Method, device and equipment for selecting aircraft power system based on multi-objective optimization

Abstract

The present application provides a method, device and equipment for selecting an aircraft power system based on multi-objective optimization. The method provided by the present application includes: constructing a target optimization problem with the goal of maximizing global fitness, with the thrust of the power system meeting the preset minimum flight requirement as the first constraint, the total weight of the power system cannot exceed the structural upper limit as the second constraint, and the total cost of the power system cannot be lower than the preset maximum economic cost as the third constraint; constructing a fitness function for the target optimization problem; according to the fitness function, with the goal of maximizing global fitness, solving the optimal solution of the target optimization problem, and determining the solved optimal solution as the selected power system; continuous variables update positions based on a particle swarm optimization algorithm. The aircraft power system selection method based on multi-objective optimization provided by the present application is used to automatically select the most suitable power system in the early stage of aircraft design to achieve intelligent matching of the power system.

Description

Technical Field

[0001] The present application relates to the technical field of aircraft power systems, and in particular to a method, device and equipment for selecting an aircraft power system based on multi-objective optimization. Background Art

[0002] With the rapid development of aviation technology, especially the widespread application of small aircraft such as drones and light aircraft, the design and optimization of aircraft power systems have become key factors in improving aircraft performance, reducing operating costs and reducing environmental impact. In the current aviation market, the demand for more efficient, more environmentally friendly and more economical power systems is growing, which has promoted the continuous innovation and progress of power system technology.

[0003] However, the selection of existing aircraft power systems is mainly based on the experience of designers, lacking a systematic and intelligent decision-making method, and not fully combining the multi-dimensional performance parameters of the aircraft for comprehensive analysis. Therefore, how to complete the selection of appropriate power systems to optimize the various performance indicators of the power system while meeting the flight mission requirements has become a hot topic in current research. Summary of the invention

[0004] In view of this, the present application provides an aircraft power system selection method, device and equipment based on multi-objective optimization, which is used to automatically select the most suitable power system in the early stage of aircraft design and realize intelligent matching of the power system.

[0005] Specifically, the present application is implemented through the following technical solutions:

[0006] In a first aspect, the present application provides a method for selecting an aircraft power system based on multi-objective optimization, the method comprising:

[0007] A target optimization problem is constructed with the goal of maximizing global fitness, with the thrust of the power system meeting the preset minimum flight requirement as the first constraint, the total weight of the power system cannot exceed the structural upper limit as the second constraint, and the total cost of the power system cannot be lower than the preset maximum economic cost as the third constraint; wherein the target optimization problem maps the configuration parameters of the power system to the search space of the particle swarm, each particle includes a set of configuration parameters of a power system, and the position of each particle is composed of a discrete variable for characterizing the type of the power system and a continuous variable for characterizing the configuration parameters of the power system;

[0008] Constructing a fitness function for the target optimization problem; wherein the fitness function is a fitness function that comprehensively considers range, output power, weight, reliability and cost;

[0009] According to the fitness function, with the goal of maximizing the global fitness, solving the optimal solution of the target optimization problem, and determining the solved optimal solution as the selected power system;

[0010] Wherein, in the process of solving the optimal solution of the target optimization problem, in each iteration, the discrete variable updates the position based on Bayesian reasoning; and the continuous variable updates the position based on the particle swarm optimization algorithm.

[0011] The second aspect of the present application provides an aircraft power system selection device based on multi-objective optimization, the device comprising a processing module, a building module and a computing module; wherein:

[0012] The processing module is used to construct a target optimization problem with the goal of maximizing global fitness, with the thrust of the power system meeting the preset minimum flight requirement as the first constraint condition, the total weight of the power system cannot exceed the structural upper limit as the second constraint condition, and the total cost of the power system cannot be lower than the preset maximum economic cost as the third constraint condition; wherein the target optimization problem maps the configuration parameters of the power system to the search space of the particle swarm, each particle includes a set of configuration parameters of a power system, and the position of each particle is composed of a discrete variable for characterizing the type of the power system and a continuous variable for characterizing the configuration parameters of the power system;

[0013] The construction module is used to construct a fitness function of the target optimization problem; wherein the fitness function is a fitness function that comprehensively considers range, output power, weight, reliability and cost;

[0014] The calculation module is used to solve the optimal solution of the target optimization problem according to the fitness function with the goal of maximizing the global fitness, and determine the solved optimal solution as the selected power system;

[0015] Wherein, in the process of solving the optimal solution of the target optimization problem, in each iteration, the discrete variable updates the position based on Bayesian reasoning; and the continuous variable updates the position based on the particle swarm optimization algorithm.

[0016] The aircraft power system selection method, device and equipment provided in the present application based on multi-objective optimization construct a target optimization problem by taking maximizing global fitness as the goal, taking the thrust of the power system meeting the preset minimum flight requirement as the first constraint, the total weight of the power system cannot exceed the structural upper limit as the second constraint, and the total cost of the power system cannot be lower than the preset maximum economic cost as the third constraint, and then constructing a fitness function for the target optimization problem, so as to solve the optimal solution of the target optimization problem based on the fitness function with the goal of maximizing the global fitness, and determine the solved optimal solution as the selected power system. In this way, on the first hand, by constructing the fitness function, multiple optimization goals such as range, output power, weight, reliability and cost are comprehensively considered to ensure that the selected power system achieves the optimization of overall performance while meeting the flight mission requirements. In other words, with the goal of maximizing global fitness, through the particle swarm optimization algorithm, the optimal solution is continuously approached and optimized in the iterative process. Find the optimal solution and ensure that the selected power system is the global optimal rather than the local optimal; secondly, by automatically generating the initial particle swarm, updating the power system type and configuration parameters and other steps, the intelligent selection of the power system is realized. Through the automated selection and optimization process, not only the time cycle of power system selection is greatly shortened and the design efficiency is improved, but also the accuracy and reliability of the selected power system can be ensured, and the design risk can be reduced; thirdly, in each iteration process, the discrete variables used to characterize the power system type are updated based on Bayesian reasoning, and the continuous variables used to characterize the configuration parameters are updated based on the particle swarm optimization algorithm. The combination of particle swarm optimization algorithm and Bayesian reasoning can find the optimal power system and configuration parameters under complex constraints, greatly improving the search efficiency and accuracy, and providing effective decision support for the design of aircraft power systems; fourthly, this method is applicable to various types of aircraft and has good scalability and applicability. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flowchart of a first embodiment of a method for selecting an aircraft power system based on multi-objective optimization provided in this application;

[0018] Figure 2 A hardware structure diagram of an aircraft power system selection device based on multi-objective optimization in which the aircraft power system selection device based on multi-objective optimization of the present application is located;

[0019] Figure 3 This is a structural schematic diagram of the first embodiment of the aircraft power system selection device based on multi-objective optimization provided in this application. DETAILED DESCRIPTION

[0020] Here, exemplary embodiments are described in detail, and examples thereof are shown in the accompanying drawings. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The implementations described in the following exemplary embodiments do not represent all implementations consistent with the present application.

[0021] The terms used in this application are only for the purpose of describing specific embodiments and are not intended to limit this application. The singular forms of "a", "said" and "the" used in this application are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in this article refers to and includes any or all possible combinations of one or more associated listed items.

[0022] It should be understood that although the terms first, second, third, etc. may be used in the present application to describe various information, these information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0023] Specific embodiments are given below to introduce the technical solution of the present application in detail.

[0024] Figure 1 This is a flowchart of the first embodiment of the aircraft power system selection method based on multi-objective optimization provided by this application. Figure 1 , the method provided in this embodiment may include:

[0025] S101. Construct a target optimization problem with the goal of maximizing global fitness, the thrust of the power system meeting the preset minimum flight requirement as the first constraint, the total weight of the power system not exceeding the structural upper limit as the second constraint, and the total cost of the power system not being lower than the preset maximum economic cost as the third constraint.

[0026] The target optimization problem maps the configuration parameters of the power system to the search space of the particle swarm, each particle includes a set of configuration parameters of a power system, and the position of each particle is composed of a discrete variable for characterizing the type of the power system and a continuous variable for characterizing the configuration parameters of the power system.

[0027] Specifically, when selecting an aircraft power system, since the aircraft power system involves multiple types such as turboshaft-electric hybrid power system, piston-electric hybrid power system and pure electric power, each type can be set with different configuration parameters. Therefore, when selecting the power system, it is necessary to select the best power system and the best configuration parameters to ensure that the aircraft can meet the requirements of the flight mission to the greatest extent.

[0028] In this application, by comprehensively optimizing the multi-dimensional configuration parameters of different types of power systems (turboshaft-electric hybrid power system, piston-electric hybrid power system, pure electric power system) to meet the performance requirements of complex flight missions, the most suitable power system can be automatically selected in the early stage of aircraft design to achieve intelligent matching of the power system.

[0029] Furthermore, in order to select the optimal power system, in this application, the problem of selecting the optimal power system is transformed into a solution process of the target optimization problem, the configuration parameters of the power system are mapped to the search space of the particle swarm, and an improved particle swarm optimization algorithm is used. It is possible to achieve the optimal design of the power system parameters while satisfying multiple constraints.

[0030] Specifically, a target optimization problem is constructed with the goal of maximizing global fitness, with the thrust of the power system meeting the preset minimum flight requirement as the first constraint, the total weight of the power system cannot exceed the structural upper limit as the second constraint, and the total cost of the power system cannot be lower than the preset maximum economic cost as the third constraint. By solving the optimal solution of the target optimization problem, the optimal power system and the corresponding configuration parameters can be selected.

[0031] It should be noted that global fitness is a comprehensive indicator used to evaluate the pros and cons of all possible solutions. It can be understood that by solving the maximum value of global fitness, the optimal power system selection can be solved.

[0032] Furthermore, the constraints of the target optimization problem are as follows:

[0033] The first constraint is that the thrust of the power system meets the preset minimum flight requirements; the second constraint is that the total weight of the power system cannot exceed the structural upper limit; the third constraint is that the total cost of the power system cannot be lower than the preset maximum economic cost.

[0034] It should be noted that the preset minimum flight requirements are pre-set according to the actual model of the aircraft, and the preset minimum flight requirements are used to ensure that the aircraft can fly normally.

[0035] Furthermore, the structural upper limit is the maximum weight limit that the aircraft structure design can bear. For example, in one embodiment, the structural upper limit is characterized by weight to ensure the structural safety and performance stability of the aircraft. In addition, the preset maximum economic cost is also set according to actual needs, and in this embodiment, it is not limited.

[0036] It should be noted that the first constraint is used to ensure that the aircraft can perform basic flight missions; the second constraint is used to ensure the structural safety and stable performance of the aircraft; and the third constraint is used to ensure the economic feasibility and profitability of the project.

[0037] Furthermore, each particle represents a possible solution, and its position represents a set of configuration parameters of a dynamical system, which updates its position and velocity based on its own experience and the experience of the group.

[0038] Referring to the above description, it can be understood that the position of each particle is composed of two parts: the type of the power system and the configuration parameters of the power system. The type of the power system is a certain power system among the preset multiple power system types; further, the configuration parameters of the power system are parameters used to describe the characteristics and performance of the power system, such as range, power system weight, fuel efficiency, cost, output power, reliability, etc.

[0039] Optionally, in one possible implementation, the set of configuration parameters includes the following parameters: flight speed, battery energy density, average power consumption of the aircraft, battery weight, power system weight, total fuel weight, output power, reliability and cost.

[0040] It should be noted that the preset multiple power types are set according to actual needs and are not limited in this embodiment. For example, in one embodiment, the preset multiple power systems may include a turbine engine, a piston engine, a turboshaft-electric hybrid system, a piston-electric hybrid system, or a pure electric power system.

[0041] It can be understood that, for a particle, the type of the power system therein is selected from a plurality of preset power systems, which is a discrete variable, and the configuration parameter thereof is a continuous variable.

[0042] In other words, each particle carries a discrete variable and a set of continuous variables, one discrete variable corresponds to a dynamic system, and a set of continuous variables corresponds to a set of configuration parameters.

[0043] Furthermore, the value set of the discrete variable xi is {1, 2, ..., N}, where N is the total number of preset multiple power system types. For example, the preset multiple power types include five types: A, B, C, D, and E, that is, N is 5, when it is 1, it means that power system A is selected, and when it is 2, it means that power system B is selected. Further, combined with the above example, for example, a particle represents power system A and a set of configuration parameters A1 corresponding to power system A, and another particle represents power system A and a set of configuration parameters A2 corresponding to power system A.

[0044] Combining the above example, for example, the position of a particle is given by It indicates that, Contains discrete variables, Contains continuous variables. It should be noted that discrete variables represent the type of power system, which is one-dimensional. Furthermore, continuous variables represent a set of configuration parameters, and their dimension is consistent with the dimension of the configuration parameters.

[0045] S102, constructing a fitness function for the target optimization problem; wherein the fitness function is a fitness function that comprehensively considers range, output power, weight, reliability and cost.

[0046] Specifically, the fitness function is a function used to evaluate the quality of each solution (i.e., particle). The value of the fitness function determines the solution's ability to survive and the chance of being selected during the optimization process.

[0047] In specific implementation, according to the mission requirements of the aircraft and the performance characteristics of the power system, the objectives to be optimized are determined to include range, output power, weight, reliability and cost, and then a fitness function is constructed by comprehensively considering multiple factors such as range, output power, weight, reliability and cost. For example, in one embodiment, the weighted sum or weighted product of range, output power, weight, reliability and cost can be determined as the fitness function, and the fitness value of different particles can be calculated through the fitness function to characterize the comprehensive performance of the power system and configuration parameters corresponding to the particle in multi-dimensional performance indicators.

[0048] A possible implementation is given below to introduce the constructed fitness function in detail:

[0049] Optionally, in a possible implementation, the fitness function is:

[0050] ;

[0051] Among them, the , , , are respectively preset weight coefficients; To maximize the range, the Calculated according to the first formula; To minimize weight, the Calculated according to the second formula; For maximizing output power, the Calculated according to the third formula; To maximize reliability, the Calculated according to the fourth formula; To minimize the cost, the Calculate according to the fifth formula.

[0052] It should be noted that the specific value of the preset weight coefficient is dynamically allocated according to task requirements, and .

[0053] The dimensions and numerical ranges of different optimization objectives are usually different. In order to eliminate the influence of the dimension, each optimization objective needs to be normalized so that the value range of all optimization objectives is between 0 and 1, making the fitness function after weighted combination more physically meaningful.

[0054] Optionally, in a possible implementation manner, the first formula is:

[0055] ;

[0056] ;

[0057] Wherein, R is the range; is the minimum range of the search space; is the maximum range of the search space; v is the flight speed; is the total fuel weight; W is the fuel consumption rate per unit time; is the weight of the battery; is the energy density of the battery; is the average power consumption of the aircraft.

[0058] Furthermore, the weight of the power system directly affects the take-off weight and fuel consumption of the aircraft, which is usually obtained through design calculations. In other words, the second formula is:

[0059] ;

[0060] Where W is the weight value of the power system generated by the particle swarm, is the maximum weight of the search space; is the minimum weight of the search space.

[0061] Furthermore, the weight of the powertrain It can be calculated by the following formula:

[0062] ;

[0063] Where W is the fuel consumption rate per unit time, is the engine weight, is the total fuel weight, is the weight of the battery.

[0064] Furthermore, output power is an important performance indicator of the power system, which determines the thrust and speed of the aircraft. Among them, output power is a maximization target, and the third formula is:

[0065] ;

[0066] Wherein, P is the output power generated by the particle swarm; is the maximum power of the search space; is the minimum power of the search space.

[0067] Furthermore, reliability measures the failure probability of the power system within the design life cycle, usually calculated based on a statistical model. In addition, reliability is a maximization goal, and the fourth formula is:

[0068] ;

[0069] Wherein, S is the system reliability of the aircraft; is the minimum failure probability of the power system within its design life cycle; is the maximum failure probability of the power system during its design life cycle;

[0070] Similarly, the fifth formula is:

[0071] ;

[0072] Among them, the is the maximum cost of the power system; is the minimum cost of the power system.

[0073] It should be noted that the fitness function takes into account multiple factors such as range, weight, output power, reliability and cost. In this way, the best balance point can be found based on multiple factors to avoid performance imbalance problems that may be caused by a single factor. The fitness function that comprehensively considers range, output power, weight, reliability and cost can achieve the best balance between performance, weight, reliability and cost in the design and optimization of the power system, thereby improving the overall efficiency and competitiveness of the system.

[0074] S103, according to the fitness function, with the goal of maximizing the global fitness, solving the optimal solution of the target optimization problem, and determining the solved optimal solution as the selected power system.

[0075] Wherein, in the process of solving the optimal solution of the target optimization problem, in each iteration, the discrete variable updates the position based on Bayesian reasoning; and the continuous variable updates the position based on the particle swarm optimization algorithm.

[0076] Specifically, when the value of the fitness function is the largest, the optimal solution obtained is the selected power system.

[0077] It should be noted that when the fitness function is used to maximize the global fitness calculation, the particle swarm needs to be continuously iterated, and the global fitness is recalculated after each iteration. After multiple iterations, the optimal solution obtained is the final selected power system and configuration parameters.

[0078] Furthermore, in the process of solving the optimal solution of the target optimization problem, in each iteration, the discrete variables update positions based on Bayesian reasoning; and the continuous variables update positions based on a particle swarm optimization algorithm.

[0079] It should be noted that Bayesian reasoning is a reasoning method based on probability theory, which is used to update the estimate of the probability of an event occurring under given observation data. In specific implementation, the probability distribution of discrete variables is updated based on Bayesian reasoning.

[0080] Furthermore, the particle swarm optimization algorithm finds the optimal solution through the interaction and information sharing between particles in the group. In specific implementation, the position of the continuous variable is updated based on the particle swarm optimization algorithm.

[0081] A specific embodiment is given below to describe in detail the process of solving the optimal solution of the target optimization problem. Specifically, in a possible implementation, solving the optimal solution of the target optimization problem according to the fitness function with the goal of maximizing the global fitness includes:

[0082] (1) A chaos initialization strategy is adopted to obtain an initial particle swarm, and particles that do not meet the constraints are screened out according to the first constraint condition, the second constraint condition and the third constraint condition.

[0083] Specifically, the chaos initialization strategy is used to generate the initial particle swarm. The chaos initialization strategy uses the randomness and ergodicity of the chaotic system to generate initial solutions with uniform distribution and obvious differences, which helps to enhance the search ability and efficiency of the algorithm.

[0084] In the specific implementation, a series of random and uniformly distributed initial particles are randomly generated based on the chaotic initialization strategy. The first, second and third constraints are determined according to the thrust requirements, total weight requirements and total cost requirements of the aircraft. Each particle in the initial particle group is evaluated to check whether it meets all the defined constraints. If the particle does not meet any constraint, it is removed from the particle group. In this way, it is ensured that the remaining particles meet the requirements of practical applications and reduce invalid searches and waste of computing resources.

[0085] (2) For each particle, based on the position of the particle and the fitness function, determine the fitness value of the particle, and according to the fitness value of each particle, determine the current global optimal solution and the local optimal solution of each particle.

[0086] In the specific implementation, the position information of each particle (including discrete variables used to characterize the power system and continuous variables used to characterize the configuration parameters) is input into the fitness function, and the fitness value corresponding to the particle is calculated.

[0087] It can be understood that the fitness value corresponding to a particle reflects the comprehensive performance of the power system configuration represented by the particle in multiple performance indicators.

[0088] Furthermore, for all particles, all fitness values ​​calculated in the current iteration are obtained, and the largest fitness value is determined as the current global optimal solution.

[0089] Furthermore, for a particle, all fitness values ​​calculated for the particle in each iteration are obtained, and the largest fitness value is determined as the current local optimal solution of the particle.

[0090] (3) According to the current number of iterations, determine the random perturbation factor in the current iteration process.

[0091] Specifically, the current number of iterations refers to the number of complete cycles that the algorithm has completed from the beginning of execution to the current moment. In each iteration, the algorithm searches for new solutions based on the current solution space and tries to improve the quality of the solution.

[0092] Furthermore, the random perturbation factor is used to increase the diversity of the search and avoid premature convergence. The random perturbation factor is a random number or random variable used to make small random adjustments to the solution during the algorithm iteration process.

[0093] In the specific implementation, in each iteration process, a random perturbation factor is determined according to the current number of iterations and the preset random perturbation strategy. The random perturbation factor is a random number that varies within a specific range. Its size and distribution can be adjusted according to the requirements of the algorithm and the characteristics of the search space. In this way, the random perturbation factor is used to increase the diversity and robustness of the search, making it more likely to find the global optimal solution.

[0094] A specific embodiment is given below to introduce the process of determining the random perturbation factor in detail. Optionally, in a possible implementation process, determining the random perturbation factor in the current iteration process according to the current iteration number includes:

[0095] Step 1: Determine the current exploration stage based on the current number of iterations and the preset maximum number of iterations; Step 2: Determine the random perturbation standard deviation corresponding to the exploration stage based on the current exploration stage; Step 3: Take 0 as the mean of the Gaussian distribution and the random perturbation standard deviation as the standard deviation of the Gaussian distribution, randomly generate a random number based on the Gaussian distribution, and determine the random number as the random perturbation factor in the current iteration process.

[0096] Specifically, the maximum number of iterations represents the number of iteration terminations. In specific implementation, different exploration stages can be determined by a preset maximum number of iterations. For example, in one embodiment, the exploration stage is divided into two stages, the initial stage and the later stage, and the maximum number of iterations is evenly divided. When the current number of iterations is less than or equal to half of the maximum number of iterations, it is determined that the current stage is the initial exploration stage; when the current number of iterations is greater than half of the maximum number of iterations, it is determined that the current stage is the later exploration stage.

[0097] In specific implementation, different exploration stages may have different random perturbation standard deviations. For example, in one embodiment, in the initial exploration stage, the random perturbation standard deviation is a fixed value σ1, and in the later exploration stage, the random perturbation standard deviation is a fixed value σ2. For another example, in another embodiment, in the initial exploration stage, the random perturbation standard deviation is the current number of iterations multiplied by parameter a, and in the later exploration stage, the random perturbation standard deviation is the current number of iterations multiplied by parameter b.

[0098] It can be understood that the standard deviation σ of the disturbance can take a larger value when the number of iterations is small (early exploration stage) to enhance the search range; when the number of iterations is large (later convergence stage), it can take a smaller value to improve the convergence accuracy.

[0099] In the specific implementation, the mean of the Gaussian distribution is set to 0, and a random number that conforms to the Gaussian distribution is randomly generated based on the standard deviation of the random perturbation. This random number is used as the random perturbation factor in the current iteration process, that is, the random perturbation self-simulation .

[0100] It should be noted that by comparing the current number of iterations with the preset maximum number of iterations, the current exploration stage can be determined. This dynamic adjustment strategy enables the algorithm to search the solution space more extensively in the early stage, increasing the possibility of finding the global optimal solution; while in the later stage, it focuses more on fine-tuning near the searched better solution to improve the convergence accuracy of the solution. At the same time, by introducing a random perturbation factor, a certain degree of randomness can be maintained in the search process, which helps to discover new and better solutions, so as to maintain an appropriate search range and search accuracy at different stages.

[0101] (4) Determine the inertia weight in the current iteration process according to the current number of iterations, the preset initial inertia weight, the preset minimum inertia weight, and the preset maximum number of iterations.

[0102] Specifically, the current number of iterations is the number of iterations that the algorithm has completed during execution. The current number of iterations can be used to track the calculation process of the algorithm and serve as a basis for adjusting algorithm parameters.

[0103] Furthermore, the inertia weight is used to control the inertia of the particle in the search space. It should be noted that a larger inertia weight means that the particle is more inclined to maintain its current direction and speed, while a smaller inertia weight makes the particle more susceptible to the influence of other particles or the global optimal solution and change direction.

[0104] In specific implementation, the inertia weight in the current iteration process can be calculated based on the following formula:

[0105] ;

[0106] Among them, w is the inertia weight in the current iteration process, is the initial inertia weight, is the minimum inertia weight, is the current iteration number, is the maximum number of iterations.

[0107] (5) For each continuous variable in a particle, based on a particle swarm optimization algorithm, the velocity of the particle is updated according to the inertia weight, and the position of the particle is updated according to the updated velocity of the particle, the current position of the particle and the random perturbation factor.

[0108] In specific implementation, the continuous variables in the particles can include performance indicators such as range, power system weight, fuel efficiency, cost, output power, reliability, etc.

[0109] Furthermore, the particle velocity represents the movement trend of the particle in the search space. The particle velocity is updated according to the particle's historical optimal position, global optimal position and inertia weight.

[0110] In specific implementation, for example, in one embodiment, the velocity of the particle may be updated based on the following formula:

[0111] ;

[0112] in, is the updated velocity of the ith particle, w is the inertia weight, is the current speed of the ith particle, c1 and c2 are the preset learning factors, is the local optimal solution of the ith particle, is the current position of the ith particle, r1 and r2 are random numbers in the range of 0 to 1, is the current global optimal solution.

[0113] Furthermore, after obtaining the updated speed of the particle, the position of the particle can be updated according to the updated speed of the particle, the current position of the particle and the random perturbation factor. In specific implementation, the position of the particle is updated according to the sixth formula; wherein the sixth formula is:

[0114] ;

[0115] in, is the updated position of the i-th particle; is the current position of the i-th particle; is the updated velocity of the ith particle, and ε is the random perturbation factor.

[0116] It should be noted that compared with the traditional method without random perturbation factors, particles may quickly converge to the vicinity of the local optimal solution, causing the search process to end prematurely and failing to find the global optimal solution. By introducing random perturbation factors, more potential high-quality solutions can be discovered, thereby increasing the possibility of finding the global optimal solution and helping to improve the overall performance of the aircraft power system finally selected.

[0117] (6) For each discrete variable in a particle, the position of the particle is updated based on Bayesian reasoning.

[0118] In the specific implementation, Bayesian theorem is used to combine prior probability and observation data to calculate the posterior probability value of each particle taking each discrete variable. This probability value is used to characterize the possibility of each power type becoming the optimal choice. Based on the posterior probability value, the discrete variable position of each particle is resampled.

[0119] A specific embodiment is given below to introduce in detail the process of updating the position of a particle based on Bayesian reasoning. Optionally, in a possible implementation, for each discrete variable in a particle, the position of the particle is updated based on Bayesian reasoning, including:

[0120] Step 1: For each particle, calculate the conditional probability of the particle selecting each power system type according to the sub-fitness function pre-set for the power system type.

[0121] Specifically, the sub-fitness function is an evaluation function set for each power system type, which is used to calculate the fitness or performance of the particle when the power system type is selected.

[0122] In specific implementation, it can be determined based on the design input parameters of the aircraft and the performance parameters of the power system: flight altitude, flight speed, range, power system weight, fuel efficiency, etc.

[0123] Furthermore, the conditional probability value refers to the probability value calculated based on the sub-fitness function when a particle selects a certain power system type.

[0124] In specific implementation, for each particle, the conditional probability value of the particle selecting each power system type is calculated based on its current position and sub-fitness function. The conditional probability value can be calculated based on the following formula:

[0125] ;

[0126] in, is the conditional probability value of the ith particle, is the sub-fitness function, and α is the adjustment parameter.

[0127] Step 2: Using the Bayesian formula, according to the prior distribution values ​​initially assigned to each power system type and the conditional probability values ​​of the particle selecting each power system type, determine the posterior probability values ​​of the particle selecting each power system type.

[0128] In the specific implementation, the Bayesian formula is used to combine the prior distribution values ​​initially assigned to each power system type and the conditional probability values ​​of the particle selecting each power system type to calculate the posterior probability values ​​of the particle selecting each power system.

[0129] It can be appreciated that the posterior probability values ​​provide an updated assessment of the likelihood of each powertrain type being selected after the conditional probability values ​​are obtained.

[0130] In a possible implementation, the posterior probability value of each power system selected by the particle is calculated according to the seventh formula; wherein the seventh formula is:

[0131] ;

[0132] Among them, the The posterior probability value of the kth power system is selected for the i-th particle; k ranges from 1 to N; N is the total number of power system types; The conditional probability value of the kth dynamic system is selected for the i-th particle; is the prior distribution value assigned to the kth power system type at the initial time.

[0133] Specifically, the prior distribution value is determined according to the number of power systems. In specific implementation, the same prior distribution value can be set for each power system, and the sum of the prior distribution values ​​of all power systems is set to 1. For example, in one embodiment, the preset multiple power types include A, B, and C. At this time, the value of N is 3, so the prior distribution values ​​of power systems A, B, and C are all 1 / 3.

[0134] In specific implementation, for a certain particle, its position can be updated as follows:

[0135] first, = ;

[0136] ;

[0137] ;

[0138] Secondly, = =1 / 3;

[0139] Further:

[0140] ;

[0141] ;

[0142] ;

[0143] In this way, we can calculate , , .

[0144] Step 3: Select the posterior probability value of each power system according to the particle, and select the posterior probability value corresponding to the power system type with the largest probability value as the updated position of the particle.

[0145] Specifically, according to the calculated posterior probability value, the posterior probability value corresponding to the power system type with the largest probability value is selected as the updated position of the particle. In other words, the particle is moved to the position representing the optimal power system type in the solution space according to the posterior probability value.

[0146] Combining the above example, for example, the posterior probability value of particle i selecting dynamic system A is calculated is 25%, the posterior probability value of particle i selecting dynamic system B is 35%, the posterior probability value of particle i selecting dynamic system C is 40%, which means that the probability of the particle selecting power system C is the highest. Determines the updated position for the particle.

[0147] It should be noted that, on the one hand, by pre-setting the sub-fitness function for the power system type, the performance of each particle when selecting different power system types can be accurately evaluated based on the sub-fitness function. As the iterative process proceeds, the particle will continuously adjust the power system type it selects according to its current selection and gradually approach the global optimal solution; on the other hand, using the Bayesian formula, the prior distribution value initially assigned to each power system type is combined with the conditional probability value of the particle selecting each power system type, thereby obtaining the posterior probability value of the particle selecting each power system. In this way, not only the prior knowledge (that is, the designer or expert's initial judgment on the power system type) is considered, but also the new evidence (that is, the performance of the particle in the iterative process) is integrated, making the posterior probability value more accurate and reliable, which helps to improve the efficiency and accuracy of the optimization algorithm, thereby selecting the most suitable power system type for the aircraft in a shorter time.

[0148] (7) When the preset termination condition is met and the global optimal solution is output as the optimal solution to the target optimization problem, if the preset termination condition is not met, the step of screening out particles that do not meet the constraint conditions according to the first constraint condition, the second constraint condition and the third constraint condition is performed again.

[0149] Specifically, the preset termination condition is used to determine whether the optimization has found a sufficiently good solution or whether the limitation of computing resources has been reached.

[0150] It can be understood that, during the iterative calculation process, when the number of iterations reaches a preset value, it can be determined that the iterative calculation process is completed, and the maximum global fitness in the entire calculation process is obtained.

[0151] Furthermore, the convergence condition is pre-set as , where δ is a fixed small value, that is, after D consecutive iterations, when the global optimal solution does not change significantly, it can be determined that the iterative process is terminated.

[0152] In specific implementation, for example, in one embodiment, the preset termination condition can be determined by the number of iterations: the maximum number of iterations is set to 2000 times, and when the current number of iterations reaches the maximum number of iterations, it is determined that the preset termination condition is satisfied, and the global optimal solution is output. For another example, in another embodiment, the preset termination condition can be determined by the change of the global optimal solution during the iteration process. When the global optimal solution does not change in 50 consecutive iterations, it is determined that the preset termination condition is satisfied, and the global optimal solution is output. It should be noted that the present application proposes a method for selecting an aircraft power system based on multi-objective optimization, which aims to comprehensively consider factors such as range, output power, weight, reliability and cost, and select the optimal power system and configuration parameters.

[0153] The aircraft power system selection method based on multi-objective optimization provided in this embodiment constructs a target optimization problem by taking maximizing global fitness as the goal, taking the thrust of the power system meeting the preset minimum flight requirement as the first constraint, the total weight of the power system cannot exceed the structural upper limit as the second constraint, and the total cost of the power system cannot be lower than the preset maximum economic cost as the third constraint, and then constructs a fitness function for the target optimization problem, so as to solve the optimal solution of the target optimization problem with the goal of maximizing the global fitness according to the fitness function, and determine the solved optimal solution as the selected power system. In this way, on the one hand, by constructing the fitness function, multiple optimization goals such as range, output power, weight, reliability and cost are comprehensively considered to ensure that the selected power system achieves the optimization of overall performance while meeting the flight mission requirements. In other words, with the goal of maximizing global fitness, the particle swarm optimization algorithm is used to continuously approximate and find the optimal solution in the iterative process. The optimal solution ensures that the selected power system is the global optimal rather than the local optimal. Secondly, by automatically generating the initial particle swarm, updating the power system type and configuration parameters and other steps, the intelligent selection of the power system is realized. Through the automated selection and optimization process, not only the time cycle of power system selection is greatly shortened and the design efficiency is improved, but also the accuracy and reliability of the selected power system are ensured, and the design risk is reduced. Thirdly, in each iteration process, the discrete variables used to characterize the power system type are updated based on Bayesian reasoning, and the continuous variables used to characterize the configuration parameters are updated based on the particle swarm optimization algorithm. The combination of particle swarm optimization algorithm and Bayesian reasoning can find the optimal power system and configuration parameters under complex constraints, greatly improving the search efficiency and accuracy, and providing effective decision support for the design of aircraft power systems. Fourthly, this method is applicable to various types of aircraft and has good scalability and applicability.

[0154] Corresponding to the aforementioned embodiment of an aircraft power system selection method based on multi-objective optimization, the present application also provides an embodiment of an aircraft power system selection device based on multi-objective optimization.

[0155] The present application discloses an embodiment of an aircraft power system selection device based on multi-objective optimization, which can be applied to an aircraft power system selection device based on multi-objective optimization. The device embodiment can be implemented by software, or by hardware or a combination of software and hardware. Taking software implementation as an example, as a device in a logical sense, the processor of the aircraft power system selection device based on multi-objective optimization reads the corresponding computer program instructions in the non-volatile memory into the memory and runs them. From a hardware perspective, if Figure 2 As shown, it is a hardware structure diagram of the aircraft power system selection device based on multi-objective optimization of the present application, in which the aircraft power system selection device based on multi-objective optimization is located, except Figure 2 In addition to the processor, memory, network interface, and non-volatile memory shown, the aircraft power system selection device based on multi-objective optimization in which the device in the embodiment is located may also include other hardware, which will not be elaborated here.

[0156] Figure 3 This is a structural diagram of the first embodiment of the aircraft power system selection device based on multi-objective optimization provided by this application. Figure 3 The device provided in this embodiment includes a processing module 310, a construction module 320 and a calculation module 330; wherein:

[0157] The processing module 310 is used to construct a target optimization problem with the goal of maximizing global fitness, with the thrust of the power system meeting the preset minimum flight requirement as the first constraint condition, the total weight of the power system not exceeding the structural upper limit as the second constraint condition, and the total cost of the power system not being lower than the preset maximum economic cost as the third constraint condition; wherein the target optimization problem maps the configuration parameters of the power system to the search space of the particle swarm, each particle includes a set of configuration parameters of a power system, and the position of each particle is composed of a discrete variable for characterizing the type of the power system and a continuous variable for characterizing the configuration parameters of the power system;

[0158] The construction module 320 is used to construct a fitness function of the target optimization problem; wherein the fitness function is a fitness function that comprehensively considers range, output power, weight, reliability and cost;

[0159] The calculation module 330 is used to solve the optimal solution of the target optimization problem according to the fitness function with the goal of maximizing the global fitness, and determine the solved optimal solution as the selected power system;

[0160] Wherein, in the process of solving the optimal solution of the target optimization problem, in each iteration, the discrete variable updates the position based on Bayesian reasoning; and the continuous variable updates the position based on the particle swarm optimization algorithm.

[0161] The device of this embodiment can be used to perform Figure 1 The steps, specific implementation principles and implementation processes of the method embodiment shown are similar and will not be repeated here.

[0162] Please continue to refer to Figure 2 The present application also provides an aircraft power system selection device based on multi-objective optimization, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the steps of any one of the methods provided in the first aspect of the present application are implemented.

[0163] The present application also provides a computer-readable storage medium having a computer program stored thereon, and when the program is executed by a processor, the steps of any one of the methods provided in the present application are implemented.

[0164] The implementation process of the functions and effects of each unit in the above-mentioned device is specifically described in the implementation process of the corresponding steps in the above-mentioned method, and will not be repeated here.

[0165] For the device embodiment, since it basically corresponds to the method embodiment, the relevant parts can refer to the partial description of the method embodiment. The device embodiment described above is only schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the present application scheme. A person of ordinary skill in the art can understand and implement it without paying any creative work.

[0166] The above description is only a preferred embodiment of the present application and is not intended to limit the present application. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present application shall be included in the scope of protection of the present application.

Claims

1. A method for selecting an aircraft power system based on multi-objective optimization, characterized in that: The method comprises: A target optimization problem is constructed with the goal of maximizing global fitness, with the thrust of the power system meeting the preset minimum flight requirement as the first constraint, the total weight of the power system cannot exceed the structural upper limit as the second constraint, and the total cost of the power system cannot be lower than the preset maximum economic cost as the third constraint; wherein the target optimization problem maps the configuration parameters of the power system to the search space of the particle swarm, each particle includes a set of configuration parameters of a power system, and the position of each particle is composed of a discrete variable for characterizing the type of the power system and a continuous variable for characterizing the configuration parameters of the power system; Constructing a fitness function for the target optimization problem; wherein the fitness function is a fitness function that comprehensively considers range, output power, weight, reliability and cost; According to the fitness function, with the goal of maximizing the global fitness, solving the optimal solution of the target optimization problem, and determining the solved optimal solution as the selected power system; The step of solving the optimal solution of the target optimization problem according to the fitness function and taking maximizing the global fitness as the goal includes: Using a chaotic initialization strategy to obtain an initial particle swarm, and according to the first constraint condition, the second constraint condition and the third constraint condition, screening out particles that do not meet the constraint conditions; For each particle, based on the position of the particle and the fitness function, determine the fitness value of the particle, and determine the current global optimal solution and the local optimal solution of each particle according to the fitness value of each particle; According to the current number of iterations, determine the random disturbance factor in the current iteration process; Determine the inertia weight in the current iteration process according to the current iteration number, the preset initial inertia weight, the preset minimum inertia weight and the preset maximum iteration number; For the continuous variables in each particle, based on the particle swarm optimization algorithm, the speed of the particle is updated according to the inertia weight, and the position of the particle is updated according to the updated speed of the particle, the current position of the particle and the random disturbance factor; For each discrete variable in a particle, the position of the particle is updated based on Bayesian reasoning; The method of updating the position of each particle based on Bayesian reasoning for the discrete variables in the particle includes: For each particle, the conditional probability value of the particle selecting each power system type is calculated according to the sub-fitness function pre-set for the power system type; Using the Bayesian formula, according to the prior distribution values ​​initially assigned to each power system type and the conditional probability values ​​of the particle selecting each power system type, the posterior probability values ​​of the particle selecting each power system type are determined; Select the posterior probability value of each power system according to the particle, and select the posterior probability value corresponding to the power system type with the largest probability value as the updated position of the particle; When the preset termination condition is met and the global optimal solution is output as the optimal solution of the target optimization problem, if the preset termination condition is not met, the step of screening out particles that do not meet the constraint conditions according to the first constraint condition, the second constraint condition and the third constraint condition is performed again; Wherein, in the process of solving the optimal solution of the target optimization problem, in each iteration, the discrete variable updates the position based on Bayesian reasoning; and the continuous variable updates the position based on the particle swarm optimization algorithm.

2. The method according to claim 1, characterized in that The fitness function is: ； Among them, the , , , are respectively preset weight coefficients; To maximize the range, the Calculated according to the first formula; To minimize weight, the Calculated according to the second formula; For maximizing output power, the Calculated according to the third formula; For maximum reliability, the Calculated according to the fourth formula; To minimize the cost, the Calculate according to the fifth formula; The first formula is: ； ； Wherein, R is the range; is the minimum range of the search space; is the maximum range of the search space; v is the flight speed; is the total fuel weight; W is the fuel consumption rate per unit time; is the weight of the battery; is the energy density of the battery; is the average power consumption of the aircraft; The second formula is: ； Among them, W2 is the weight value of the power system generated by the particle swarm, is the maximum weight of the search space; is the minimum weight of the search space; The third formula is: ； Wherein, P is the output power generated by the particle swarm; is the maximum power of the search space; is the minimum power of the search space; The fourth formula is: ； Wherein, S is the system reliability of the aircraft; is the minimum failure probability of the power system within its design life cycle; is the maximum failure probability of the power system during its design life cycle; The fifth formula is: ； Among them, the is the maximum cost of the power system; is the minimum cost of the power system.

3. The method according to claim 1, characterized in that The updating of the position of the particle according to the updated speed of the particle, the current position of the particle and the random disturbance factor comprises: The position of the particle is updated according to the sixth formula; wherein the sixth formula is: ； Among them, the is the updated position of the i-th particle; is the current position of the ith particle; is the updated velocity of the ith particle, and ε is a Gaussian distributed random number.

4. The method according to claim 1, characterized in that: Determining the random disturbance factor in the current iteration process according to the current iteration number includes: Determine the current exploration stage based on the current number of iterations and the preset maximum number of iterations; According to the current exploration stage, determining a random perturbation standard deviation corresponding to the exploration stage; Taking 0 as the mean of the Gaussian distribution and the standard deviation of the random perturbation as the standard deviation of the Gaussian distribution, a random number is randomly generated based on the Gaussian distribution, and the random number is determined as the random perturbation factor in the current iteration process.

5. The method according to claim 1, characterized in that The method of using the Bayesian formula to determine the posterior probability value of the particle selecting each power system according to the prior distribution value initially assigned to each power system type and the conditional probability value of the particle selecting each power system type includes: The posterior probability value of each power system selected by the particle is calculated according to the seventh formula; wherein the seventh formula is: ； Among them, the The posterior probability value of the kth power system is selected for the i-th particle; k ranges from 1 to N; N is the total number of power system types; The conditional probability value of the kth dynamic system is selected for the i-th particle; is the prior distribution value assigned to the kth power system type at the initial time.

6. The method according to claim 1, characterized in that The set of configuration parameters includes the following parameters: flight speed, energy density of the battery, average power consumption of the aircraft, weight of the battery, weight of the power system, total fuel weight, output power, reliability and cost.

7. An aircraft power system selection device based on multi-objective optimization, characterized in that: The device comprises a processing module, a building module and a computing module; wherein: The processing module is used to construct a target optimization problem with the goal of maximizing global fitness, with the thrust of the power system meeting the preset minimum flight requirement as the first constraint condition, the total weight of the power system cannot exceed the structural upper limit as the second constraint condition, and the total cost of the power system cannot be lower than the preset maximum economic cost as the third constraint condition; wherein the target optimization problem maps the configuration parameters of the power system to the search space of the particle swarm, each particle includes a set of configuration parameters of a power system, and the position of each particle is composed of a discrete variable for characterizing the type of the power system and a continuous variable for characterizing the configuration parameters of the power system; The construction module is used to construct a fitness function of the target optimization problem; wherein the fitness function is a fitness function that comprehensively considers range, output power, weight, reliability and cost; The calculation module is used to solve the optimal solution of the target optimization problem according to the fitness function with the goal of maximizing the global fitness, and determine the solved optimal solution as the selected power system; The step of solving the optimal solution of the target optimization problem according to the fitness function and taking maximizing the global fitness as the goal includes: Using a chaotic initialization strategy to obtain an initial particle swarm, and screening out particles that do not meet the constraints according to the first constraint condition, the second constraint condition and the third constraint condition; For each particle, based on the position of the particle and the fitness function, determine the fitness value of the particle, and determine the current global optimal solution and the local optimal solution of each particle according to the fitness value of each particle; According to the current number of iterations, determine the random disturbance factor in the current iteration process; Determine the inertia weight in the current iteration process according to the current iteration number, the preset initial inertia weight, the preset minimum inertia weight and the preset maximum iteration number; For the continuous variables in each particle, based on the particle swarm optimization algorithm, the speed of the particle is updated according to the inertia weight, and the position of the particle is updated according to the updated speed of the particle, the current position of the particle and the random disturbance factor; For each discrete variable in a particle, the position of the particle is updated based on Bayesian reasoning; The method of updating the position of each particle based on Bayesian reasoning for the discrete variables in the particle includes: For each particle, the conditional probability value of the particle selecting each power system type is calculated according to the sub-fitness function pre-set for the power system type; Using the Bayesian formula, according to the prior distribution values ​​initially assigned to each power system type and the conditional probability values ​​of the particle selecting each power system type, the posterior probability values ​​of the particle selecting each power system type are determined; Select the posterior probability value of each power system according to the particle, and select the posterior probability value corresponding to the power system type with the largest probability value as the updated position of the particle; When the preset termination condition is met and the global optimal solution is output as the optimal solution of the target optimization problem, if the preset termination condition is not met, the step of screening out particles that do not meet the constraint conditions according to the first constraint condition, the second constraint condition and the third constraint condition is performed again; Wherein, in the process of solving the optimal solution of the target optimization problem, in each iteration, the discrete variable updates the position based on Bayesian reasoning; and the continuous variable updates the position based on the particle swarm optimization algorithm.

8. An aircraft power system selection device based on multi-objective optimization, characterized in that: The method comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps of the method according to any one of claims 1 to 6 when executing the program.

Images