solution set is used as the initial solution of the annealing optimization algorithm, and an initial temperature value T ₀ higher than the preset threshold is set to ensure that the algorithm has sufficient flexibility in the early stage of the search and can jump out of the local optimal solution. Next, a new solution is randomly searched in the neighborhood of the current solution to explore possible better solutions around the current solution. Then, the Metropolis criterion is used to determine whether to accept the new solution. This criterion allows the algorithm to accept a poor solution with a certain probability during the search process, thereby avoiding falling into the local optimum. As the search proceeds, the temperature is gradually lowered according to the preset cooling strategy, which means that the algorithm gradually reduces the probability of accepting poor solutions, making the search more focused on high-quality solutions. By continuously repeating the judgment process of the neighborhood search and the acceptance criterion until the termination condition is met, the current solution is used as a candidate set for the global optimal solution. This method combines the advantages of global search and local fine search, and can further approach the optimal solution on the basis of the initial optimal solution, thereby improving the accuracy and efficiency of optimization.

Further optionally, the step of judging whether to accept the new solution according to the Metropolis criterion includes:

The fitness values of the current solution and the new solution are evaluated respectively in combination with the motor performance evaluation model and the fitness function;

Calculate the difference Δf between the fitness value of the new solution and the current solution, the formula is: , where f(x′) is the fitness value of the new solution x′, and f(x) is the fitness value of the current solution x;

like , then accept the new solution and update the current solution so that and ;

like , then the preset acceptance probability is used to decide whether to accept the new solution, and the preset acceptance probability is equal to , where T is the current temperature.

Understandably, the motor performance evaluation model and fitness function are used to evaluate the fitness values of the current solution x and the new solution x′, and the fitness value difference Δf between the new solution and the current solution is calculated. If Δf is less than 0, it means that the new solution is better than the current solution, then the new solution is accepted unconditionally, and the current solution and the corresponding fitness value are updated. If Δf is greater than or equal to 0, that is, the new solution is not better than the current solution, then the decision on whether to accept the new solution is made based on the preset acceptance probability. This method allows the algorithm to jump out of the local optimum with a certain probability during the search process, increasing the diversity of the search, and thus making it more likely to find the global optimal solution.

Further optionally, the step of gradually lowering the temperature according to a preset cooling strategy includes:

Update the temperature according to the temperature drop rate. The formula is: , where α is the temperature drop rate, T′ is the updated temperature, and T is the temperature before the update.

Understandably, the temperature is updated according to the temperature drop rate. α represents the temperature drop rate, which is a value between 0 and 1 and is used to control the temperature drop rate. By continuously lowering the temperature according to this cooling strategy, the algorithm can gradually reduce the probability of accepting poor solutions, making the search process more focused on the high-quality solution area, thus helping to find the global optimal solution. This cooling method ensures both the breadth and accuracy of the search.

Embodiment 2: Referring to FIG. 2 , the present invention further proposes a motor intelligent drive optimization system, the system comprising:

Optimization target module: used to determine the optimization target of the motor drive and quantify the optimization target;

Performance evaluation module: used to construct a motor performance evaluation model based on the optimization target, and define a fitness function closely related to the optimization target according to the output of the motor performance evaluation model to evaluate the pros and cons of each driving scheme;

Drive scheme optimization module: used to combine the evaluation of the fitness of the drive scheme, and optimize the drive scheme through evolutionary optimization algorithm and annealing optimization algorithm to output the optimal motor drive scheme.

Further optionally, the performance evaluation module is also used to:

Collecting the operation data and motor performance data of the motor under different driving schemes, wherein the operation data includes speed, current, voltage and temperature;

Extracting features directly related to the optimization objective from the operating data;

The extracted features and the corresponding motor performance data are used as a training set to train the deep learning model to obtain a motor performance evaluation model, which is used to predict the motor performance evaluation results under different driving schemes based on the input motor operation data.

Further optionally, the performance evaluation module is also used to:

The fitness function is used to weight and calculate various performance indicators of the motor performance evaluation result.

Further optionally, the drive scheme optimization module is also used for:

Combined with the evaluation of the fitness of the driving scheme, and through the selection, crossover and mutation operations of the evolutionary optimization algorithm, a global search is performed in the driving scheme space to find an initial optimal solution set that matches the optimization goal;

Based on the initial optimal solution set, combined with the evaluation of the fitness of the solution, and through the annealing optimization algorithm, a local search is performed in the neighborhood of the initial optimal solution to approximate the optimal solution and obtain the global optimal solution;

The global optimal solution is taken as the optimal motor driving solution.

Further optionally, the drive scheme optimization module is also used for:

Randomly generate a number of initial driving schemes, and encode the driving schemes into gene sequences as individuals of the initial population;

Evaluating the fitness value of each individual in the population by combining the motor performance evaluation model and the fitness function;

According to a preset screening ratio, excellent individuals with high fitness values in the population are screened out as the next generation of the population;

Randomly select two individuals from the population after the selection operation, perform a crossover operation, generate a crossover individual, and add the crossover individual to the population;

Performing a mutation operation on the individuals in the population after the crossover operation to generate mutant individuals, and adding the mutant individuals to the population;

Sort by fitness value and select individuals with high fitness value to form a new population;

Repeat the iteration until the termination condition of the evolutionary optimization algorithm is reached, then stop the iteration, and use the current population as the initial optimal solution set, and the individuals in the current population as the initial optimal solutions.

Further optionally, the drive scheme optimization module is also used for:

Randomly select two individuals from the population after the selection operation as individuals to be crossed;

At the intersection point of the two individuals to be crossed, exchanging genes to create two crossover individuals, wherein the intersection point is one or more positions randomly selected in the gene sequence of the individuals to be crossed;

Determine whether the new crossover individual meets the constraints;

If satisfied, the crossover individual is added to the population.

Further optionally, the drive scheme optimization module is also used for:

Using the individuals in the population after the crossover operation as individuals to be mutated;

Randomly selecting one or more mutation points in the gene sequence of the individual to be mutated;

Randomly adjusting the gene at the mutation point to generate a mutant individual;

Determine whether the variant individual satisfies the constraint condition;

If satisfied, the variant individual is added to the population.

Further optionally, the drive scheme optimization module is also used for:

Using the initial optimal solution in the initial optimal solution set as the initial solution of the annealing optimization algorithm;

Setting an initial temperature value T ₀ higher than a preset threshold;

Randomly search for new solutions in the neighborhood of the current solution;

Decide whether to accept the new solution based on the Metropolis criteria;

The temperature is gradually lowered according to the preset cooling strategy, and the neighborhood search and acceptance criterion judgment process are repeated until the termination condition is met, and the current solution is used as the candidate set of the global optimal solution.

Further optionally, the drive scheme optimization module is also used for:

The fitness values of the current solution and the new solution are evaluated respectively in combination with the motor performance evaluation model and the fitness function;

Calculate the difference Δf between the fitness value of the new solution and the current solution, the formula is: , where f(x′) is the fitness value of the new solution x′, and f(x) is the fitness value of the current solution x;

like , then accept the new solution and update the current solution so that and ;

like , then the preset acceptance probability is used to decide whether to accept the new solution, and the preset acceptance probability is equal to , where T is the current temperature.

The above-mentioned embodiments only express several implementation methods of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the patent of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention. Therefore, the protection scope of the patent of the present invention shall be subject to the attached claims.

Claims (9)

1. A motor intelligent drive optimization method, characterized in that the method comprises: Determining an optimization target for the motor drive and quantifying the optimization target; Building a motor performance evaluation model based on the optimization target, and defining a fitness function closely related to the optimization target according to the output of the motor performance evaluation model to evaluate the pros and cons of each driving scheme; Combined with the evaluation of the fitness of the drive scheme, the drive scheme is optimized through the evolutionary optimization algorithm and the annealing optimization algorithm to output the optimal motor drive scheme; The step of combining the evaluation of the fitness of the driving scheme and optimizing the driving scheme by using an evolutionary optimization algorithm and an annealing optimization algorithm to output the optimal motor driving scheme includes: Combined with the evaluation of the fitness of the driving scheme, and through the selection, crossover and mutation operations of the evolutionary optimization algorithm, a global search is performed in the driving scheme space to find an initial optimal solution set that matches the optimization goal; Based on the initial optimal solution set, combined with the evaluation of the fitness of the solution, and through the annealing optimization algorithm, a local search is performed in the neighborhood of the initial optimal solution to approximate the optimal solution and obtain the global optimal solution; The global optimal solution is taken as the optimal motor driving solution. 2. The motor intelligent drive optimization method according to claim 1, characterized in that the step of constructing a motor performance evaluation model based on the optimization target comprises: Collecting the operation data and motor performance data of the motor under different driving schemes, wherein the operation data includes speed, current, voltage and temperature; Extracting features directly related to the optimization objective from the operating data; The extracted features and the corresponding motor performance data are used as a training set to train the deep learning model to obtain a motor performance evaluation model, which is used to predict the motor performance evaluation results under different driving schemes based on the input motor operation data. 3. The motor intelligent drive optimization method according to claim 2, characterized in that in the step of defining a fitness function closely related to the optimization target according to the output of the motor performance evaluation model: The fitness function is used to weight and calculate various performance indicators of the motor performance evaluation result. 4. The motor intelligent drive optimization method according to claim 1 is characterized in that the step of combining the evaluation of the fitness of the drive scheme and performing a global search in the drive scheme space through the selection, crossover and mutation operations of the evolutionary optimization algorithm to find an initial optimal solution set matching the optimization target comprises: Randomly generate a number of initial driving schemes, and encode the driving schemes into gene sequences as individuals of the initial population; Evaluate the fitness value of each individual in the population by combining the motor performance evaluation model and the fitness function; According to a preset screening ratio, excellent individuals with high fitness values in the population are screened out as the next generation of the population; Randomly select two individuals from the population after the selection operation, perform a crossover operation, generate a crossover individual, and add the crossover individual to the population; Performing a mutation operation on the individuals in the population after the crossover operation to generate mutant individuals, and adding the mutant individuals to the population; Sort by fitness value and select individuals with high fitness value to form a new population; Repeat the iteration until the termination condition of the evolutionary optimization algorithm is reached, then stop the iteration, and use the current population as the initial optimal solution set, and the individuals in the current population as the initial optimal solutions. 5. The motor intelligent drive optimization method according to claim 4, characterized in that the step of randomly selecting two individuals from the population after the selection operation, performing a crossover operation, generating a crossover individual, and adding the crossover individuals to the population comprises: Randomly select two individuals from the population after the selection operation as individuals to be crossed; At the intersection point of the two individuals to be crossed, exchanging genes to create two crossover individuals, wherein the intersection point is one or more positions randomly selected in the gene sequence of the individuals to be crossed; Determine whether the new crossover individual meets the constraints; If satisfied, the crossover individual is added to the population. 6. The motor intelligent drive optimization method according to claim 4, characterized in that the step of performing a mutation operation on the individuals in the population after the crossover operation to generate a mutant individual and adding the mutant individual to the population comprises: Using the individuals in the population after the crossover operation as individuals to be mutated; Randomly selecting one or more mutation points in the gene sequence of the individual to be mutated; Randomly adjusting the gene at the mutation point to generate a mutant individual; Determine whether the variant individual satisfies the constraint condition; If satisfied, the variant individual is added to the population. 7. The motor intelligent drive optimization method according to claim 1 is characterized in that the step of taking the initial optimal solution set as a basis, combining the evaluation of the fitness of the solution, and performing a local search in the neighborhood of the initial optimal solution through the annealing optimization algorithm to approximate the optimal solution to obtain the global optimal solution comprises: Using the initial optimal solution in the initial optimal solution set as the initial solution of the annealing optimization algorithm; Setting an initial temperature value T ₀ higher than a preset threshold; Randomly search for new solutions in the neighborhood of the current solution; Decide whether to accept the new solution based on the Metropolis criteria; The temperature is gradually lowered according to the preset cooling strategy, and the neighborhood search and acceptance criterion judgment process are repeated until the termination condition is met, and the current solution is used as the candidate set of the global optimal solution. 8. The motor intelligent drive optimization method according to claim 7, characterized in that the step of judging whether to accept the new solution according to the Metropolis criterion comprises: The fitness values of the current solution and the new solution are evaluated respectively in combination with the motor performance evaluation model and the fitness function; Calculate the difference Δf between the fitness value of the new solution and the current solution, the formula is: , where f(x′) is the fitness value of the new solution x′, and f(x) is the fitness value of the current solution x; like , then accept the new solution and update the current solution so that and ; like , then the preset acceptance probability is used to decide whether to accept the new solution, and the preset acceptance probability is equal to , where T is the current temperature. 9. A motor intelligent drive optimization system, characterized by comprising: Optimization target module: used to determine the optimization target of the motor drive and quantify the optimization target; Performance evaluation module: used to construct a motor performance evaluation model based on the optimization target, and define a fitness function closely related to the optimization target according to the output of the motor performance evaluation model to evaluate the pros and cons of each driving scheme; Drive scheme optimization module: used to combine the fitness evaluation of the drive scheme and optimize the drive scheme through the evolutionary optimization algorithm and the annealing optimization algorithm to output the optimal motor drive scheme; Wherein, the driving scheme optimization module is also used for: Combined with the evaluation of the fitness of the driving scheme, and through the selection, crossover and mutation operations of the evolutionary optimization algorithm, a global search is performed in the driving scheme space to find an initial optimal solution set that matches the optimization goal; Based on the initial optimal solution set, combined with the evaluation of the fitness of the solution, and through the annealing optimization algorithm, a local search is performed in the neighborhood of the initial optimal solution to approximate the optimal solution and obtain the global optimal solution; The global optimal solution is taken as the optimal motor driving solution.