A control method for marine permanent magnet synchronous motor based on hybrid grey wolf whale optimization algorithm

Abstract

The application relates to the technical field of marine permanent magnet synchronous motors, and discloses a marine permanent magnet synchronous motor control method based on a hybrid grey wolf whale algorithm, which comprises the following steps: firstly, a decoupling mathematical model of a marine permanent magnet synchronous motor under a d-q synchronous rotating coordinate system is constructed; an id=0 vector control strategy is adopted, so that the electromagnetic torque Te and the q-axis current iq are in a linear relationship; a multi-target optimization function J is designed as a target function of the hybrid grey wolf whale algorithm; subsequently, the hybrid grey wolf whale algorithm is executed to perform online or offline optimization on the kp and ki parameters of a PI controller; and finally, precise speed tracking and anti-disturbance control of the marine PMSM are realized. The application realizes multi-index collaborative optimization; under typical working conditions of a ship, the dynamic response speed of the system is obviously faster than that of a traditional PID control, the steady-state speed control is more stable, torque fluctuation is significantly reduced, and the combination of fast response and high-precision control is realized.

Description

Technical Field

[0001] This invention relates to the field of marine permanent magnet synchronous motor technology, specifically a control method for marine permanent magnet synchronous motors based on the hybrid gray wolf whale algorithm. Background Technology

[0002] Permanent magnet synchronous motors (PMSMs) have been widely used in marine electric propulsion systems due to their high reliability and cost advantages. However, because PMSMs are inherently multivariable, strongly coupled nonlinear systems, their operation is susceptible to fluctuations in system parameters and external load disturbances. Traditional PI controllers are highly dependent on motor parameters, and using this controller alone is insufficient to meet the accuracy requirements of speed control systems. To improve control performance, various nonlinear control techniques have been gradually applied to PMSM speed control systems, mainly covering model predictive control, sliding mode control, active disturbance rejection control, and linear optimal control.

[0003] In the aforementioned control framework, the introduction of swarm intelligence optimization algorithms such as the Grey Wolf Optimizer (GWO) and the Whale Optimization Algorithm (WOA) effectively solves the challenge of controller parameter tuning. The Grey Wolf Optimizer, with its simple optimization principle and fast convergence speed, is often used to optimize controller parameters in backpropagation control and PID control. It achieves adaptive speed tracking through iterative optimization, improving the system's stability and robustness under parameter perturbations and load disturbances. The Whale Optimizer, characterized by its outstanding global search capability, has been improved by introducing strategies such as nonlinear inertial weights. This allows for precise tuning of the weight coefficients of PI controllers or model predictive control, significantly reducing speed overshoot, lowering the traditional PI control overshoot from 5% to 1.5%, while also enhancing its resistance to load disturbances. Neither algorithm relies on a precise motor model. When combined with nonlinear techniques such as sliding mode control, they can further compensate for the shortcomings of traditional control, providing a more accurate and reliable control scheme for PMSM speed control systems.

[0004] The application of the traditional Grey Wolf algorithm in the control of marine permanent magnet synchronous motors still has several limitations. First, the algorithm's convergence speed is slow, making it difficult to quickly adjust controller parameters in situations requiring high real-time performance in marine propulsion systems, especially during sudden load changes or operating condition transitions, resulting in lag in dynamic response. Second, the Grey Wolf algorithm is sensitive to initial parameter values, easily getting trapped in local optima under complex conditions such as motor parameter perturbations, load fluctuations, and wind and wave interference, leading to decreased control accuracy. Furthermore, the algorithm's performance significantly deteriorates when handling high-dimensional optimization problems, making it difficult to simultaneously optimize multiple control parameters such as the current loop and speed loop, affecting the overall control effect. Simultaneously, its optimization process lacks adaptability to the nonlinear characteristics of the motor, resulting in insufficient control robustness under low-speed heavy-load or high-speed weak-field conditions, potentially causing continuous fluctuations in speed and torque, making it difficult to meet the wide-condition, high-reliability operation requirements of marine propulsion systems.

[0005] Traditional whale algorithms also have significant shortcomings in the application of marine permanent magnet synchronous motor (PMSM) control. First, their convergence speed is generally slow. When ship propulsion systems frequently encounter load changes, wind and wave disturbances, and operating condition switching, they struggle to quickly optimize controller parameters online, leading to sluggish dynamic response and increased speed and torque fluctuations. Second, the whale algorithm has limited optimization accuracy. When dealing with complex characteristics such as motor parameter perturbations, back-EMF cross-coupling, and nonlinear friction, it is prone to inaccurate parameter tuning, thus reducing the system's steady-state performance. Furthermore, the algorithm is sensitive to initial parameter settings and lacks robustness under different navigation conditions, especially in low-speed, heavy-load or high-speed, weak-field regions, where control performance is easily affected. Simultaneously, the whale algorithm has weak search capabilities in high-dimensional optimization spaces, making it difficult to simultaneously optimize multiple control objectives, such as dynamic response speed, steady-state accuracy, and disturbance rejection capability, thus limiting its further application in wide-condition speed control systems for marine PMSMs. Summary of the Invention

[0006] The purpose of this invention is to primarily solve the following problems: 1. Parameter tuning is difficult and has poor adaptability. Traditional PID controller parameters are mainly determined by experience or trial and error, making it difficult to balance dynamic response and steady-state accuracy under the wide operating conditions of ships. When faced with changes in motor parameters, sudden load changes, and wind and wave disturbances, traditional PID controllers lack effective online self-tuning capabilities, resulting in a significant decline in control performance.

[0007] 2. It has weak anti-interference ability and cannot effectively compensate for nonlinear factors. Traditional PID controllers lack direct compensation mechanisms for nonlinear characteristics such as back EMF, cross-coupling, and magnetic circuit saturation, making them prone to speed and torque fluctuations in medium- and high-speed or weak magnetic regions. Furthermore, they lack robustness to perturbations of propulsion system parameters and the superposition of multiple disturbances, failing to meet the stringent stability and reliability requirements of marine propulsion systems. Therefore, a control method for marine permanent magnet synchronous motors based on the hybrid gray wolf-whale algorithm is proposed.

[0008] The technical solution of the present invention to solve the above-mentioned technical problems is as follows: A control method for marine permanent magnet synchronous motors based on the hybrid gray wolf-whale algorithm includes the following steps: S1. Construct a decoupled mathematical model of a marine permanent magnet synchronous motor in the dq synchronous rotating coordinate system, and adopt an id=0 vector control strategy to make the electromagnetic torque Te and the q-axis current iq linearly related. S2. Design a multi-objective optimization function J as the objective function of the hybrid gray wolf whale algorithm; S3. Execute the hybrid gray wolf whale algorithm to optimize the kp and ki parameters of the PI controller online or offline. The hybrid gray wolf whale algorithm includes the following optimization mechanisms: I. Initialize the population using an improved Tent chaotic map; II. The search behavior is dynamically adjusted using a nonlinear convergence factor; III. In each iteration, a WOA bubble network spiral approximation mechanism is introduced to update the α wolf position of the gray wolf algorithm; IV. At the end of each iteration, perform Cauchy mutation on the global optimal solution; S4. The optimal kp and ki parameters obtained by the hybrid gray wolf whale algorithm are loaded into the PI speed loop controller in real time, and combined with the id=0 current loop to form a dual closed-loop vector control system, so as to realize the precise tracking and disturbance rejection control of the speed of the marine PMSM.

[0009] Based on the above technical solution, the present invention can be further improved as follows.

[0010] Preferably, the expression for the multi-objective optimization function J is as follows: ; in, For a given rotational speed, This refers to the actual rotational speed. This represents the torque ripple amplitude. To adjust the system time, This is the q-axis voltage. , , , All are weighted coefficients.

[0011] Preferably, the initialization of the population using the improved Tent chaotic map includes the following steps: when hour, ;when hour, or ,generate The chaotic sequence with enhanced ergodicity within the interval is mapped to the feasible region of parameters kp and ki through a linear transformation.

[0012] Preferably, the nonlinear convergence factor a dynamically adjusted search behavior is calculated using the following formula: ; ; ; in, Indicates the convergence factor; express The initial value; express The value at the end of the iteration; This indicates the total number of iterations preset by the algorithm; This indicates the current iteration round of the algorithm.

[0013] Preferably, the WOA bubble network spiral approximation mechanism introduced for updating the α wolf position of the gray wolf algorithm in each iteration is calculated by the following formula: dα represents the dynamic distance between the current candidate solution and wolf α; xa represents the position of the alpha wolf α, i.e. the global optimal solution in the t-th iteration; x represents the position of the current candidate solution; c1 represents the random coefficient vector; p1 represents a random number uniformly distributed in the interval [0,1]; ρ represents the scaling coefficient; b represents the spiral shape parameter; l represents the random coefficient.

[0014] Preferably, the position update process of the hybrid gray wolf whale algorithm in step S3 further includes: The three individuals with the best fitness in the population were selected as α wolf, β wolf, and δ wolf, respectively. For each individual, calculate using the Grey Wolf algorithm; The Whale Algorithm selects an update strategy with probability p∈[0,1]: when p<0.5 and |A|<1, it performs prey encirclement update; when p<0.5 and |A|≥1, it performs random search update; when p≥0.5, it performs spiral update. The results of the gray wolf algorithm and the whale algorithm are combined; Finally, boundary correction and Cauchy mutation are performed to complete a single iteration.

[0015] Compared with the prior art, the technical solution of this application has the following beneficial technical effects: 1. This invention integrates the global search capability of GWO and the local convergence characteristics of WOA by using TCGWOWOA. It automatically adjusts the search intensity at different iteration stages through a nonlinear convergence factor. The multi-objective optimization function comprehensively considers speed error, torque fluctuation and response speed to achieve synergistic optimization of multiple indicators. Under typical ship operating conditions, the system's dynamic response speed is significantly faster than traditional PID control, the steady-state speed control is more stable, and the torque fluctuation is significantly reduced, achieving a balance between fast response and high-precision control.

[0016] 2. This invention uses the Cauchy mutation mechanism to help the algorithm quickly escape local optima after disturbances; improves the Tent mapping to enhance population diversity and make the search range more comprehensive; the working condition adaptive triggering mechanism can detect disturbances in real time and start optimization immediately to achieve rapid response. When faced with large load changes, motor parameter drift and bus voltage fluctuations caused by wind and waves during ship navigation, the system can quickly recover stability without obvious overshoot or oscillation.

[0017] 3. This invention ensures uniform initial population distribution by improving the Tent mapping, thereby enhancing the search robustness under complex operating conditions. The nonlinear compensation term is introduced into the optimization objective to effectively offset the nonlinear effects of the motor under different operating conditions. Under typical ship operating conditions such as low-speed heavy load, high-speed weak field and frequent start-stop, the system can maintain stable and accurate control performance without significant attenuation during operating condition switching.

[0018] 4. This invention reduces the computational burden through lightweight design using TCGWOWOA, and the parameter range conforms to engineering practice; the algorithm is essentially a general parameter optimization engine, which can flexibly connect to the key parameters of different control structures without modifying existing hardware, and can be deployed only through software upgrades; the algorithm has high execution efficiency, meeting the real-time requirements of motor control; it is compatible with multiple architectures such as PID, vector control, and model predictive control, and can be extended to marine auxiliary motor systems. Attached Figure Description

[0019] Figure 1 The flowchart of the improved TCGWOWOA algorithm of this invention (steps one to four); Figure 2 The flowchart of the improved TCGWOWOA algorithm of this invention (steps 5 to 6) is shown. Figure 3 The flowchart of the improved TCGWOWOA algorithm of this invention (steps 7 to 8) is shown. Figure 4 This is a diagram of the TCGWOWOA control system of the present invention; Figure 5 This is a block diagram of the PMSM speed control system of the present invention; Figure 6 This is an overall diagram showing the change in motor speed according to the present invention; Figure 7 This is a partially enlarged view of the sudden load applied to this invention; Figure 8 This is a partially enlarged view of the sudden load unloading of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] A control method for marine permanent magnet synchronous motors based on the hybrid gray wolf-whale algorithm includes the following steps: S1. Construct a decoupled mathematical model of a marine permanent magnet synchronous motor in the dq synchronous rotating coordinate system, and adopt an id=0 vector control strategy to make the electromagnetic torque Te and the q-axis current iq linearly related. S2. Design a multi-objective optimization function J as the objective function of the hybrid gray wolf whale algorithm; S3. Execute the hybrid gray wolf whale algorithm to optimize the kp and ki parameters of the PI controller online or offline. The hybrid gray wolf whale algorithm includes the following optimization mechanisms: I. Initialize the population using an improved Tent chaotic map; II. The search behavior is dynamically adjusted using a nonlinear convergence factor; III. In each iteration, a WOA bubble network spiral approximation mechanism is introduced to update the α wolf position of the gray wolf algorithm; IV. At the end of each iteration, perform Cauchy mutation on the global optimal solution; S4. The optimal kp and ki parameters obtained by the hybrid gray wolf whale algorithm are loaded into the PI speed loop controller in real time, and combined with the id=0 current loop to form a dual closed-loop vector control system, so as to realize the precise tracking and disturbance rejection control of the speed of the marine PMSM.

[0022] The expression for the multi-objective optimization function J is as follows: ; in, For a given rotational speed, This refers to the actual rotational speed. This represents the torque ripple amplitude. To adjust the system time, This is the q-axis voltage. , , , All are weighted coefficients.

[0023] The process of initializing the population using the improved Tent chaotic map includes the following steps: when hour, ;when hour, or ,generate The chaotic sequence with enhanced ergodicity within the interval is mapped to the feasible region of parameters kp and ki through a linear transformation.

[0024] The nonlinear convergence factor a is used to dynamically adjust the search behavior, which is calculated by the following formula: ; ; ; in, Indicates the convergence factor; express The initial value; express The value at the end of the iteration; This indicates the total number of iterations preset by the algorithm; This indicates the current iteration round of the algorithm.

[0025] The WOA bubble network spiral approximation mechanism is introduced to update the α wolf position of the gray wolf algorithm in each iteration, and is calculated by the following formula: ; dα represents the dynamic distance between the current candidate solution and wolf α; xa represents the position of the alpha wolf α, i.e. the global optimal solution in the t-th iteration; x represents the position of the current candidate solution; c1 represents the random coefficient vector; p1 represents a random number uniformly distributed in the interval [0,1]; ρ represents the scaling coefficient; b represents the spiral shape parameter; l represents the random coefficient.

[0026] The position update process of the hybrid gray wolf whale algorithm in step S3 further includes: The three individuals with the best fitness in the population were selected as α wolf, β wolf, and δ wolf, respectively. For each individual, calculate using the Grey Wolf algorithm; The Whale Algorithm selects an update strategy with probability p∈[0,1]: when p<0.5 and |A|<1, it performs prey encirclement update; when p<0.5 and |A|≥1, it performs random search update; when p≥0.5, it performs spiral update. The results of the gray wolf algorithm and the whale algorithm are combined; Finally, boundary correction and Cauchy mutation are performed to complete a single iteration.

[0027] The TCGWOWOA algorithm (also known as the hybrid gray wolf-whale algorithm) deeply integrates the global exploration advantages of GWO with the local exploitation characteristics of WOA, and incorporates improved Tent chaotic mapping, nonlinear convergence factors, and Cauchy mutation optimization strategies to construct a hybrid optimization framework that combines efficiency and robustness. The core logic of this algorithm is to balance population diversity and convergence speed through a multi-stage collaborative update mechanism, avoiding getting trapped in local optima, and ultimately achieving an accurate solution to complex optimization problems. Its detailed algorithm flow is as follows: Algorithm TCGWOWOA Inputs: maximum number of iterations (max_epoch), population size (pop_size), problem boundaries (problem_bounds) (including upper bounds ub and lower bounds lb for each dimension), and objective function obj_fun (used to evaluate the fitness of the solution; this paper aims to minimize the maximum completion time MakeSpan). Output: the global optimal solution (best_solution) and the optimal fitness value (best_fitness). Step 1: Startup and Parameter Input Start executing the algorithm; Input core parameters: maximum number of iterations max_epoch, population size pop_size, problem boundaries problem_bounds (including upper and lower bounds lb / ub for each dimension), and objective function obj_fun (used to evaluate the fitness of the solution).

[0028] Step 2: Population initialization and initial optimal solution setting Population initialization based on improved Tent chaotic mapping: generate pop_size chaotic values ​​in the range [0,1], and map them to the problem_bounds solution space through linear transformation to ensure uniform distribution of initial solutions; The objective function obj_fun is called to calculate the fitness value for each initial solution; Sort the population in ascending order of fitness value (the smaller the fitness value, the better the quality of the solution). Initialize the global optimal solution (best individual after sorting) and the optimal fitness value (best fitness). Initialize the iteration count epoch=0.

[0029] Step 3: Iteration loop judgment Determine if the epoch is less than max_epoch: If not, directly output the global optimal solution best_solution and the optimal fitness value best_fitness, and the algorithm terminates; if yes, enter the iterative core process.

[0030] Step 4: Iterative Core Computation (GWO related) Calculate the nonlinear convergence factor a: the formula is a=2×(1-epoch / (2×max_epoch)), to achieve a dynamic balance between global exploration and local development; The top 3 optimal solutions of the population are selected as α wolf (global best), β wolf (second best), and δ wolf (relatively best) to guide the population's search for optimality. For each solution in the population, perform the following operations: Generate 6 random numbers r1~r6 in the range [0,1], and calculate the random coefficient vectors A1=2a×r1-a, A2=2a×r2-a, A3=2a×r3-a, C1=2×r4, C2=2×r5, C3=2×r6; Calculate the position X1 guided by α wolf: X1 = alpha - A1 × C1 × abs (alpha - current solution); Calculate the position X2 guided by the β wolf: X2 = beta - A2 × C2 × abs (beta - current solution); Calculate the position X3 guided by the δ wolf: X3 = delta - A3 × C3 × abs (delta - current solution); The location of the integrated GWO third-party bootloader is: gwo_new_pos=(X1+X2+X3) / 3.

[0031] Step 5: WOA Multi-Strategy Position Update Generate random numbers p (0 ≤ p ≤ 1), and select the WOA update strategy according to probability: If p < 0.5 and |A| < 1: execute the prey encirclement strategy, woa_pos = alpha - A × abs (alpha - current solution); If p < 0.5 and |A| ≥ 1: execute the random search strategy, randomly select an individual rand_sol that is not the current solution in the population, and woa_pos = rand_sol - A × C × abs (rand_sol - current solution); If p≥0.5: Execute the spiral update strategy, calculate D=abs(alpha-current solution), b=1 (spiral parameter), l=rand(-1,1), woa_pos=D×exp(b×l)×cos(2×π×l)+alpha.

[0032] Step 6: Defusion and Optimization Merge GWO and WOA positions: candidate_pos = 0.5 × gwo_new_pos + 0.5 × woa_pos; Boundary correction: If candidate_pos exceeds the upper and lower bounds of problem_bounds, it is set to lb if it is less than the lower bound and to ub if it is greater than the upper bound; Cauchy mutation optimization: mutated_pos = candidate_pos + 0.1 × cauchy(0,1) (0.1 is the scale parameter, cauchy(0,1) is a random number from the standard Cauchy distribution). Call obj_fun to calculate the fitness value mutated_fitness of the mutated solution mutated_pos.

[0033] Step 7: Greedy Selection and Population Renewal 22. Greedy selection: If `mutated_fitness` < the fitness value of the current solution, replace the current solution with `mutated_pos`; otherwise, retain the current solution. 23. Update the population (a new population is obtained after replacement / retention); 24. Update the global optimal solution: Traverse the updated population. If there is an individual with a fitness value better than best_fitness, update best_solution to that individual and best_fitness to the corresponding fitness value. Iteration count update: epoch = epoch + 1, return to step 3 and continue the loop.

[0034] Step 8: Algorithm Termination When the number of iterations reaches max_epoch, the loop ends and the global optimal solution best_solution and the optimal fitness value best_fitness are output.

[0035] The improved hybrid gray wolf whale optimization algorithm (TCGWOWOA) based on Tent chaotic sequences and Cauchy mutations is as follows: Figure 1-3 As shown in the figure, in the field of marine permanent magnet synchronous motor control, the fusion characteristics of this algorithm demonstrate its significant adaptability: marine motors need to face scenarios of sudden load changes and complex operating conditions during navigation. TCGWOWOA, by leveraging the initial population diversity of the improved Tent mapping, can quickly cover the solution space of control parameters, avoiding control inaccuracies caused by initial parameter deviations. Simultaneously, the synergistic effect of its nonlinear convergence factor and Cauchy mutation can accurately optimize PI controller parameters during the dynamic adjustment phase of the motor, improving the dynamic response speed of speed and torque, and deepen local optimization during steady-state operation, reducing motor torque ripple and energy consumption. Compared with traditional control algorithms, TCGWOWOA can more efficiently balance the control accuracy and operational stability of marine motors, providing a superior algorithmic solution for the reliable and energy-saving operation of ship propulsion systems. This process also intuitively demonstrates the adaptability of the algorithm throughout the entire process from initialization to iterative optimization.

[0036] The control system of WOA-PI in PMSM, such as Figure 4 As shown, the block diagram of the PMSM speed control system based on TCGWOWOA-PI is as follows: Figure 5 As shown.

[0037] To verify the effectiveness of the designed control strategy, a simulation model was built in Matlab / Simulink software, and simulation experiments were conducted. The PMSM parameters were set as follows: stator resistance R = 2.875Ω, d-axis and q-axis inductance L = 8.5mH, permanent magnet chain ψf = 0.175, moment of inertia J = 0.003 kg·m², damping coefficient B = 0.008 N / (m / s), rated speed n = 1000 r / min, and number of pole pairs P = 4. Operating conditions were set as follows: a 5 N·m load was applied suddenly in 0.3 s, and the load was removed suddenly in 0.6 s. The simulation results are as follows. Figure 6-8 As shown. Figure 6 This is a graph showing the overall changes in motor speed under three control strategies: PI, WOA-PI, and TCGWOWOA-PI. Figure 7 This is a magnified view of a portion of the area affected by the sudden load. Figure 8 Enlarged view of a section where the load was suddenly unloaded.

[0038] Depend on Figure 6-8 It can be seen that during the startup phase of the control system, traditional PI control suffers from a lack of adaptive optimization capabilities in parameters, resulting in a speed overshoot of about 5% and a relatively long settling time. WOA-PI uses the Whale Optimization Algorithm to perform preliminary optimization of PI parameters, reducing the overshoot to about 3%, but the algorithm's global search accuracy is limited, and the improvement in stabilization speed is not significant. In contrast, TCGWOWOA-PI incorporates improved strategies (such as chaotic initialization and greedy mechanisms) into the basic WOA algorithm, greatly improving the accuracy and speed of parameter optimization. Therefore, the overshoot is only about 1.5%, and the time to reach stability is much shorter than the former two.

[0039] Under the conditions of a sudden 5 N·m load applied in 0.3 s and a sudden load unloaded in 0.6 s, the PI control has the largest speed fluctuation due to its weak disturbance rejection capability; although WOA-PI has a certain disturbance rejection capability, its algorithm has insufficient response speed to dynamic disturbances, resulting in a small overshoot; TCGWOWOA-PI, on the other hand, has almost no overshoot due to its better parameter adaptability and dynamic adjustment capability, and its stability against load disturbances is significantly superior, with overall performance far better than the first two control strategies.

[0040] This application, "Control Method for Marine Permanent Magnet Synchronous Motors Based on TCGWOWOA Algorithm," addresses the core requirements of marine PMSM control systems for wide operating conditions, multiple disturbances, and high precision. It overcomes existing technological bottlenecks through three-dimensional innovation in algorithms, control, and engineering, with its advanced features concentrated in four dimensions: 1. Algorithm architecture This innovative approach combines the strong global exploration capabilities of GWO with the fast local convergence characteristics of WOA, and couples the alpha wolf position update mechanism with the bubble net predation strategy to address the pain points of single algorithms being prone to getting stuck in local optima or slow convergence. It also features three key improvements: an improved Tent chaotic mapping to optimize population initialization and enhance parameter space coverage uniformity; a nonlinear convergence factor to dynamically balance global exploration and local exploitation; and a Cauchy mutation mechanism to help escape local optima. Benchmark tests have verified that the algorithm's optimization accuracy, convergence speed, and robustness are significantly superior to traditional and existing improved algorithms.

[0041] 2. Control logic aspects The multi-objective optimization function is reconstructed, integrating indicators such as speed error and torque fluctuation. By weighting, dynamic response and steady-state accuracy are balanced, solving the performance imbalance problem of traditional PID controllers. An adaptive triggering mechanism is designed to collect multi-source operating condition information in real time, start and stop optimization as needed, and adjust the iteration frequency, resolving the conflict between the real-time performance of intelligent algorithms and optimization performance, and adapting to embedded hardware constraints.

[0042] 3. Interference immunity and operating condition adaptation Leveraging Cauchy variation and nonlinear convergence factors, it rapidly responds to multi-source disturbances such as load abrupt changes and parameter perturbations, effectively suppressing speed and torque fluctuations even under extreme conditions. By improving the Tent mapping to enhance population diversity, it achieves stable control across all operating conditions, from low-speed heavy load to high-speed field weakening, maintaining high-precision operation without performance degradation during operating condition switching.

[0043] 4. Engineering Applications The algorithm features a lightweight design, optimizing iterative logic and parameter scale to meet the real-time control requirements of marine motors. Deployment requires only a software upgrade, eliminating the need for hardware modifications and reducing application costs. It is also compatible with multiple control architectures and can be extended to marine auxiliary motors, providing a unified solution for upgrading marine electric propulsion systems.

[0044] The core technology of the TCGWOWOA algorithm in this application provides key technical support for the enterprise's marine permanent magnet synchronous motor control series products. It accurately adapts to the actual operating requirements of ships under wide operating conditions and multiple disturbances, significantly improving the product's dynamic response speed, steady-state control accuracy, and anti-disturbance capability, solving the pain points of traditional products' performance imbalance and poor operating condition adaptability. The lightweight design of the algorithm requires no modification to existing hardware; product iteration can be enabled solely through software upgrades, reducing R&D and production costs. It is also compatible with multiple control architectures and can be extended to auxiliary motors, helping enterprises enrich their product portfolio, strengthen core competitiveness, and lay a solid foundation for the market promotion and technological upgrading of high-end marine motor control products.

[0045] The technology applied for is expected to bring significant economic benefits to enterprises: product iteration can be achieved through software upgrades, significantly reducing hardware modification costs and R&D investment. Optimized product control precision and anti-interference performance are improved, enhancing competitiveness in the high-end market and helping to increase product premium and market share. At the same time, it reduces energy consumption and maintenance costs of marine motors for end customers, increases customer loyalty, drives continuous revenue and profit growth, and forms a virtuous cycle of profitability.

[0046] The application of the technology in this application is expected to significantly improve the operational stability and safety of ship propulsion systems, reduce energy consumption and emissions, and promote the development of the shipping industry towards green and intelligent directions. By improving the control precision and anti-interference capability of motors, it reduces the risk of failure, ensures maritime transportation safety, promotes the technological upgrading of ship supporting industries, and enhances my country's competitiveness in the field of high-end ship electrical systems, thus having significant social benefits and industry demonstration value.

[0047] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0048] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A control method for marine permanent magnet synchronous motors based on a hybrid gray wolf-whale algorithm, characterized in that, Includes the following steps: S1. Construct a decoupled mathematical model of a marine permanent magnet synchronous motor in the dq synchronous rotating coordinate system, and adopt an id=0 vector control strategy to make the electromagnetic torque Te and the q-axis current iq linearly related. S2. Design a multi-objective optimization function J as the objective function of the hybrid gray wolf whale algorithm; S3. Execute the hybrid gray wolf whale algorithm to optimize the kp and ki parameters of the PI controller online or offline. The hybrid gray wolf whale algorithm includes the following optimization mechanisms: I. Initialize the population using an improved Tent chaotic map; II. The search behavior is dynamically adjusted using a nonlinear convergence factor; III. In each iteration, a WOA bubble mesh spiral approximation mechanism is introduced to update the α wolf position of the gray wolf algorithm; IV. At the end of each iteration, perform Cauchy mutation on the global optimal solution; S4. The optimal kp and ki parameters obtained by the hybrid gray wolf whale algorithm are loaded into the PI speed loop controller in real time, and combined with the id=0 current loop to form a dual closed-loop vector control system, so as to realize the precise tracking and disturbance rejection control of the speed of the marine PMSM.

2. The control method for marine permanent magnet synchronous motors based on the hybrid gray wolf-whale algorithm according to claim 1, characterized in that, The expression for the multi-objective optimization function J is as follows: ； in, For a given rotational speed, This refers to the actual rotational speed. This represents the torque ripple amplitude. To adjust the system time, This is the q-axis voltage. , , , All are weighted coefficients.

3. The control method for marine permanent magnet synchronous motors based on the hybrid gray wolf-whale algorithm according to claim 1, characterized in that, The process of initializing the population using the improved Tent chaotic map includes the following steps: when hour, ;when hour, or ,generate The chaotic sequence with enhanced ergodicity within the interval is mapped to the feasible region of parameters kp and ki through a linear transformation.

4. The control method for marine permanent magnet synchronous motors based on the hybrid gray wolf-whale algorithm according to claim 1, characterized in that, The nonlinear convergence factor a is used to dynamically adjust the search behavior, which is calculated by the following formula: ； ； ； in, Indicates the convergence factor; express The initial value; express The value at the end of the iteration; This indicates the total number of iterations preset by the algorithm; This indicates the current iteration round of the algorithm.

5. The control method for marine permanent magnet synchronous motors based on the hybrid gray wolf-whale algorithm according to claim 1, characterized in that, The WOA bubble network spiral approximation mechanism is introduced to update the α wolf position of the gray wolf algorithm in each iteration, and is calculated by the following formula: ； dα represents the dynamic distance between the current candidate solution and wolf α; xa represents the position of the alpha wolf α, i.e. the global optimal solution in the t-th iteration; x represents the position of the current candidate solution; c1 represents the random coefficient vector; p1 represents a random number uniformly distributed in the interval [0,1]; ρ represents the scaling coefficient; b represents the spiral shape parameter; l represents the random coefficient.

6. The control method for marine permanent magnet synchronous motors based on the hybrid gray wolf-whale algorithm according to claim 1, characterized in that: The position update process of the hybrid gray wolf whale algorithm in step S3 further includes: The three individuals with the best fitness in the population were selected as α wolf, β wolf, and δ wolf, respectively. For each individual, calculate using the Grey Wolf algorithm; The Whale Algorithm selects an update strategy with probability p∈[0,1]: when p<0.5 and |A|<1, it performs prey encirclement update; when p<0.5 and |A|≥1, it performs random search update; when p≥0.5, it performs spiral update. The results of the gray wolf algorithm and the whale algorithm are combined; Finally, boundary correction and Cauchy mutation are performed to complete a single iteration.

Images