Controller parameter tuning method, system, and torsional vibration suppression method for electric drives

By constructing standardized test condition sequences and optimizing algorithms to automatically tune controller parameters, the problem of low parameter tuning efficiency in electric drive torsional vibration suppression strategies is solved, achieving efficient and accurate torsional vibration suppression and engineering adaptability.

CN123113352APending Publication Date: 2026-11-17HEFEI UNIV OF TECH
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202611350141.5
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2026-09-02
Publication Date
2026-11-17

AI Technical Summary

Technical Problem

Existing electric drive torsional vibration suppression strategies suffer from low controller parameter tuning efficiency, poor repeatability, and a lack of systematic test conditions and quantitative indicators, making it difficult to comprehensively evaluate the adaptability of torsional vibration suppression strategies in complex engineering scenarios.

Method used

A standardized test condition sequence containing various time-varying disturbances and parameter perturbation boundaries is constructed. Multidimensional performance indicators are extracted through simulation and a comprehensive score is calculated. The controller parameters are automatically iteratively tuned using an optimization algorithm.

Benefits of technology

It improves the efficiency and accuracy of parameter tuning, provides systematic test conditions and quantitative indicators, realizes quantitative evaluation and engineering adaptability of torsional vibration suppression strategies, and reduces manual intervention.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN123113352A_ABST
    Figure CN123113352A_ABST
Patent Text Reader

Abstract

This invention relates to a controller parameter tuning method, system, and method for torsional vibration suppression in the field of torsional vibration suppression technology for electrically driven systems. The controller parameter tuning method includes: constructing a test condition sequence containing various time-varying disturbances and parameter perturbation boundaries; performing simulations based on the test condition sequence to obtain multi-dimensional performance indicators; weighted summing of the multi-dimensional performance indicators to obtain a comprehensive score; and using the comprehensive score as the objective function, iteratively obtaining the parameters of the electrically driven torsional vibration controller using an optimization algorithm. This invention, by constructing a standardized test condition sequence containing time-varying disturbances and parameter perturbation boundaries, can comprehensively simulate complex disturbance scenarios in actual engineering, improving the engineering adaptability of the evaluation. By normalizing and weighting the multi-dimensional performance indicators to calculate the comprehensive score, and automatically iteratively tuning the controller parameters based on the comprehensive score using an optimization algorithm, the efficiency and accuracy of parameter tuning are significantly improved, reducing manual intervention and increasing parameter tuning efficiency.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] This invention relates to the field of torsional vibration suppression technology, and in particular to a controller parameter tuning method, system, and torsional vibration suppression method for electric drive torsional vibration suppression. Background Technology

[0002] During operation, electric drive systems are prone to torsional vibration due to load variations, parameter perturbations, and external disturbances, affecting system stability and comfort. Existing torsional vibration suppression strategies typically rely on a torsional vibration controller to control the motor torque output and suppress torsional vibration in the electric drive system when encountering different types of disturbances. Therefore, the parameter tuning of the torsional vibration controller is crucial. However, existing parameter tuning often depends on testing under a single operating condition, such as using only step response or constant load conditions. Furthermore, parameter tuning often relies on human experience and manual adjustment. For example, before tuning the controller parameters, a system identification experiment is first conducted. A swept-frequency torque excitation is applied while the load acceleration response is collected, and the resonant frequency of the system under different inertias is obtained through spectrum analysis. The tuning process begins by debugging the PID controller, gradually adjusting the parameters at a nominal inertia of 1.1 kg·m², using step response and impact response as indicators. After 30 sets of experiments, a set of optimal parameters is found. Then, the load inertia is changed to 1.5 kg·m², and the above parameters are retested. If a significant performance degradation is found, gain scheduling is developed to address this issue. Parameter tuning was performed at two inertia points of 1.1 and 1.5 kg·m², obtaining two sets of parameters, which were then linearly interpolated to obtain the scheduling table. Furthermore, to further improve performance, a notch filter was added in the experiment. A fixed-frequency notch filter was connected in series with the original PID controller, and multiple sets of experiments were conducted to tune the controller parameters. It is evident that the existing electric drive torsional vibration suppression strategy suffers from low controller parameter tuning efficiency, poor repeatability, and a lack of systematic test conditions and quantitative indicators, making it difficult to comprehensively evaluate the adaptability of the torsional vibration suppression strategy in complex engineering scenarios. Summary of the Invention

[0003] To address the technical problems of low controller parameter tuning efficiency, poor repeatability, lack of systematic test conditions and quantitative indicators in existing electric drive torsional vibration suppression strategies, which makes it difficult to comprehensively evaluate the adaptability of torsional vibration suppression strategies in engineering practice, this invention provides a controller parameter tuning method, system, and torsional vibration suppression method for electric drives.

[0004] In a first aspect, the present invention provides a controller parameter tuning method for an electric drive torsional vibration suppression strategy, which is used for adaptive parameter tuning of the electric drive torsional vibration controller under all operating conditions, all suppression strategies, and the entire life cycle. The controller parameter tuning method includes the following steps: A test condition sequence containing various time-varying disturbances and parameter perturbation boundaries is constructed. Simulations are performed based on the test condition sequence to obtain response data for each test condition, and multidimensional performance indicators are calculated based on the response data to quantify the performance of the torsional vibration suppression strategy. The multidimensional performance indicators are weighted and summed to obtain a comprehensive score reflecting the overall performance of the torsional vibration suppression strategy across multiple dimensions. The comprehensive score is used as the objective function, and an optimization algorithm is used iteratively to obtain the parameters of the electrically driven torsional vibration controller.

[0005] Secondly, this invention proposes a controller parameter tuning system for an electric drive torsional vibration suppression strategy, which uses the controller parameter tuning method for the electric drive torsional vibration suppression strategy described in the first aspect. This controller parameter tuning system includes: a test condition construction module, a simulation extraction module, a comprehensive score calculation module, and a parameter tuning module. The test condition construction module is used to construct a test condition sequence containing various time-varying disturbances and parameter perturbation boundaries. The simulation extraction module is used to simulate the test condition sequence to obtain response data under each test condition, and calculate multi-dimensional performance indicators to quantify the performance of the torsional vibration suppression strategy based on the response data. The comprehensive score calculation module is used to weightedly sum the multi-dimensional performance indicators to obtain a comprehensive score reflecting the comprehensive performance of the torsional vibration suppression strategy across multiple dimensions. The parameter tuning module uses the comprehensive score as the objective function and iteratively obtains the parameters of the electric drive torsional vibration controller using an optimization algorithm.

[0006] Thirdly, the present invention provides a method for suppressing torsional vibration in an electric drive system, which uses the controller parameter tuning method of the electric drive torsional vibration suppression strategy in the first aspect to adaptively tune the parameters of the torsional vibration controller. The tuned torsional vibration controller controls the motor torque output and suppresses the torsional vibration of the electric drive system when encountering different types of disturbances.

[0007] Fourthly, the present invention provides a computer-readable storage medium having a computer program / instructions stored thereon. When executed by a processor, the computer program / instructions implement the steps of the controller parameter tuning method for the electrically driven torsional vibration suppression strategy in the first aspect.

[0008] The beneficial effects of this invention are as follows: This invention constructs a standardized test condition sequence that includes time-varying disturbances and parameter perturbation boundaries, enabling comprehensive simulation of complex disturbance scenarios in actual engineering projects and improving the engineering adaptability of the evaluation. By normalizing multi-dimensional performance indicators and calculating a weighted comprehensive score, a quantitative and objective evaluation of torsional vibration suppression strategies is achieved, avoiding the limitations of a single indicator. Based on the comprehensive score, an optimization algorithm is used to automatically iteratively tune the controller parameters, significantly improving the efficiency and accuracy of parameter tuning, reducing manual intervention, and increasing parameter tuning efficiency. Therefore, this invention provides a systematic test condition, quantitative indicator, and standardized, reusable parameter tuning platform. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a flowchart of the controller parameter tuning method for suppressing torsional vibration in electric drives; Figure 2 This is the architecture diagram of the controller parameter tuning system for electric drive torsional vibration suppression; Figure 3 This is a structural diagram of the test condition construction module. Detailed Implementation

[0011] The technical solutions in the embodiments of the present invention will be clearly and completely described below. 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.

[0012] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention pertains. The terminology used herein in the specification of this invention is for the purpose of describing particular embodiments only and is not intended to be limiting of the invention. The term "or / and" as used herein includes any and all combinations of one or more of the associated listed items.

[0013] This embodiment provides a controller parameter tuning method for an electric drive torsional vibration suppression strategy, which is used for adaptive parameter tuning of the electric drive torsional vibration controller under all operating conditions, all suppression strategies, and the entire life cycle. The electric drive torsional vibration suppression strategy includes several general suppression strategies, such as PID controllers, notch filters, LQR controllers, and adaptive controllers. Regardless of the suppression strategy chosen, the core of this controller parameter tuning method lies in constructing a core framework of "building a standardized test condition sequence—extracting multi-dimensional performance indicators—calculating a comprehensive score—automatically iteratively tuning parameters." By constructing a standardized test condition sequence, various engineering scenarios including time-varying disturbances and parameter perturbation boundaries can be simulated, thus providing a comprehensive testing environment for the torsional vibration suppression strategy. Multi-dimensional performance indicators are extracted through simulation, and a comprehensive score is calculated based on normalization and weighting. Then, an optimization algorithm is used to automatically iteratively tune the controller parameters, achieving full automation from operating condition construction to parameter optimization.

[0014] Please refer to Figure 1 The construction of the test condition sequence is the basis for parameter tuning. In this embodiment, the test condition sequence may include disturbance signals of various types of time-varying disturbances, such as step, sine, pulse and random disturbance time sequence combinations. These signals can simulate scenarios such as load changes, periodic fluctuations, instantaneous impacts and random noise that may occur in actual operation, as shown in Table 1.

[0015] Table 1: Physical meaning and parameter settings for each type of interference. Interference type Physical meaning Interference parameter settings Step Load changes Trigger a step jump of a specified amplitude within a set time period sine Motor eccentric rotation sine wave random Uneven road surface The amplitude is 0.5 N•m, and the random seed is 42. pulse Gear collision The amplitude is 0.2 N•m, the width is 0.2s, and the period is 0.4s.

[0016] Among these, step disturbances can simulate sudden load changes. For example, during the operation of an electric vehicle, when the driver accelerates or decelerates rapidly, the rapid change in the motor's output torque can trigger torsional vibration in the system. By constructing a step signal, the response speed and suppression capability of the torsional vibration suppression strategy to sudden operating conditions can be tested. Sinusoidal disturbances can simulate periodically changing loads. For example, when a vehicle is traveling on undulating roads or affected by periodic wind resistance, this disturbance will continuously excite the system. By varying the frequency and amplitude of the sinusoidal signal, the robustness and filtering characteristics of the strategy at different frequencies can be evaluated. Random disturbances can simulate unpredictable and complex disturbances in actual engineering, such as the superposition of multiple noise sources or uneven road surfaces. Random signals can comprehensively test the adaptability of the strategy in uncertain environments. Impulse disturbances can simulate short-duration, high-intensity impacts, such as the instant of gear meshing or road bumps. These impacts will excite the system's intrinsic modes. Pulse signals can be used to test the strategy's suppression effect on transient impacts. By combining these disturbance signals according to a preset time sequence, a test condition sequence covering time-varying disturbances and parameter perturbation boundaries can be generated. To further explain the test condition sequence, in one embodiment, the test condition sequence can first apply a step disturbance to simulate a sudden increase in load (testing response speed), then superimpose a sinusoidal disturbance to simulate periodic load changes (testing steady-state performance), followed by an impulse disturbance to simulate instantaneous impact (testing shock resistance), and finally introduce a random disturbance to simulate environmental noise (testing overall robustness). For example, a test condition sequence lasting 10 seconds can be generated, where a step disturbance with an amplitude of 10 Nm is applied for the first 2 seconds, followed by a sinusoidal disturbance with a frequency of 5 Hz and an amplitude of 5 Nm superimposed for 3 seconds, then an impulse disturbance with an amplitude of 20 Nm is applied for 1 second, and finally a random disturbance with an amplitude of 3 Nm is applied for 4 seconds. This combination method can comprehensively cover various typical disturbance scenarios that electric drive systems may encounter in actual operation, thereby comprehensively evaluating the response characteristics of torsional vibration suppression strategies under different disturbance types. The parameter perturbation boundaries can include variations within a certain range of key parameters such as motor parameters (e.g., inductance, resistance), mechanical parameters (e.g., moment of inertia, damping coefficient, stiffness), and controller parameters (e.g., gain, time constant). These can be incorporated into the test condition sequence through parameter scanning or Monte Carlo simulation. The setting of these parameter perturbation boundaries can be based on the system's physical constraints or historical operating data. These variations within a certain range can reflect parameter drift caused by factors such as temperature and aging during actual operation. By incorporating these parameter perturbation boundaries into the test conditions, the robustness of the suppression strategy to parameter uncertainties can be evaluated. By combining time-varying disturbances with parameter perturbation boundaries, standardized test condition sequences can realistically reflect the operating state of the electric drive system in complex engineering environments, providing a reliable benchmark for evaluating torsional vibration suppression strategies.Furthermore, to further expand the applicability of the controller parameter tuning method of this invention, it is not limited to the aforementioned perturbation signal forms. Other types of perturbation signals, such as ramp signals, square wave signals, or frequency sweep signals of specific frequencies, can be introduced according to actual engineering needs. These signals can also be combined according to different timing logics. The parameter perturbation boundary can also be adjusted according to different application scenarios. For example, in the field of electric vehicles, a corresponding parameter perturbation range can be set to address the perturbation characteristics of motor parameters with temperature changes. In the field of industrial transmission, a wider parameter perturbation boundary can be set to address situations with large variations in load inertia. By flexibly adjusting the composition of the test condition sequence and the parameter perturbation boundary, this method can adapt to electric drive systems of different fields and specifications, thereby improving the versatility and reliability of its engineering adaptability assessment.

[0017] After constructing the standardized test condition sequence, the simulation extraction phase begins, where simulations are run to extract multi-dimensional performance indicators used to quantify the performance of the torsional vibration suppression strategy. These multi-dimensional performance indicators include: root mean square torsional acceleration, impact intensity, control energy consumption, and settling time, which quantify the performance of the torsional vibration suppression strategy from different perspectives. Specifically, the root mean square torsional acceleration reflects the overall intensity of torsional vibration during operation; a smaller value indicates a better suppression effect. Impact intensity measures the severity of torque changes and is closely related to ride comfort. It is calculated by differentiating the acceleration signal to obtain the rate of change of acceleration. A lower value indicates that the suppression strategy generates less impact when dealing with sudden changes in operating conditions, resulting in higher ride comfort. In particular, the control energy consumption parameter among the performance indicators assesses the energy consumed during torsional vibration suppression, such as the energy of the controller output signal or the additional power consumption of actuators (such as motors). A lower value indicates higher energy efficiency of the strategy and less impact on system range or operating costs, which is of great significance for energy-saving design. It can be calculated based on the controller's output signal. Settling time characterizes the speed from disturbance occurrence to recovery to stability, reflecting the response capability of the suppression strategy. A shorter settling time indicates a faster response speed and better dynamic performance of the system. Of course, performance indicators are not limited to those mentioned above; other relevant indicators such as torsional peak value, overshoot, steady-state error, and frequency domain response characteristics can also be included to further enrich the evaluation dimensions. By quantitatively evaluating from multiple dimensions such as time domain, frequency domain, energy, and response speed, multi-dimensional performance indicators can comprehensively assess the performance of the suppression strategy in terms of engineering adaptability. Furthermore, the simulation system in this stage can employ a mathematical model or a high-fidelity simulation model based on the electric drive system. The electric drive system in the simulation model is simplified to a dual-inertia model, consisting of motor-side inertia and load-side inertia connected by an elastic shaft. The dynamic equation of this dual-inertia model is: Motor side: .

[0018] Load side: .

[0019] Shaft torque: .

[0020] In the formula, J m This represents the motor's inertia. ω m This represents the angular velocity of the motor. τ c To control torque. τ s This is the shaft torque. B m is the coefficient of viscous friction. F cm is the Coulomb friction coefficient at the motor end. sign (·) is a symbolic function. J l This represents the load inertia. ω l This represents the load angular velocity. τ d This is the disturbance torque. F cl is the Coulomb friction coefficient at the load end. k s This is the shaft stiffness coefficient. θ s This represents the difference in torsional angle between the motor and the load. c s is the damping coefficient. ω s This is the difference in torsional angular velocity between the motor and the load. θ m This refers to the motor angle. θ l The load angle is used. During the simulation, the response data for each test condition sequence can be recorded first, and the above performance indicators can be calculated based on the response data. For example, the root mean square of torsional acceleration can be calculated by performing root mean square calculation on the acceleration signal, while the impact can be obtained by filtering and peak detection of the torque change rate. The extraction process of these performance indicators can be fully automated, thus avoiding the subjectivity and errors of manual measurement and calculation.

[0021] A comprehensive score is calculated based on the extracted multi-dimensional performance indicators to quantify the overall performance of the torsional vibration suppression strategy. Specifically, each performance indicator is first normalized to a uniform scale, such as between 0 and 1, to eliminate the influence of different indicator dimensions and magnitudes. Normalization methods can include range normalization, Z-score normalization, or nonlinear mapping functions based on expert experience. Subsequently, weights can be assigned to each performance indicator according to engineering requirements. For example, in scenarios with high comfort and energy-saving requirements, impact intensity and energy consumption control can be assigned higher weights, such as 0.4 each. In this case, the root mean square of torsional acceleration and settling time can be assigned lower weights, such as 0.1 each. A comprehensive score is obtained by weighted summation of the multi-dimensional performance indicators based on the assigned weights. This score reflects the comprehensive performance of the torsional vibration suppression strategy across multiple dimensions. The lower the comprehensive score, the better the overall performance of the torsional vibration suppression strategy, providing a quantitative basis for subsequent parameter tuning. In other embodiments, the weights of each indicator in the weighted summation of multi-dimensional performance indicators can be determined based on expert experience, or using the analytic hierarchy process (AHP) or entropy weighting method.

[0022] After obtaining the overall score, an optimization algorithm is used to automatically iteratively tune the controller parameters to optimize the performance of the torsional vibration suppression strategy. In the optimization algorithm, the overall score is used as the objective function, and the algorithm iteratively searches the controller parameter space to find the parameter combination that yields the best overall score. For example, the optimization algorithm can employ particle swarm optimization (PSO) or Bayesian optimization, which are suitable for high-dimensional, non-convex optimization problems and do not require gradient information. PSO can simulate the foraging behavior of a flock of birds, searching for the target solution through information sharing and cooperation among individuals in the group. During parameter tuning, each particle represents a set of controller parameters (e.g., the proportional, integral, and derivative coefficients of a PID controller). The particle's position represents the parameter value, and the particle's velocity represents the direction and step size of parameter adjustment. The particles fly in the parameter space, and by iteratively updating the particle's velocity and position and evaluating the overall score corresponding to each particle, the particle swarm can gradually converge to the parameter combination with the lowest overall score. Bayesian optimization is suitable for scenarios with high evaluation costs (e.g., each simulation requires a long time). In Bayesian optimization algorithms, by constructing a probabilistic model of the objective function and selecting the next evaluation point based on the acquisition function, optimal parameters can be found within a relatively small number of iterations. The new controller parameters generated by the optimization algorithm are fed back into the simulation model, and the simulation extraction and comprehensive score calculation steps are repeated until the convergence condition is met. The convergence condition may include the change in the comprehensive score being less than a preset threshold or reaching the maximum number of iterations. Through automatic iteration using particle swarm optimization or Bayesian optimization algorithms, the controller parameter space can be searched efficiently to find the parameter combination that yields the best comprehensive score for torsional vibration suppression in actual electric drive systems, thereby achieving automatic tuning of the torsional vibration suppression strategy parameters. This automatic iterative process can significantly improve the efficiency and effectiveness of parameter tuning, avoiding the inefficiency and subjectivity of traditional manual adjustments. On the other hand, after the controller parameters are tuned, another set of independent test condition sequences can be used to verify the optimized controller until the verification results are consistent, ensuring its robustness under different operating conditions and further ensuring that the tuned parameters have good engineering adaptability.

[0023] In summary, the controller parameter tuning method for the electric drive torsional vibration suppression strategy in this invention, by constructing a standardized test condition sequence including time-varying disturbances and parameter perturbation boundaries, can comprehensively simulate complex disturbance scenarios in actual engineering, improving the engineering adaptability of the evaluation. Through multi-dimensional performance index normalization and weighted calculation of the comprehensive score, a quantitative and objective evaluation of the torsional vibration suppression strategy is achieved, avoiding the one-sidedness of a single index. Based on the comprehensive score, the controller parameters are automatically iteratively tuned using an optimization algorithm, significantly improving the efficiency and accuracy of parameter tuning, reducing manual intervention, and increasing parameter tuning efficiency. This provides a systematic test condition, quantitative index, and standardized, reusable parameter tuning platform, decoupling the complex torsional vibration suppression strategy development process into independently optimizable and replaceable modules, and enabling a comprehensive evaluation of the adaptability of the torsional vibration suppression strategy in actual engineering.

[0024] To further verify the advantages of this invention over existing technologies, a performance comparison was conducted with traditional experimental methods. First, an electric drive system torsional vibration test bench was constructed, and its main parameters are shown in Table 2.

[0025] Table 2: Main parameters of the experimental bench. project Specification project Specification Motor rated power 50kW Motor rated speed 3000rpm Load inertia simulation range <![CDATA[0.5~2.0 kg•m 2 ]]> Torque sensor range ±200 N•m Accelerometer range <![CDATA[±400 rad / s 2 ]]> Sampling frequency 10kHz

[0026] Table 2 shows the comparison of control performance after the controller parameters were tuned using traditional experimental methods and the method of this invention.

[0027] Table 3: Comparison of control performance between the method of the present invention and traditional experimental methods. index The method of this invention is optimal Traditional method optimal contrast <![CDATA[Torsional vibration RMS (rad / s 2) > 62.14 64.87 4.39% Control RMS (N•m) 126.82 129.36 2.00% Inhibition rate (%) 30.31 27.25 10.10% Adjustment time (s) 0.85 1.5 43.33%

[0028] As shown in Table 3, the torsional vibration RMS of the traditional experimental method is 64.87 rad / s², while the torsional vibration RMS of the method of the present invention is 62.23 rad / s², a difference of 4.23%. This indicates that the method of the present invention is basically consistent with the actual situation. At the same time, the adjustment time is shorter, only 0.85s, while the traditional experimental method requires 1.5s, resulting in a slower response speed, which is not conducive to repeated experiments.

[0029] The method of this invention is compared with traditional experimental methods in terms of development efficiency and cost, and the results are shown in Table 4.

[0030] Table 4: Comparison of development efficiency between the method of this invention and traditional experimental methods. Comparison items Platform Approach Traditional experimental methods Increase Total development time 8 hours 184 hours 95.65% Number of experiments 0 times 80 times 100% Total development cost Approximately 0 yuan 15,000 yuan 100% Types of development strategies 5 types 2 types 150%

[0031] As shown in Table 4, the method of this invention completed all the work from strategy selection and parameter optimization to robustness verification within 8 hours, while the traditional experimental method took 184 hours and cost as much as 15,000 yuan. The method of this invention saves more than 95% of the time and cost. In addition, the hybrid strategy finally adopted by the method of this invention achieved a torsional acceleration RMS of 62.14 rad / s² and a suppression rate of 30.31% under complete operating conditions, both of which are better than the 64.87 rad / s² and 27.25% obtained by the traditional method. At the same time, the method of this invention has a faster response time.

[0032] In another embodiment, a controller parameter tuning system for an electric drive torsional vibration suppression strategy is also proposed, which uses the controller parameter tuning method for the electric drive torsional vibration suppression strategy described in the above embodiments. Please refer to... Figure 2 The controller parameter tuning system includes: a test condition construction module, a simulation extraction module, a comprehensive score calculation module, and a parameter tuning module. The test condition construction module generates a standardized test condition sequence. This sequence can include time-varying disturbances and parameter perturbation boundaries, and can be passed to the simulation extraction module in the form of data files or signal streams. The simulation extraction module runs a simulation model of the electric drive system, records response data, and extracts multi-dimensional performance indicators. The comprehensive score calculation module receives these multi-dimensional performance indicators, performs normalization and weighted calculations, and outputs a comprehensive score. The parameter tuning module generates new controller parameters based on the comprehensive score using optimization algorithms and feeds these parameters back into the simulation model, forming a closed-loop iteration. The entire controller parameter tuning process requires no manual intervention, significantly improving the efficiency and objectivity of evaluation and tuning. Furthermore, the controller parameter tuning system may include a user interface to display test conditions, performance indicators, comprehensive scores, and parameter tuning results, facilitating engineers to monitor and adjust the evaluation process. Through this modular design, the controller parameter tuning system of the present invention can flexibly adapt to different electric drive systems and torsional vibration suppression strategies, and has high versatility and scalability.

[0033] Furthermore, the test condition construction module, as the starting point of the controller parameter tuning system, primarily functions to construct a standardized test condition sequence that includes time-varying disturbances and parameter perturbation boundaries. For example... Figure 3As shown, the test condition construction module includes a test condition sequence generation unit and a boundary condition setting unit. The test condition sequence generation unit can generate various types of disturbance signals, such as step signals, sinusoidal signals, pulse signals, and random signals, and combine these disturbance signals according to a preset time sequence to form a test sequence capable of simulating complex disturbances in real engineering environments. The boundary condition setting unit can be used to set the boundary range of parameter perturbations, such as the variation range of motor moment of inertia, the fluctuation amplitude of load torque, and the perturbation range of the system damping coefficient. The parameter perturbation boundary can be set according to different application scenarios or randomly generated according to a preset probability distribution function. By combining time-varying disturbances with parameter perturbation boundaries, this module can construct standardized test conditions covering a wide range of engineering scenarios, thus providing comprehensive and representative input conditions for subsequent simulation evaluation. In addition, the test condition construction module may also include a test condition library to store various predefined standard test condition templates. Users can select or combine these templates as needed to quickly generate the required test condition sequences.

[0034] The simulation extraction module is connected to the test condition construction module to receive a pre-constructed standardized test condition sequence and run simulations to extract multi-dimensional performance indicators. This module can incorporate a high-fidelity electric drive system simulation model, which may include a motor model, transmission system model, load model, and torsional vibration suppression controller model. Upon receiving the test condition sequence, the module executes simulation calculations sequentially to simulate the dynamic response of the electric drive system under specific disturbances and parameter perturbations. During the simulation, the module can collect key state variables of the electric drive system in real time, such as motor speed, torque, torsional acceleration, and controller output signals. Based on this collected response data, the module further calculates a series of multi-dimensional performance indicators, which can quantify the engineering adaptability of the torsional vibration suppression strategy from different perspectives. For example, the module can calculate the root mean square value of torsional acceleration to evaluate the overall suppression effect; calculate the impact to evaluate the smoothness of the system during transient processes; calculate control energy consumption to evaluate the energy efficiency of the suppression strategy; and calculate the settling time to evaluate the system response speed. In addition, the simulation extraction module can also support parallel simulation, that is, run multiple simulation tasks under different working conditions at the same time, thereby greatly improving the extraction efficiency of performance indicators.

[0035] The comprehensive score calculation module is connected to the simulation extraction module to receive the extracted multi-dimensional performance indicators and calculate the comprehensive score based on normalization and weighting. Since the dimensions and numerical ranges of different performance indicators may vary significantly—for example, the root mean square of torsional vibration acceleration might be measured in radians per second (RMS), while control energy consumption might be measured in joules (J)—directly comparing or weighting these indicators is unreasonable. The comprehensive score calculation module can first normalize each performance indicator, mapping it to a uniform numerical range, such as 0 to 1. Normalization methods can include maximum / minimum value normalization, Z-score standardization, or target value-based normalization. After normalization, the comprehensive score calculation module can further perform a weighted summation of the normalized performance indicators according to preset weighting coefficients to obtain a comprehensive score. These weighting coefficients reflect the relative importance of different performance indicators in engineering adaptability assessment. For example, in some application scenarios, torsional vibration suppression may be more important than control energy consumption, thus requiring a higher weight. In another embodiment, the comprehensive score calculation module can automatically adjust the weights of various performance indicators based on the severity of the current test conditions. For example, when the test conditions are in a high-frequency, large-amplitude disturbance phase, the weight of the root mean square of torsional acceleration can be automatically increased to emphasize the suppression capability of the strategy. Conversely, when the conditions are in a stable operating phase, the weight of control energy consumption can be increased to emphasize the economic efficiency of the strategy. This dynamic weighting mechanism enables the comprehensive score to more objectively reflect the overall performance of the strategy under all operating conditions. In another embodiment, the comprehensive score calculation module can also have an outlier removal function. For extreme indicator data caused by unstable simulation values, median filtering or statistical methods can be used for smoothing, thereby avoiding the misleading effect of outlier data on parameter tuning direction. In some embodiments, the comprehensive score calculation module can also compare the calculated comprehensive score with a preset performance benchmark and automatically generate a performance evaluation report, providing engineers with intuitive decision support.

[0036] The parameter tuning module is connected to the comprehensive score calculation module and is used to automatically iteratively tune the controller parameters based on the calculated comprehensive score using an optimization algorithm. The parameter tuning module can receive the comprehensive score as the value of the optimization objective function and use it to drive the optimization algorithm to search for a better combination of controller parameters. The optimization algorithm can include particle swarm optimization (PSO) or Bayesian optimization. In PSO, each particle represents a set of controller parameters. The particle swarm flies in the parameter space, updating its velocity and position by tracking the historical best positions of individuals and the group, thus gradually approaching the target solution. In Bayesian optimization, a probabilistic surrogate model of the objective function is constructed, and the next most promising sampling point is selected using a sampling function, thereby finding better parameters in fewer iterations. The parameter tuning module can output the new set of controller parameters generated by the optimization algorithm to the simulation extraction module, triggering a new round of simulation and comprehensive score calculation. Through this closed-loop iterative process, the controller parameters that result in a better comprehensive score can be automatically searched, thus achieving engineering-adaptive automatic optimization of the torsional vibration suppression strategy. In addition, the parameter tuning module also includes a parameter constraint unit, which is used to set the range of controller parameters to ensure that the optimization results conform to the actual engineering situation.

[0037] It is worth mentioning that this controller parameter tuning system also supports torsional vibration suppression evaluation extended to multivariable coupled systems. For example, in multi-motor driven systems, the torque distribution and mutual interference between motors can make torsional vibration phenomena more complex. The test condition construction module can generate complex sequences containing cooperative disturbances from multiple motors, while the simulation extraction module can extract coupling indicators such as multi-axis synchronization error and total system energy consumption. The parameter tuning module can simultaneously optimize the parameters of multiple controllers to achieve optimal global system performance. This ability to support complex systems makes this invention of significant application value in the field of high-end equipment manufacturing.

[0038] In the implementation of the controller parameter tuning system, the modules can be integrated into a single physical device, such as a high-performance computer or embedded controller. Alternatively, these modules can be distributed across different physical devices, communicating and collaborating via a network. For example, the test condition construction module and the comprehensive score calculation module can be deployed on a cloud server, while the simulation extraction module and the parameter tuning module can be deployed on a local test platform. This distributed architecture fully leverages the powerful computing capabilities of the cloud and the real-time advantages of the local platform. In practical engineering applications, the controller parameter tuning system can be used in the powertrain development phase of electric vehicles. Engineers can use the test condition construction module to simulate the vehicle's operation under extreme conditions such as icy roads, bumpy roads, and rapid acceleration and deceleration. The simulation extraction module can extract the impact of motor torque fluctuations on the drive shaft and the torsional acceleration perceived by passengers. The comprehensive score calculation module can calculate the comprehensive score by setting different weighting coefficients according to the vehicle's positioning (e.g., comfort or sport). The parameter tuning module can automatically optimize the gain parameters of the torsional vibration suppression controller, ensuring good ride comfort and power response under various road conditions. The controller parameter tuning system can also be applied in the field of industrial motor control, such as torsional vibration suppression of wind turbine drive trains, shaft torsional vibration suppression of ship propulsion systems, and vibration suppression of industrial robot joints.

[0039] In summary, the controller parameter tuning system for the electric drive torsional vibration suppression strategy proposed in this invention addresses the pain points of traditional evaluation methods, such as incomplete operating condition coverage, single indicators, and low parameter tuning efficiency, through modular design and automated processes. The test condition construction module ensures the comprehensiveness of the evaluation environment through a combination of time-varying disturbances and parameter perturbations. The simulation extraction module ensures the scientific depth of the evaluation through the extraction of multi-dimensional indicators. The comprehensive score calculation module ensures the intuitiveness of the evaluation results through normalization and weighted calculation. The parameter tuning module ensures the objectivity and efficiency of parameter optimization through automatic iteration of the optimization algorithm. These modules collaborate to form a closed-loop, intelligent parameter tuning platform, which can significantly improve the R&D quality and engineering adaptability of torsional vibration suppression strategies for electric drive systems, providing strong technical support for the continuous advancement of electric drive technology.

[0040] In another embodiment, a torsional vibration suppression method for an electric drive system is proposed, which uses the controller parameter tuning method of the electric drive torsional vibration suppression strategy in the above embodiments to adaptively tune the parameters of the torsional vibration controller. The tuned torsional vibration controller controls the motor torque output and suppresses the torsional vibration of the electric drive system when encountering different types of disturbances.

[0041] In another embodiment, a computer-readable storage medium storing a computer program / instructions is also proposed. When the computer program / instructions are executed by a processor, the steps of the controller parameter tuning method for the electrically driven torsional vibration suppression strategy in the above embodiments are implemented. The computer-readable storage medium may include, but is not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof.

[0042] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0043] The embodiments described above are merely illustrative of several implementations of the present invention, and while the descriptions are relatively specific and detailed, they should not be construed as limiting the scope of the invention patent. 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 all fall within the protection scope of the present invention. Therefore, the protection scope of this invention patent should be determined by the appended claims.

Claims

1. A method for tuning controller parameters for an electric drive torsional vibration suppression strategy, characterized in that, It is used for adaptive parameter tuning of electric drive torsional vibration controllers under all operating conditions, all suppression strategies, and the entire life cycle; the controller parameter tuning method includes the following steps: Construct a test condition sequence that includes multiple time-varying perturbations and parameter perturbation boundaries; The response data for each test condition is obtained by simulation based on the test condition sequence, and a multi-dimensional performance index for quantifying the performance of the torsional vibration suppression strategy is calculated based on the response data. The weighted summation of the multidimensional performance indicators yields a comprehensive score that reflects the overall performance of the torsional vibration suppression strategy across multiple dimensions. The parameters of the electric drive torsional vibration controller are obtained by using the comprehensive score as the objective function and iterative optimization algorithm.

2. The controller parameter tuning method for the electric drive torsional vibration suppression strategy according to claim 1, characterized in that, The test condition sequence is a time series combination including step, sine, pulse and random disturbance; The parameter perturbation boundary is introduced into the test condition sequence through parameter scanning or Monte Carlo simulation.

3. The controller parameter tuning method for the electric drive torsional vibration suppression strategy according to claim 1, characterized in that, The multidimensional performance indicators include: root mean square torsional acceleration, impact intensity, control energy consumption, and adjustment time.

4. The controller parameter tuning method for the electric drive torsional vibration suppression strategy according to claim 1, characterized in that, Simulations are performed using a mathematical model or a high-fidelity simulation model based on the electric drive system. The electric drive system in the simulation model is simplified to a dual-inertia model, which consists of the motor-side inertia and the load-side inertia connected by an elastic shaft.

5. The controller parameter tuning method for the electric drive torsional vibration suppression strategy according to claim 1, characterized in that, When calculating the overall score, each performance indicator is first normalized to a uniform scale, and then a weighted sum is performed. And / or, based on expert experience, or by using the analytic hierarchy process or entropy weighting method, determine the weights of each indicator in the weighted summation of multidimensional performance indicators.

6. The controller parameter tuning method for the electric drive torsional vibration suppression strategy according to claim 1, characterized in that, The optimization algorithm for calculating the parameters of the electric drive torsional vibration controller adopts either particle swarm optimization or Bayesian optimization algorithm.

7. The controller parameter tuning method for the electric drive torsional vibration suppression strategy according to claim 1, characterized in that, After the controller parameters are tuned, the optimized controller is verified using another set of independent test condition sequences until the verification results are consistent.

8. A controller parameter tuning system for an electric drive torsional vibration suppression strategy, characterized in that, It employs a controller parameter tuning method using the electric drive torsional vibration suppression strategy as described in any one of claims 1 to 7; the controller parameter tuning system includes: The test condition construction module is used to construct a sequence of test conditions that include various time-varying disturbances and parameter perturbation boundaries; The simulation extraction module is used to simulate the response data under each test condition based on the test condition sequence, and to calculate the multi-dimensional performance index for quantifying the performance of the torsional vibration suppression strategy based on the response data. The comprehensive score calculation module is used to weight and sum the multi-dimensional performance indicators to obtain a comprehensive score that reflects the overall performance of the torsional vibration suppression strategy in multiple dimensions. The parameter tuning module is used to obtain the parameters of the electric drive torsional vibration controller by using the comprehensive score as the objective function and iteratively optimizing the algorithm.

9. A method for suppressing torsional vibration in an electric drive system, characterized in that, It uses the controller parameter tuning method of the electric drive torsional vibration suppression strategy as described in any one of claims 1 to 7 to perform adaptive parameter tuning of the torsional vibration controller; the tuned torsional vibration controller controls the motor torque output and suppresses the torsional vibration of the electric drive system when encountering different types of disturbances.

10. A computer-readable storage medium having a computer program / instructions stored thereon, characterized in that, When the computer program / instruction is executed by the processor, it implements the steps of the controller parameter tuning method for the electric drive torsional vibration suppression strategy as described in any one of claims 1 to 7.