A water jet propeller control method and system based on optimization algorithm

Abstract

The application provides a water jet propeller control method and system based on an optimization algorithm, which comprises the following steps: calculating a comprehensive index based on a speed error, a pressure fluctuation, a load change and a dynamic response characteristic, and dividing a working condition into three complexity levels; searching for an initial optimization parameter according to the complexity level; determining a fusion coefficient according to a control performance and a complexity change trend, and weighting and combining the initial parameter; outputting the fused parameter to a hydraulic controller after applying a single-step change rate constraint to the fused parameter, monitoring the performance, and iteratively optimizing the parameter if the performance is not up to the standard. The application constructs a working condition complexity quantification model based on multi-dimensional feature weighted fusion, takes the working condition complexity level as a core decision variable throughout the whole optimization process, and improves the environmental self-adaptive ability of the water jet propeller hydraulic control system in a nonlinear, time-varying and strong disturbance environment.

Description

Technical Field

[0001] This invention relates to the field of ship propulsion control technology, and in particular to a waterjet propulsion control method and system based on optimization algorithms. Background Technology

[0002] Waterjet propulsion, with its advantages of high efficiency, low noise, cavitation resistance, and high maneuverability, is gradually becoming the mainstream propulsion method for marine engineering equipment and ship power. Its hydraulic control system uses hydraulic cylinders to drive nozzles or rudders to change the direction of the water jet, enabling flexible ship steering. Simultaneously, it manipulates the reverse bucket to change the jet direction, giving the ship reverse thrust. In actual operation, the system is affected by factors such as changes in ship speed, external load disturbances, changes in the marine environment, and the nonlinearity of hydraulic components, causing the system parameters and dynamic characteristics to constantly change with the operating conditions.

[0003] Traditional PID control methods typically employ fixed control parameters, achieving good control performance under single design conditions. However, when the system enters a complex, variable operating environment, fixed parameters struggle to balance fast response, stability, and disturbance rejection. In recent years, intelligent optimization methods such as particle swarm optimization and genetic algorithms have been widely used for control parameter optimization. However, existing methods usually pre-specify a single optimization algorithm type, lacking the ability to recognize the complexity of the current system state and dynamically adjust the optimization strategy according to changes in operating conditions. When the operating environment changes, a single optimization algorithm is prone to problems such as decreased convergence speed, insufficient search capability, or getting trapped in local optima.

[0004] Some studies have attempted to improve optimization results through the fusion of multiple algorithms. However, the determination of the fusion coefficients is mostly based on the performance differences of the algorithms at the current moment, lacking forward-looking prediction of the changing trends of operating conditions. This results in a lag in the response of the fusion results to sudden changes in operating conditions. In addition, direct switching between different optimization algorithms often causes sudden changes in PID parameters, which in turn leads to hydraulic pressure shocks and actuator oscillations, affecting the safety and stability of system operation. Summary of the Invention

[0005] In view of the above, the main objective of this invention is to propose a waterjet propulsion control method and system based on an optimization algorithm to solve the aforementioned technical problems.

[0006] This invention proposes a waterjet propulsion control method based on an optimization algorithm, the method comprising the following steps: Step 1: Utilize the speed error, pressure fluctuation, load change, and dynamic response characteristics of the hydraulic actuator during the operation of the water jet propulsion system to calculate a comprehensive index characterizing the current operational complexity. Based on the comprehensive index value, classify the current operating condition into first complexity, second complexity, or third complexity to obtain the complexity level of the current operating condition. Step 2: When the current operating condition complexity level is the first level, perform a local search based on the current control parameters to obtain the local search results under the first complexity. When the current operating condition complexity level is the second level, start the improved adaptive genetic algorithm to perform a global parameter search to obtain the genetic algorithm search results under the second complexity. When the current operating condition complexity level is the third level, first start the improved adaptive genetic algorithm to perform a global coarse search, and continuously monitor the fitness improvement rate of each generation during the global coarse search. When the fitness improvement rate of the preset number of generations is lower than the preset threshold, terminate the global coarse search. Use the optimal sub-region found at the termination as the constraint boundary, and then start the improved particle swarm algorithm to perform a fine search within the constraint boundary to obtain the search results under the third complexity. Based on the search results, obtain the preliminary optimization parameters corresponding to the current operating condition complexity level. Step 3: Calculate the control performance of the preliminary optimized parameters based on the current operating conditions; determine the basic fusion coefficients based on the control performance and the complexity of the current operating conditions, and calculate the difference between the complexity of the current operating conditions and the complexity of the previous time step as the complexity change trend; generate a forward-looking correction term based on the complexity change trend, superimpose the forward-looking correction term onto the basic fusion coefficients, and use the superimposed fusion coefficients to perform a weighted combination of the preliminary optimized parameters to obtain the fused control parameters; Step 4: After applying a single-step rate of change constraint to the fused control parameters, output them to the hydraulic controller of the waterjet propulsion unit for closed-loop adjustment. Monitor whether the control performance after closed-loop adjustment meets the preset requirements. If it meets the preset requirements, keep the current fused control parameters as the final output control parameters. If it does not meet the preset requirements, re-acquire data and iteratively execute steps 1 to 3 to obtain the final output control parameters.

[0007] This invention also proposes a waterjet propulsion control system based on an optimization algorithm, the system comprising: The complexity level classification module is used for: By utilizing the speed error, pressure fluctuation, load change, and dynamic response characteristics of the hydraulic actuator of the water jet propulsion system during operation, a comprehensive index characterizing the current operational complexity is calculated. Based on the comprehensive index value, the current operating condition is divided into first complexity, second complexity, or third complexity to obtain the complexity level of the current operating condition. The preliminary parameter optimization module is used for: When the current operating condition complexity level is the first level, a local search is performed based on the current control parameters to obtain the local search results under the first complexity level. When the current operating condition complexity level is the second level, an improved adaptive genetic algorithm is started to perform a global parameter search to obtain the genetic algorithm search results under the second complexity level. When the current operating condition complexity level is the third level, the improved adaptive genetic algorithm is first started to perform a global coarse search, and the fitness improvement rate of each generation is continuously monitored during the global coarse search process. When the fitness improvement rate of the preset number of generations is lower than the preset threshold, the global coarse search is terminated. The optimal sub-region found at the termination is used as the constraint boundary, and then the improved particle swarm algorithm is started to perform a fine search within the constraint boundary to obtain the search results under the third complexity level. Based on the search results, preliminary optimization parameters corresponding to the current operating condition complexity level are obtained. The control parameter fusion module is used for: Based on the current operating conditions, calculate the control performance of the preliminary optimized parameters; determine the basic fusion coefficients based on the control performance and the current operating condition complexity, and calculate the difference between the current operating condition complexity and the operating condition complexity at the previous time step as the complexity change trend; generate a forward-looking correction term according to the complexity change trend, superimpose the forward-looking correction term onto the basic fusion coefficients, and use the superimposed fusion coefficients to perform a weighted combination of the preliminary optimized parameters to obtain the fused control parameters; The control parameter verification module is used for: After applying a single-step rate of change constraint to the fused control parameters, the output is sent to the hydraulic controller of the waterjet propulsion unit for closed-loop adjustment. The control performance after closed-loop adjustment is monitored to see if it meets the preset requirements. If it does, the current fused control parameters are kept as the final output control parameters. If it does not meet the preset requirements, the data is re-acquired and the complexity level division module is repeatedly executed to the control parameter fusion module to obtain the final output control parameters.

[0008] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention constructs a quantitative model of working condition complexity based on multi-dimensional feature weighted fusion, and takes the working condition complexity level as the core decision variable throughout the entire optimization process, so that the control parameter optimization process always keeps in real time matching with the current actual operating state. This solves the problem that existing methods lack the ability to recognize the complexity of the system state, and improves the environmental adaptability of the water jet propulsion hydraulic control system in nonlinear, time-varying and strong disturbance environments. 2. This invention constructs a three-level progressive optimization decision tree based on different levels of working condition complexity. Under low complexity, it performs local fast search; under medium complexity, it initiates an improved genetic algorithm to explore global parameter combinations; and under high complexity, it triggers phased collaborative optimization of the genetic algorithm and particle swarm optimization algorithm. Furthermore, it adaptively triggers switching based on the fitness improvement rate during the iteration process of the genetic algorithm, so that the optimization strategy is dynamically adjusted with changes in working conditions. This avoids the convergence speed of a single algorithm decreasing or getting stuck in local optima under complex and variable working conditions, thereby improving the global optimization capability and parameter optimization efficiency in complex marine environments. 3. This invention uses the trend of changing working conditions as a forward-looking correction factor superimposed on the fusion coefficient, enabling the control system to adjust the optimization direction in advance according to the deterioration or mitigation trend of the operating environment. At the same time, the single-step rate of change constraint mechanism ensures the continuous and smooth transition of control parameters, effectively suppressing the hydraulic shock and mechanism oscillation caused by sudden changes in the fusion coefficient. While ensuring the rapid response of control parameters, it maintains the stability of system operation, realizing the smooth switching and high-quality control of the water jet propulsion hydraulic control system in the entire working condition range. Attached Figure Description

[0009] Figure 1 This is a flowchart of a waterjet propulsion control method based on an optimization algorithm proposed in this invention; Figure 2 This is a schematic diagram of the overall framework of a waterjet propulsion control system based on an optimization algorithm proposed in this invention. Detailed Implementation

[0010] Embodiments of the present invention are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and should not be construed as limiting the present invention.

[0011] These and other aspects of the embodiments of the present invention will become clear from the following description and accompanying drawings. In these descriptions and drawings, some specific embodiments of the present invention are specifically disclosed to illustrate some ways of implementing the principles of the embodiments of the present invention; however, it should be understood that the scope of the embodiments of the present invention is not limited thereto.

[0012] Please see Figure 1 This invention proposes a waterjet propulsion control method based on an optimization algorithm, which includes the following steps: Step 1: Utilize the speed error, pressure fluctuation, load change, and dynamic response characteristics of the hydraulic actuator during the operation of the water jet propulsion system to calculate a comprehensive index characterizing the current operational complexity. Based on the comprehensive index value, classify the current operating condition into first complexity, second complexity, or third complexity to obtain the complexity level of the current operating condition. In step 1, the speed error, pressure fluctuation, load change, and dynamic response characteristics of the hydraulic actuator of the waterjet propulsion system during operation are used to calculate a comprehensive index characterizing the current operational complexity. Based on the comprehensive index value, the current operating condition is divided into first complexity, second complexity, or third complexity to obtain the complexity level of the current operating condition. The specific steps are as follows: The absolute value of the control error is calculated based on the speed error, the standard deviation of the pressure fluctuation is calculated based on the pressure fluctuation, the relative amplitude of the load change is calculated based on the load change, and the error change rate and pressure change rate are calculated based on the dynamic response characteristics to obtain the original characteristics of the working condition in five dimensions. Preset initial weights are assigned to the original features of the five dimensions of the operating condition, and the original features of the five dimensions are weighted and summed using the preset initial weights to obtain the comprehensive index value at the current moment. Among them, the initial weights of the five dimensions are determined according to the degree of influence of each feature on the hydraulic control performance of the waterjet propulsion system. Control error and pressure fluctuation directly reflect the closed-loop control accuracy and hydraulic system stability, and are given higher weights; load change and dynamic change are used to describe the speed of external disturbance and state change, and are given medium weights; system uncertainty is used as a supplementary index of model bias and is given lower weights. In particular, the initial weights in this invention are set to 0.35, 0.25, 0.15, 0.5 and 0.1 respectively. Based on the real-time change of the original characteristics of each working condition relative to the historical average at the current moment, the preset initial weights of the original characteristics of each working condition corresponding to the comprehensive index value at the current moment are dynamically corrected, and all corrected weights are normalized to obtain the updated comprehensive index value. The updated comprehensive index value is compared with a preset first complexity threshold and a second complexity threshold. When the updated comprehensive index value is lower than the first complexity threshold, it is classified as the first complexity level; when the updated comprehensive index value is between the first complexity threshold and the second complexity threshold, it is classified as the second complexity level; and when the updated comprehensive index value is higher than the second complexity threshold, it is classified as the third complexity level, so as to obtain the current working condition complexity level. The first complexity threshold and the second complexity threshold are determined statistically based on historical operating data of the waterjet propulsion hydraulic system under different sea conditions and load conditions. In particular, in this invention, the first complexity threshold is set to 0.35 and the second complexity threshold is set to 0.65.

[0013] Step 2: When the current operating condition complexity level is the first level, perform a local search based on the current control parameters to obtain the local search results under the first complexity. When the current operating condition complexity level is the second level, start the improved adaptive genetic algorithm to perform a global parameter search to obtain the genetic algorithm search results under the second complexity. When the current operating condition complexity level is the third level, first start the improved adaptive genetic algorithm to perform a global coarse search, and continuously monitor the fitness improvement rate of each generation during the global coarse search. When the fitness improvement rate of the preset number of generations is lower than the preset threshold, terminate the global coarse search. Use the optimal sub-region found at the termination as the constraint boundary, and then start the improved particle swarm algorithm to perform a fine search within the constraint boundary to obtain the search results under the third complexity. Based on the search results, obtain the preliminary optimization parameters corresponding to the current operating condition complexity level. In step 2, preliminary optimization parameters corresponding to the complexity level of the current operating condition are obtained based on the search results. The specific steps are as follows: When the current operating condition complexity level is the first complexity, it is determined that the current operation is in a stable state. Local search is performed using the current control parameters to track parameter drift and obtain local search results under the first complexity. When the complexity level of the current working condition is the second complexity, if it is determined that the local search cannot cover the global optimal region, the improved adaptive genetic algorithm is started to perform a global parameter search in the entire search space to obtain the genetic algorithm search results under the second complexity. When the current working condition complexity level is the third complexity, it is determined that the current environment is a strongly disturbed coupling environment. Then, the phased collaborative optimization strategy of genetic algorithm and particle swarm algorithm is started to obtain the search results under the third complexity. The search results corresponding to the local search results under the first complexity, the genetic algorithm search results under the second complexity, or the search results under the third complexity are used as the initial optimization parameters. In the process of initiating the phased collaborative optimization strategy of genetic algorithm and particle swarm optimization algorithm, the improved adaptive genetic algorithm is first started to perform a global coarse search, and then the improved particle swarm optimization algorithm is started to perform a fine search within the constraint boundary. The specific steps are as follows: Based on collaborative optimization triggering decision, the improved adaptive genetic algorithm is launched to perform a global coarse search under the third complexity condition, so as to carry out the real-time iterative search process of the genetic algorithm. In the real-time iterative search process of the genetic algorithm, the fitness of the best individual in each generation is monitored generation by generation, and the relative improvement rate between the best fitness in the current generation and the best fitness in the previous iteration is calculated to obtain the fitness improvement rate sequence. When the fitness improvement rate of multiple consecutive generations is lower than the preset switching threshold, it is determined that the global coarse search has identified the effective sub-region of the optimal parameters, and a switching trigger signal is generated. The preset switching threshold is 0.01, that is, when the fitness improvement rate is lower than 1% for 5 consecutive generations, the switching is triggered. The above value can be fine-tuned during the debugging phase according to the actual system response characteristics. By switching the trigger signal, the genetic algorithm is terminated, and the local neighborhood of the best individual found by the current genetic algorithm is extracted as the prior constraint boundary. The improved particle swarm optimization algorithm is launched, and a preset proportion of particles are initialized within the prior constraint boundary, while the remaining particles are chaotically initialized in the entire original search space to obtain a hybrid initialized particle swarm. The preset proportion is set to 80%, that is, 80% of the particles are initialized within the prior constraint boundary to utilize the search results of the genetic algorithm, and the remaining 20% ​​of the particles are chaotically initialized in the entire original search space to take into account global diversity. Perform fine particle swarm search on the mixed initial particle swarm until convergence to obtain the search results under the third complexity.

[0014] Specifically, the improved adaptive genetic algorithm is initiated to perform a global parameter search. The execution process of the improved adaptive genetic algorithm includes the following steps: An initial population containing control parameters is randomly generated, where each set of control parameters is treated as a chromosome individual, to obtain the initial population for the genetic algorithm. It should be noted that the standard genetic algorithm is a global optimization algorithm that simulates natural selection and genetic mechanisms, and its core operations include selection, crossover, and mutation. The improvement of this invention lies in the adaptive allocation mechanism of crossover rate and mutation rate, as well as the soft catastrophe recovery mechanism after premature convergence detection. The fitness of each individual in the initial population of the genetic algorithm is calculated. Individuals are then sorted from highest to lowest fitness, and crossover and mutation rates are assigned based on their relative ranking to obtain the genetic operation parameters. Individuals with higher rankings receive the first crossover and first mutation rates, while those with lower rankings receive the second crossover and second mutation rates. The first crossover rate is lower than the second crossover rate, and the first mutation rate is lower than the second mutation rate. Specifically, the fitness function is defined as a quantitative index evaluating the quality of a set of control parameters. In this invention, the fitness function is constructed based on the system response characteristics, specifically as follows: ; in, Represents the fitness value. This represents the tracking error at time t. Indicates overshoot. Indicates the adjustment time. , and This represents the corresponding weighting coefficient; a larger fitness value indicates better control parameters. Crossover and mutation operations are performed based on genetic operation parameters to generate offspring individuals and obtain the current population. Different variable asynchronous lengths are set for the three dimensions of the control parameters: proportion, integral, and derivative. The sensitivity coefficients of each dimension are obtained by pre-calibrating through sensitivity analysis. The first variable asynchronous length is used for the dimensions with high sensitivity coefficients, and the second variable asynchronous length is used for the dimensions with low sensitivity coefficients. The first variable asynchronous length is less than the second variable asynchronous length, resulting in offspring individuals generated based on genetic operation parameters and differential variable asynchronous lengths. The individual with the highest fitness among the offspring of the current population is selected as the best individual of the current generation. The fitness of the best individual of the current generation is taken as the optimal fitness of the current generation. The change between the optimal fitness of the current generation and the optimal fitness of the previous generation is calculated. When the change is less than a preset premature convergence threshold for several consecutive generations, the current population is determined to be prematurely convergent. The best individual in the current generation is retained, and the remaining individuals are re-initialized using chaotic mapping. At the same time, the mutation rate is increased to obtain the population after soft catastrophe. The premature convergence threshold is set to 0.001, and the number of consecutive generations is set to 5. That is, when the change in the optimal fitness is less than 0.001 for 5 consecutive generations, soft catastrophe is triggered. The population after the soft catastrophe is used as the current population for the next iteration until the termination condition is met, at which point the improved adaptive genetic algorithm has finished executing.

[0015] Specifically, during the process of initiating the improved particle swarm optimization algorithm to perform a fine search within the constraint boundaries, the execution process of the improved particle swarm optimization algorithm also includes the following steps: The position of each particle in the initial swarm is generated by a chaotic mapping sequence to ensure that all particles are evenly distributed in the search space, resulting in a chaotically initialized particle swarm. It should be noted that the standard particle swarm optimization algorithm is an optimization algorithm based on swarm intelligence, which finds the optimal solution by simulating the flight of individuals in the solution space. Its core lies in the balance between individual memory and swarm sharing. The improvements of this invention are the chaotic initialization strategy, the inertial weight modulation based on swarm activity, and the boundary reflection strategy. Calculate the degree of dispersion of all particles in the particle swarm relative to the population mean in each dimension after chaotic initialization to obtain the population activity. The inertial weight is dynamically adjusted based on the population activity level, and the current working condition complexity is introduced as an external modulation coefficient during the adjustment process. Under the third complexity working condition, the decay time constant of the inertial weight is increased to prolong the maintenance generation of the strong exploration state, so as to obtain the modulated inertial weight. Obtain the preset initial values ​​of individual cognitive factors, social cognitive factors, and the total number of iterations. Make the individual cognitive factors decrease non-linearly from the initial value of individual cognitive factors as the iteration progresses, and make the social cognitive factors increase non-linearly from the initial value of social cognitive factors as the iteration progresses, so as to obtain the learning factor of the current iteration step. The velocity and position of each particle are updated based on the modulated inertial weights and the learning factor of the current iteration step. When a particle goes out of bounds, a random reflection strategy is used to bounce back along the boundary. At the same time, every preset number of iterations, a chaotic small perturbation with an amplitude decreasing with the number of iterations is applied to the global optimal particle to obtain the particle swarm iteration result. Repeat the particle swarm iteration until the termination condition is met to complete the execution of the improved particle swarm algorithm.

[0016] Step 3: Calculate the control performance of the preliminary optimized parameters based on the current operating conditions; determine the basic fusion coefficients based on the control performance and the complexity of the current operating conditions, and calculate the difference between the complexity of the current operating conditions and the complexity of the previous time step as the complexity change trend; generate a forward-looking correction term based on the complexity change trend, superimpose the forward-looking correction term onto the basic fusion coefficients, and use the superimposed fusion coefficients to perform a weighted combination of the preliminary optimized parameters to obtain the fused control parameters; In step 3, the basic fusion coefficients are determined based on control performance and current operating condition complexity, and the difference between the current operating condition complexity and the operating condition complexity at the previous time step is calculated as the complexity change trend. A forward-looking correction term is generated based on the complexity change trend, and this forward-looking correction term is superimposed on the basic fusion coefficients. The superimposed fusion coefficients are then used to weight and combine the preliminary optimization parameters to obtain the fused control parameters. The specific steps are as follows: Substitute the control parameters from different algorithms in the initial optimization parameters into the current operating conditions to calculate the tracking error, overshoot, settling time, pressure fluctuation, and control input change corresponding to each set of control parameters. After normalizing each index in the tracking error, overshoot, settling time, pressure fluctuation, and control input change, perform a weighted sum to obtain the comprehensive performance value of different algorithms. Sensitivity coefficients are determined based on the normalized differences between the comprehensive performance values ​​of different algorithms and the complexity of the current operating conditions; the higher the complexity of the current operating conditions, the larger the sensitivity coefficient; the sensitivity coefficients and normalized differences are substituted into the logical function to obtain the basic fusion coefficients; Determine whether the current time is the first execution time. If the current time is the first execution time, use the preset initial working condition complexity as the working condition complexity of the previous time. If the current time is not the first execution time, obtain the actual stored working condition complexity of the previous time to obtain the historical working condition complexity. The preset initial working condition complexity is set to 0, that is, the working condition of the previous time is in the lowest complexity state by default when the execution is first performed. The difference between the current operating condition complexity and the historical operating condition complexity is calculated as the complexity change trend, and a forward-looking correction term is obtained based on the complexity change trend; wherein, a positive forward-looking correction term is generated when the difference between the current operating condition complexity and the historical operating condition complexity is positive, and a negative forward-looking correction term is generated when the difference between the current operating condition complexity and the historical operating condition complexity is negative. The forward-looking correction term is superimposed on the base fusion coefficient to obtain the corrected fusion coefficient; The corrected fusion coefficients are smoothed and limited to obtain the final fusion coefficients. In the limiting process, the lower limit of the fusion coefficient is set to 0.10 and the upper limit is set to 0.90 to ensure that the final fusion coefficients are always within a reasonable range and to avoid parameter mutations caused by a single algorithm. The final fusion coefficients are used to weight and combine the corresponding parts of the initial optimization parameters to obtain the fused control parameters.

[0017] Specifically, the corrected fusion coefficients are smoothed and clipped. The specific steps are as follows: When the current operating condition complexity is the first complexity, a first smoothing coefficient is used; when the current operating condition complexity is the second complexity, a second smoothing coefficient is used; and when the current operating condition complexity is the third complexity, a third smoothing coefficient is used. The first smoothing coefficient is greater than the second smoothing coefficient, and the second smoothing coefficient is greater than the third smoothing coefficient, so as to obtain an adaptive smoothing coefficient. In particular, in this invention, the first smoothing coefficient is set to 0.95, the second smoothing coefficient is set to 0.80, and the third smoothing coefficient is set to 0.60. Obtain the corrected fusion coefficient and the comprehensive performance values ​​of different algorithms. Use the adaptive smoothing coefficient to perform an exponential weighted average of the final fusion coefficient and the corrected fusion coefficient at the previous time step to obtain the smoothed fusion coefficient. The absolute value of the normalized difference is obtained by numerical calculation based on the comprehensive performance of different algorithms. When the absolute value of the normalized difference is less than the preset dead zone threshold, the final fusion coefficient of the previous moment is kept unchanged and the current update is skipped to obtain the fusion coefficient after stability processing; otherwise, the smoothed fusion coefficient is used as the updated fusion coefficient to obtain the fusion coefficient after stability processing. The dead zone threshold is set to 0.05, that is, when the performance difference between the two algorithms is less than 5%, the performance of the two algorithms is considered to be close, and frequent adjustment of the fusion coefficient is meaningless. The fusion coefficient of the previous moment is maintained to prevent repeated fluctuations. The fusion coefficient after stability processing is limited to a preset minimum and maximum allowable value to complete the smoothing and limiting processing of the modified fusion coefficient.

[0018] Step 4: After applying a single-step rate of change constraint to the fused control parameters, output them to the hydraulic controller of the waterjet propulsion unit for closed-loop adjustment. Monitor whether the control performance after closed-loop adjustment meets the preset requirements. If it meets the preset requirements, keep the current fused control parameters as the final output control parameters. If it does not meet the preset requirements, re-acquire data and iteratively execute steps 1 to 3 to obtain the final output control parameters. The single-step rate of change constraint is applied as follows: The maximum permissible change for each of the three control parameters (proportional, integral, and derivative) is pre-calibrated. After obtaining the fused control parameters at each step, the absolute value of the difference between each parameter and the final output value at the previous moment is calculated. Each absolute value of the difference is compared with its corresponding maximum permissible change. If all differences do not exceed the limit, the fused control parameter is directly output. If any difference exceeds the limit, the single-step change of that parameter is scaled proportionally to within its maximum permissible change while maintaining the direction of change. The calibration method for the maximum permissible change is as follows: Under no-load conditions of the hydraulic system, a step change excitation is applied to the three control parameters (proportional, integral, and derivative), and the maximum rate of change under the condition that the actuator response is not distorted is recorded. This rate is multiplied by the control period and then by a safety factor to obtain the maximum permissible change for each parameter. Preferably, the safety factor is 0.8.

[0019] In step 4, after applying a single-step rate of change constraint to the fused control parameters, the output is sent to the hydraulic controller of the waterjet propulsion system for closed-loop adjustment. The control performance after closed-loop adjustment is monitored to see if it meets the preset requirements. If it does, the current fused control parameters are retained as the final output control parameters. If it does not meet the preset requirements, data is re-acquired and steps 1 to 3 are iteratively executed to obtain the final output control parameters. The specific steps are as follows: S1. The fused control parameters are applied to the hydraulic controller of the waterjet propulsion system. Under normal operating conditions, target speed change conditions, load change conditions, pressure disturbance conditions, and rapid turning conditions, the tracking error, overshoot, response time, steady-state error, pressure fluctuation, and control stability indicators are obtained to obtain multi-dimensional control performance test results. S2, compare the measured value of each dimension in the multi-dimensional control performance test results with its respective preset qualified threshold to obtain the dimension-by-dimensional comparison results; S3, when the measured values ​​of all dimensions in the dimensional comparison results meet the corresponding qualified thresholds, the current control parameter is deemed qualified and the current fused control parameter is kept as the final output control parameter, and a qualified judgment result is obtained; S4, When the measured value of any dimension in the dimensional comparison results does not meet the corresponding qualified threshold, obtain the weight value corresponding to each feature used to calculate the comprehensive index, as well as the control dimension corresponding to each feature; according to the degree to which the measured value of the control dimension corresponding to each feature deviates from the corresponding qualified threshold, increase the weight value corresponding to the unqualified dimension in the corresponding weight value to obtain the adjusted weight; S5, recalculate the adjusted weights as the weights of the corresponding features to obtain the adjusted comprehensive index of the current running complexity; use the adjusted comprehensive index of the current running complexity as the comprehensive index of the current running complexity, recollect data and execute steps 1 to 3 to obtain the updated fused control parameters. S6. Repeat steps S1 to S5 until the measured results of control performance in all dimensions meet their respective preset qualification thresholds to obtain the final output control parameters.

[0020] Please see Figure 2 This invention also provides a waterjet propulsion control system based on an optimization algorithm, the system comprising: The complexity level classification module is used for: By utilizing the speed error, pressure fluctuation, load change, and dynamic response characteristics of the hydraulic actuator of the water jet propulsion system during operation, a comprehensive index characterizing the current operational complexity is calculated. Based on the comprehensive index value, the current operating condition is divided into first complexity, second complexity, or third complexity to obtain the complexity level of the current operating condition. The preliminary parameter optimization module is used for: When the current operating condition complexity level is the first level, a local search is performed based on the current control parameters to obtain the local search results under the first complexity level. When the current operating condition complexity level is the second level, an improved adaptive genetic algorithm is started to perform a global parameter search to obtain the genetic algorithm search results under the second complexity level. When the current operating condition complexity level is the third level, the improved adaptive genetic algorithm is first started to perform a global coarse search, and the fitness improvement rate of each generation is continuously monitored during the global coarse search process. When the fitness improvement rate of the preset number of generations is lower than the preset threshold, the global coarse search is terminated. The optimal sub-region found at the termination is used as the constraint boundary, and then the improved particle swarm algorithm is started to perform a fine search within the constraint boundary to obtain the search results under the third complexity level. Based on the search results, preliminary optimization parameters corresponding to the current operating condition complexity level are obtained. The control parameter fusion module is used for: Based on the current operating conditions, calculate the control performance of the preliminary optimized parameters; determine the basic fusion coefficients based on the control performance and the current operating condition complexity, and calculate the difference between the current operating condition complexity and the operating condition complexity at the previous time step as the complexity change trend; generate a forward-looking correction term according to the complexity change trend, superimpose the forward-looking correction term onto the basic fusion coefficients, and use the superimposed fusion coefficients to perform a weighted combination of the preliminary optimized parameters to obtain the fused control parameters; The control parameter verification module is used for: After applying a single-step rate of change constraint to the fused control parameters, the output is sent to the hydraulic controller of the waterjet propulsion unit for closed-loop adjustment. The control performance after closed-loop adjustment is monitored to see if it meets the preset requirements. If it does, the current fused control parameters are kept as the final output control parameters. If it does not meet the preset requirements, the data is re-acquired and the complexity level division module is repeatedly executed to the control parameter fusion module to obtain the final output control parameters.

[0021] It should be understood that various parts of the present invention can be implemented in hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented in software or firmware stored in memory and executed by a suitable instruction execution system. For example, if implemented in hardware, as in another embodiment, it can be implemented using any one or a combination of the following techniques known in the art: discrete logic circuits having logic gates for implementing logical functions on data signals, application-specific integrated circuits (ASICs) having suitable combinational logic gates, programmable gate arrays (PGAs), field-programmable gate arrays (FPGAs), etc.

[0022] In the description of this specification, references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0023] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of the present invention. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of the present invention, and these modifications and improvements all fall within the scope of protection of the present invention. Therefore, the scope of protection of this patent should be determined by the appended claims.

Claims

1. A waterjet propulsion control method based on an optimization algorithm, characterized in that, The method includes the following steps: Step 1: Utilize the speed error, pressure fluctuation, load change, and dynamic response characteristics of the hydraulic actuator during the operation of the water jet propulsion system to calculate a comprehensive index characterizing the current operational complexity. Based on the comprehensive index value, classify the current operating condition into first complexity, second complexity, or third complexity to obtain the complexity level of the current operating condition. Step 2: When the current working condition complexity level is the first complexity, perform a local search based on the current control parameters to obtain the local search results under the first complexity. When the current working condition complexity level is the second complexity, start the improved adaptive genetic algorithm to perform a global parameter search to obtain the genetic algorithm search results under the second complexity. When the current working condition complexity level is the third complexity, first start the improved adaptive genetic algorithm to perform a global coarse search, and continuously monitor the fitness improvement rate of each generation during the global coarse search process. When the fitness improvement rate of the preset number of generations is lower than the preset threshold, terminate the global coarse search, take the optimal sub-region found at the termination as the constraint boundary, and then start the improved particle swarm algorithm to perform a fine search within the constraint boundary to obtain the search results under the third complexity. Preliminary optimization parameters corresponding to the current operating condition complexity level are obtained based on the search results; Step 3: Calculate the control performance of the preliminary optimized parameters based on the current operating conditions; The basic fusion coefficients are determined based on control performance and current operating condition complexity, and the difference between the current operating condition complexity and the operating condition complexity at the previous moment is calculated as the trend of complexity change. Based on the trend of complexity change, a forward-looking correction term is generated. The forward-looking correction term is superimposed on the basic fusion coefficient. The superimposed fusion coefficient is used to weight and combine the preliminary optimization parameters to obtain the fused control parameters. Step 4: After applying a single-step rate of change constraint to the fused control parameters, output them to the hydraulic controller of the waterjet propulsion unit for closed-loop adjustment. Monitor whether the control performance after closed-loop adjustment meets the preset requirements. If it meets the preset requirements, keep the current fused control parameters as the final output control parameters. If it does not meet the preset requirements, re-acquire data and iteratively execute steps 1 to 3 to obtain the final output control parameters.

2. The waterjet propulsion control method based on an optimization algorithm according to claim 1, characterized in that, In step 1, the speed error, pressure fluctuation, load change, and dynamic response characteristics of the hydraulic actuator of the waterjet propulsion system during operation are used to calculate a comprehensive index characterizing the current operational complexity. Based on the comprehensive index value, the current operating condition is divided into first complexity, second complexity, or third complexity to obtain the complexity level of the current operating condition. The specific steps are as follows: The absolute value of the control error is calculated based on the speed error, the standard deviation of the pressure fluctuation is calculated based on the pressure fluctuation, the relative amplitude of the load change is calculated based on the load change, and the error change rate and pressure change rate are calculated based on the dynamic response characteristics to obtain the original characteristics of the working condition in five dimensions. The original features of the five dimensions of the working conditions are assigned preset initial weights, and the original features of the five dimensions of the working conditions are weighted and summed using the preset initial weights to obtain the comprehensive index value at the current time. Based on the real-time change of the original characteristics of each working condition relative to the historical average at the current moment, the preset initial weights of the original characteristics of each working condition corresponding to the comprehensive index value at the current moment are dynamically corrected, and all corrected weights are normalized to obtain the updated comprehensive index value. The updated comprehensive index value is compared with the preset first complexity threshold and second complexity threshold. When the updated comprehensive index value is lower than the first complexity threshold, it is classified as the first complexity level. When the updated comprehensive index value is between the first complexity threshold and the second complexity threshold, it is classified as the second complexity level. When the updated comprehensive index value is higher than the second complexity threshold, it is classified as the third complexity level, so as to obtain the complexity level of the current working condition.

3. The waterjet propulsion control method based on an optimization algorithm according to claim 2, characterized in that, In step 2, preliminary optimization parameters corresponding to the complexity level of the current operating condition are obtained based on the search results. The specific steps are as follows: When the current operating condition complexity level is the first complexity, it is determined that the current operation is stable. Local search is performed using the current control parameters to track parameter drift and obtain local search results under the first complexity. When the complexity level of the current working condition is the second complexity, it is determined that the local search cannot cover the global optimal region. Therefore, the improved adaptive genetic algorithm is started to perform a global parameter search in the entire search space to obtain the genetic algorithm search results under the second complexity. When the current working condition complexity level is the third complexity, it is determined that the current environment is a strongly disturbed coupling environment. Then, the phased collaborative optimization strategy of genetic algorithm and particle swarm algorithm is started to obtain the search results under the third complexity. The search results corresponding to the local search results under the first complexity, the genetic algorithm search results under the second complexity, or the search results under the third complexity are used as the initial optimization parameters.

4. The waterjet propulsion control method based on an optimization algorithm according to claim 3, characterized in that, In the process of initiating the phased collaborative optimization strategy of genetic algorithm and particle swarm optimization algorithm, the improved adaptive genetic algorithm is first started to perform a global coarse search, and then the improved particle swarm optimization algorithm is started to perform a fine search within the constraint boundary. The specific steps are as follows: Based on collaborative optimization triggering decision, the improved adaptive genetic algorithm is launched to perform a global coarse search under the third complexity condition, so as to carry out the real-time iterative search process of the genetic algorithm. In the real-time iterative search process of the genetic algorithm, the fitness of the best individual in each generation is monitored generation by generation, and the relative improvement rate between the best fitness in the current generation and the best fitness in the previous iteration is calculated to obtain the fitness improvement rate sequence. When the fitness improvement rate is lower than the preset switching threshold for multiple consecutive generations in the fitness improvement rate sequence, it is determined that the global coarse search has identified the effective sub-region of the optimal parameters, and a switching trigger signal is generated. By switching the trigger signal, the genetic algorithm is terminated, and the local neighborhood of the best individual found by the current genetic algorithm is extracted as the prior constraint boundary. The improved particle swarm algorithm is started, and a preset proportion of particles are initialized within the prior constraint boundary. The remaining particles are chaotically initialized in the entire original search space to obtain a hybrid initialized particle swarm. Perform fine particle swarm search on the mixed initial particle swarm until convergence to obtain the search results under the third complexity.

5. The waterjet propulsion control method based on an optimization algorithm according to claim 4, characterized in that, In step 3, the basic fusion coefficient is determined based on the control performance and the current operating condition complexity, and the difference between the current operating condition complexity and the operating condition complexity at the previous moment is calculated as the complexity change trend. Based on the complexity change trend, a forward-looking correction term is generated. This forward-looking correction term is then superimposed on the basic fusion coefficients. The superimposed fusion coefficients are used to weight and combine the initial optimization parameters to obtain the fused control parameters. The specific steps are as follows: Substitute the control parameters from different algorithms in the initial optimization parameters into the current operating conditions to calculate the tracking error, overshoot, settling time, pressure fluctuation, and control input change corresponding to each set of control parameters. After normalizing each index in the tracking error, overshoot, settling time, pressure fluctuation, and control input change, perform a weighted sum to obtain the comprehensive performance value of different algorithms. Sensitivity coefficients are determined based on the normalized differences between the comprehensive performance values ​​of different algorithms and the complexity of the current operating conditions; the higher the complexity of the current operating conditions, the larger the sensitivity coefficient; the sensitivity coefficients and normalized differences are substituted into the logical function to obtain the basic fusion coefficients; Determine whether the current time is the first execution time. If the current time is the first execution time, use the preset initial working condition complexity as the working condition complexity of the previous time. If the current time is not the first execution time, obtain the actual stored working condition complexity of the previous time to obtain the historical working condition complexity. The difference between the current operating condition complexity and the historical operating condition complexity is calculated as the complexity change trend, and a forward-looking correction term is obtained based on the complexity change trend; wherein, a positive forward-looking correction term is generated when the difference between the current operating condition complexity and the historical operating condition complexity is positive, and a negative forward-looking correction term is generated when the difference between the current operating condition complexity and the historical operating condition complexity is negative. The forward-looking correction term is superimposed on the base fusion coefficient to obtain the corrected fusion coefficient; The corrected fusion coefficients are smoothed and clipped to obtain the final fusion coefficients; The final fusion coefficients are used to weight and combine the corresponding parts of the initial optimization parameters to obtain the fused control parameters.

6. The waterjet propulsion control method based on an optimization algorithm according to claim 5, characterized in that, The corrected fusion coefficients are smoothed and clipped. The specific steps are as follows: When the current operating condition complexity is the first complexity, the first smoothing coefficient is used; when the current operating condition complexity is the second complexity, the second smoothing coefficient is used; when the current operating condition complexity is the third complexity, the third smoothing coefficient is used. The first smoothing coefficient is greater than the second smoothing coefficient, and the second smoothing coefficient is greater than the third smoothing coefficient, so as to obtain an adaptive smoothing coefficient. Obtain the corrected fusion coefficient and the comprehensive performance values ​​of different algorithms. Use the adaptive smoothing coefficient to perform an exponential weighted average of the final fusion coefficient and the corrected fusion coefficient at the previous time step to obtain the smoothed fusion coefficient. The absolute value of the normalized difference is obtained by numerical calculation based on the comprehensive performance of different algorithms. When the absolute value of the normalized difference is less than the preset dead zone threshold, the final fusion coefficient of the previous moment is kept unchanged and the current update is skipped to obtain the fusion coefficient after stability processing; otherwise, the smoothed fusion coefficient is used as the updated fusion coefficient to obtain the fusion coefficient after stability processing. The fusion coefficient after stability processing is limited to a preset minimum and maximum allowable value to complete the smoothing and limiting processing of the modified fusion coefficient.

7. The waterjet propulsion control method based on an optimization algorithm according to claim 6, characterized in that, In step 2, the improved adaptive genetic algorithm is started to perform a global parameter search. The execution process of the improved adaptive genetic algorithm includes the following steps: An initial population containing control parameters is randomly generated, where each set of control parameters is treated as a chromosome individual, to obtain the initial population for the genetic algorithm; The fitness of each individual in the initial population of the genetic algorithm is calculated, and the individuals are sorted from high to low fitness. The crossover rate and mutation rate are assigned according to the relative ranking of each individual in the sorting to obtain the genetic operation parameters. Crossover and mutation operations are performed based on genetic parameters to generate offspring individuals, and the current population is obtained. The individual with the highest fitness among the offspring of the current population is selected as the best individual of the current generation. The fitness of the best individual of the current generation is taken as the optimal fitness of the current generation. The change between the optimal fitness of the current generation and the optimal fitness of the previous generation is calculated. When the change is less than the preset premature convergence threshold for several consecutive generations, the current population is judged to be premature. The best individual in the current generation is kept unchanged, and the remaining individuals are re-initialized using chaotic mapping. At the same time, the mutation rate is increased to obtain the population after soft catastrophe. The population after the soft catastrophe is used as the current population for the next iteration until the termination condition is met, at which point the improved adaptive genetic algorithm has finished executing.

8. The waterjet propulsion control method based on an optimization algorithm according to claim 4, characterized in that, During the process of initiating a refined search within the constraint boundaries using the improved particle swarm optimization algorithm, the execution process of the improved particle swarm optimization algorithm also includes the following steps: The position of each particle in the initial population is generated by using a chaotic mapping sequence so that all particles are uniformly distributed in the search space, resulting in a particle swarm after chaotic initialization. Calculate the degree of dispersion of all particles in the particle swarm relative to the population mean in each dimension after chaotic initialization to obtain the population activity. The inertial weight is dynamically adjusted based on the population activity level, and the current working condition complexity is introduced as an external modulation coefficient during the adjustment process. Under the third complexity working condition, the decay time constant of the inertial weight is increased to prolong the maintenance generation of the strong exploration state, so as to obtain the modulated inertial weight. Obtain the preset initial values ​​of individual cognitive factors, social cognitive factors, and the total number of iterations. Make the individual cognitive factors decrease non-linearly from the initial value of individual cognitive factors as the iteration progresses, and make the social cognitive factors increase non-linearly from the initial value of social cognitive factors as the iteration progresses, so as to obtain the learning factor of the current iteration step. The velocity and position of each particle are updated based on the modulated inertial weights and the learning factor of the current iteration step. When a particle goes out of bounds, a random reflection strategy is used to bounce back along the boundary. At the same time, every preset number of iterations, a chaotic small perturbation with an amplitude decreasing with the number of iterations is applied to the global optimal particle to obtain the particle swarm iteration result. Repeat the particle swarm iteration until the termination condition is met to complete the execution of the improved particle swarm algorithm.

9. A waterjet propulsion control method based on an optimization algorithm according to claim 8, characterized in that, In step 4, after applying a single-step rate-of-change constraint to the fused control parameters, the output is sent to the hydraulic controller of the waterjet propulsion system for closed-loop adjustment. The control performance after closed-loop adjustment is monitored to see if it meets the preset requirements. If it does, the current fused control parameters are retained as the final output control parameters. If it does not meet the preset requirements, data is re-acquired and steps 1 to 3 are iteratively executed to obtain the final output control parameters. The specific steps are as follows: S1 applies the fused control parameters to the hydraulic controller of the waterjet propulsion system. Under normal operating conditions, target speed change conditions, load change conditions, pressure disturbance conditions, and rapid turning conditions, the tracking error, overshoot, response time, steady-state error, pressure fluctuation, and control stability indicators are obtained to obtain multi-dimensional control performance test results. S2, compare the measured value of each dimension in the multi-dimensional control performance test results with its respective preset qualified threshold to obtain the dimension-by-dimensional comparison results; S3, when the measured values ​​of all dimensions in the dimension-by-dimensional comparison results meet the corresponding qualified thresholds, the current control parameter is deemed qualified and the current fused control parameter is kept as the final output control parameter, and a qualified judgment result is obtained; S4. When the measured value of any dimension in the dimension-by-dimensional comparison results does not meet the corresponding qualified threshold, obtain the weight value corresponding to each feature used to calculate the comprehensive index, as well as the control dimension corresponding to each feature; according to the degree to which the measured value of the control dimension corresponding to each feature deviates from the corresponding qualified threshold, increase the weight value corresponding to the unqualified dimension in the corresponding weight value to obtain the adjusted weight. S5, recalculate the adjusted weights as the weights of the corresponding features to obtain the adjusted comprehensive index of the current running complexity; use the adjusted comprehensive index of the current running complexity as the comprehensive index of the current running complexity, recollect data and execute steps 1 to 3 to obtain the updated fused control parameters. S6. Repeat steps S1 to S5 until the measured results of control performance in all dimensions meet their respective preset qualification thresholds to obtain the final output control parameters.

10. A waterjet propulsion control system based on an optimization algorithm, characterized in that, The system employs a waterjet propulsion control method based on an optimization algorithm as described in any one of claims 1 to 9, and the system comprises: The complexity level classification module is used for: By utilizing the speed error, pressure fluctuation, load change, and dynamic response characteristics of the hydraulic actuator of the water jet propulsion system during operation, a comprehensive index characterizing the current operational complexity is calculated. Based on the comprehensive index value, the current operating condition is divided into first complexity, second complexity, or third complexity to obtain the complexity level of the current operating condition. The preliminary parameter optimization module is used for: When the current operating condition complexity level is the first level, a local search is performed based on the current control parameters to obtain the local search results under the first complexity level. When the current operating condition complexity level is the second level, an improved adaptive genetic algorithm is started to perform a global parameter search to obtain the genetic algorithm search results under the second complexity level. When the current operating condition complexity level is the third level, the improved adaptive genetic algorithm is first started to perform a global coarse search, and the fitness improvement rate of each generation is continuously monitored during the global coarse search process. When the fitness improvement rate of the preset number of generations is lower than the preset threshold, the global coarse search is terminated. The optimal sub-region found at the termination is used as the constraint boundary, and then the improved particle swarm algorithm is started to perform a fine search within the constraint boundary to obtain the search results under the third complexity level. Based on the search results, preliminary optimization parameters corresponding to the current operating condition complexity level are obtained. The control parameter fusion module is used for: Based on the current operating conditions, calculate the control performance of the preliminary optimized parameters; determine the basic fusion coefficients based on the control performance and the current operating condition complexity, and calculate the difference between the current operating condition complexity and the operating condition complexity at the previous time step as the complexity change trend; generate a forward-looking correction term according to the complexity change trend, superimpose the forward-looking correction term onto the basic fusion coefficients, and use the superimposed fusion coefficients to perform a weighted combination of the preliminary optimized parameters to obtain the fused control parameters; The control parameter verification module is used for: After applying a single-step rate of change constraint to the fused control parameters, the output is sent to the hydraulic controller of the waterjet propulsion unit for closed-loop adjustment. The control performance after closed-loop adjustment is monitored to see if it meets the preset requirements. If it does, the current fused control parameters are kept as the final output control parameters. If it does not meet the preset requirements, the data is re-acquired and the complexity level division module is repeatedly executed to the control parameter fusion module to obtain the final output control parameters.

Images