Electromechanical system signal processing and energy conversion method
By employing multi-port signal processing and energy conversion methods, combined with smoothing functions and adaptive control, the problems of model parameter perturbation and external interference in electromechanical servo systems were solved, achieving balanced optimization of the system in dynamic and steady-state processes, and improving control accuracy and robustness.
Patent Information
- Application Number
- CN202511950240.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-12-23
- Publication Date
- 2026-02-06
AI Technical Summary
Existing technologies struggle to effectively address model parameter perturbations, unknown external disturbances, and input saturation constraints in electromechanical servo systems, resulting in insufficient control accuracy and robustness. Traditional control methods are prone to overshoot and oscillation during dynamic processes and lack energy management optimization.
By employing multi-port signal processing and energy conversion methods, a dynamic model is established, a smoothing function is introduced to handle input saturation, a collaborative optimization controller is designed, and a variable gain disturbance observer and adaptive control are combined to achieve rapid trajectory signal adjustment during the dynamic process and energy dissipation optimization during the steady-state process.
It improves the control accuracy and robustness of the electromechanical system, enables stable operation under different working conditions, reduces vibration and energy consumption, and improves the system's anti-interference performance and service life.
Smart Images
Figure CN121485533A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of electromechanical control technology, and in particular to a method for signal processing and energy conversion of electromechanical systems. Background Technology
[0002] In modern industrial automation fields such as robotics, electric vehicles, aerospace, and high-end manufacturing equipment, high-precision, high-dynamic-performance electromechanical servo systems play a central role. Permanent magnet synchronous motors (PMSMs), due to their high power density, high efficiency, and excellent control performance, have become the preferred servo drive devices in these fields. However, achieving high-performance control of PMSM servo systems faces numerous challenges.
[0003] First, the system's dynamic model is subject to parameter perturbations (such as changes in resistance, inductance, and moment of inertia), and it is inevitably affected by unknown external load disturbances during operation. These uncertainties significantly reduce the system's control accuracy and tracking performance. Second, due to physical limitations, the system's actuators (such as power inverters) suffer from input saturation constraints. When the control signal exceeds the actuator's output capacity, saturation nonlinearity problems arise, leading to system performance degradation and even instability. Traditional control methods (such as classic PI control) are inadequate in handling these complex problems.
[0004] Specifically, while PI controllers are simple in structure and easy to implement, their fixed gain makes it difficult to maintain optimal performance in both dynamic and steady-state processes. In dynamic processes, high gain set for speed can lead to overshoot, oscillation, and easily induce input saturation; in steady-state processes, gain set for accuracy and interference resistance can result in slow response. Furthermore, traditional PI control lacks a systematic mechanism for handling model uncertainties and unknown disturbances, resulting in limited robustness. Although some advanced control strategies, such as adaptive control, disturbance observers, and port Hamiltonian control, offer solutions from the perspectives of signal processing or energy management, they often focus only on one aspect and lack a synergistic optimization mechanism that can comprehensively address dynamic speed, steady-state accuracy, robustness, and energy management.
[0005] Therefore, there is an urgent need in this field for a novel control method that can comprehensively address model uncertainty, external disturbances, and input saturation constraints, and achieve a balance between dynamic response and steady-state accuracy, as well as signal tracking and energy dissipation, under all operating conditions of the system. Summary of the Invention
[0006] To address the aforementioned issues, this application provides a signal processing and energy conversion method for electromechanical systems. This method aims to solve comprehensive control challenges in electromechanical systems, such as model parameter perturbation, unknown external disturbances, and input saturation constraints. It enables rapid adjustment of the trajectory signal during dynamic processes, optimized control of energy dissipation, and precise trajectory tracking during steady-state processes. This ensures that the control variables change continuously, smoothly, and without vibration during full-condition operation, improving system robustness, anti-interference performance, and service life. Through a synergistic mechanism of signal processing and energy conversion control, it automatically balances the system's control accuracy and energy efficiency under different operating conditions.
[0007] This application provides a method for signal processing and energy conversion in an electromechanical system, the method comprising the following steps: From the perspective of signal and energy, the input-saturated uncertain electromechanical system is treated as a multi-port signal processing and energy conversion device. The first expression of the dynamic model is established. Based on the actuator input saturation problem in the actual system, a smoothing function is introduced to update the first expression to obtain the second expression. Based on the defined filtering error variable, an auxiliary system input saturation compensation mechanism is constructed to deal with the input saturation problem. A collaborative optimization controller is established; wherein the collaborative optimization controller is composed of a convex combination mechanism of a first controller based on signal processing and a second controller based on energy transformation; Based on the second expression and the auxiliary system input saturation compensation mechanism, a variable gain interference observer is established; wherein, the variable gain interference observer is used to observe the lumped unknown external interference and compensate it into the system; Based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external disturbances, a first controller is designed. Based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external disturbances, design a second controller; Substituting the designed first and second controllers into the convex combination mechanism, a collaborative optimization control input is obtained. This collaborative optimization control input is then input to the electromechanical system to achieve closed-loop collaborative optimization control.
[0008] As a preferred technical solution, the first expression of the dynamic model is: (1) In the formula, For state vectors, For having perturbation A smooth vector field, Represents the input saturation vector. To control the input gain matrix, Due to unknown external interference, For the set of real numbers,n and m It is a positive integer. State vector x The first derivative.
[0009] As a preferred technical solution, based on the effective gradient and considering the actuator input saturation problem in the actual system, a smoothing function is introduced to update the first expression, resulting in a second expression, including: The actuator input saturation problem in a real system can be described as follows: (2) In the formula, For practical design control input signals, and These are the upper and lower bounds of the physical limitations of the actuator, respectively. It is a saturation function. j It is an integer; By introducing a smoothing function, the actuator input saturation problem in the actual system can be approximated as follows: (3) In the formula, , Let `sign` be the smooth approximation function, and `sign` be the sign function. u M The boundary equilibrium function, erf It is the error function; The actuator input saturation problem in a real system can be re-expressed as the following smooth saturation function: (4) In the formula, This represents the approximation error of the saturated model, and , Let be an unknown positive constant, and satisfy the following conditions: ; Based on the smoothing saturation function and its resulting approximation error, the first expression is updated to obtain the second expression: (5) In the formula, f d For external lumped interference, , The approximation error generated by the smoothing saturation function This is a smooth approximation function.
[0010] As a preferred technical solution, the defined filtering error variable is: (6) In the formula, For filtering error variables, For trajectory tracking error, For the desired trajectory, For design parameters, Let t be an auxiliary vector, and t be time. It is a time variable; The auxiliary system input saturation compensation mechanism is expressed as follows: (7) In the formula, As an auxiliary variable, For saturation error, This is the gain matrix.
[0011] As a preferred technical solution, the convex combination mechanism is expressed as follows: (8) In the formula, It is the identity matrix. , , and The designed smooth switching function satisfies... , ; To collaboratively optimize the controller, As the first controller, As the second controller, g To control the input gain matrix; The smooth switching function and Represented as: (9) In the formula, For trajectory tracking error, The rate of change of the trajectory tracking error. and is the weighting factor, and e is the natural constant.
[0012] As a preferred technical solution, the variable gain interference observer is represented as follows: (10) In the formula, It's interference. The observation vector, It is the intermediate state vector of the designed variable gain disturbance observer. It is a nonlinear function. The time-varying gain matrix is based on the error. The derivative of the desired trajectory signal; nonlinear functions and error-based time-varying gain matrix Designed as follows: (11) In the formula, and The gain matrix is a constant unit. The time-varying gain matrix is determined. The range of adjustment amplitude, The minimum value of the time-varying gain matrix is determined, and satisfies... , , ; , and Here, is the gain coefficient, and diag is the symbol for the diagonal matrix; Define the interference observation error as The error dynamics of the variable gain disturbance observer can be described as follows: (12) In the formula, f ( x ) is a smooth vector field.
[0013] As a preferred technical solution, the design of the first controller based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external interference includes the following methods: Based on the input saturation compensation mechanism of the auxiliary system and unknown external disturbances, the error dynamics of the second expression can be described as follows: (13) In the formula, For having perturbation p A smooth vector field; Let the structural information , and Can be an unknown constant and known functions Extraction, i.e. ,and The first controller is designed as follows: (14) In the formula, , and The constant gain matrix and time-varying gain parameters are respectively designed to satisfy... , ; The first controller performs variable gain adaptive updates using the following formula: (15) In the formula, For virtual parameters, for The estimation, the estimation error is expressed as , For design parameters, , T is the matrix transpose. For design parameters, The function is known.
[0014] As a preferred technical solution, the design of the second controller based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external interference includes the following methods: Based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external disturbances, the error dynamics of the second expression are described as a dissipative EPH system model: (16) In the formula, It is a bounded and continuously differentiable Hamiltonian storage function. , Hamiltonian function The partial derivatives, Here is the damping matrix; By configuring the interconnect structure, damping injection, and energy shaping of the dissipative EPH system, we obtain: (17) In the formula The energy control feedback signal is perturbed by parameters. For parameter perturbation compensation adaptive control law; Introducing an interconnection structure , This enables the allocation of interconnection structures; among which, For the desired interconnection matrix, For interconnection matrix, Damping injection is achieved through a variable damping injection mechanism, which is expressed as follows: (18) In the formula, A virtual input signal is injected into the variable damping injection mechanism. and For state vectors, , , , and To design positive parameters, and Let be the dynamic initial value and steady-state value of the damped injection, respectively, for any bounded function. satisfy , ; Energy shaping is the process of adding energy to a system, i.e. ;in, Let Hamiltonian be the desired Hamiltonian function; The expected Hamiltonian function The construction is as follows: (19) In the formula, , For design parameters, the desired Hamiltonian function The first-order partial derivative is calculated as follows: ; Based on the dissipative EPH system model, the structure of the dissipative EPH system that completes interconnection structure allocation, damping injection, and energy shaping is described as follows: (20) In the formula, For the desired interconnection matrix, Let be the desired damping matrix; Decomposition and perturbation The relevant functions yielded the following: (twenty one) In the formula, , , and The perturbation quantity derived from the decomposition; Considering equations (16) and (20), we assume that there exists a function. satisfy: (twenty two) In the formula, It is an adaptive parameter vector; The adaptive compensation control and adaptive update law are designed as follows: (twenty three) In the formula, For adaptive compensation control law, For adaptive parameter vectors The estimate, For the designed gain parameters, , For the design of a smooth switching function; The expression for the second controller is determined based on adaptive compensation control and adaptive update law.
[0015] As a preferred technical solution, the expression for the second controller is: (twenty four).
[0016] As a preferred technical solution, the method is applied to a PMSM servo system, the dynamics of which are described as follows: (25) In the formula, , , and They are respectively Stator current and stator voltage on the shaft; , and These are the motor's mechanical angular velocity, rotor flux, and stator phase resistance, respectively. and For stator inductance, The coefficient of viscous friction is... For extreme logarithms, It is the moment of inertia; , , , and Uncertainty in system model parameters; For electromagnetic torque, For unknown load torque, , and It is an unknown external disturbance.
[0017] The electromechanical system signal processing and energy conversion methods according to the various schemes of this application have at least the following technical effects: This application utilizes errors and their rate of change to design a collaborative optimization function in a smooth convex combination mechanism, thereby improving the performance of the collaborative optimization control strategy. Secondly, it explicitly describes the input saturation constraint problem and designs an input saturation compensation mechanism to weaken or even eliminate its negative impact on the system. For model uncertainties and unknown external disturbances in the system, considering the concept of refined control, adaptive techniques and error-based variable gain disturbance observers are employed to suppress their negative effects on the system. Furthermore, based on signal processing principles, an adaptive PI (API) control algorithm is designed to improve the speed of dynamic adjustment in the system's trajectory tracking control; its structure is simple and easy to implement. Based on the concept of energy transformation, an adaptive (Error Port-controlled Hamiltonian, EPH) control method is designed, further enhancing its advantages in steady-state processes. Moreover, this application uses Lyapunov theory to systematically verify for the first time the stability of the closed-loop system employing the collaborative optimization control strategy, and that all signals are bounded. Finally, to verify the effectiveness of the proposed collaborative optimization control strategy, a permanent magnet synchronous motor (PMSM) servo system, which is widely used in many industrial fields such as robotics, industrial manufacturing, electric vehicles, transportation, and aerospace due to its high power density and high reliability, was used. The system was used to verify that the proposed method has important guiding and application value.
[0018] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description
[0019] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the following description of the embodiments taken in conjunction with the accompanying drawings, wherein: Figure 1 A flowchart of a signal processing and energy conversion method for an electromechanical system provided in this application embodiment; Figure 2 This is a schematic diagram of the collaborative optimization control strategy provided in the embodiments of this application; Figure 3 Weighting factors provided for embodiments of this application Three-dimensional curves of different smooth continuous switching functions; Figure 4 Provided for the embodiments of this application and Time-varying gain curve; Figure 5A schematic diagram of a PMSM servo system based on LINKS-RT provided in this application embodiment; Figure 6 Various response curves using PI-NSS, PI, API, EPH, and collaborative optimization methods are provided for the starting of fixed load interference in the embodiments of this application. Figure 7 Various response curves using PI-NSS, PI, API, EPH, and co-optimization methods are provided for constant value abrupt load disturbances in the embodiments of this application. Figure 8 Various response curves using PI-NSS, PI, APID, EPH, and collaborative optimization methods are provided for time-varying load disturbances in the embodiments of this application. Detailed Implementation
[0020] To enable those skilled in the art to better understand the technical solution of this application, the application will be described in detail below with reference to the accompanying drawings and specific embodiments.
[0021] This application provides a method for signal processing and energy conversion in an electromechanical system, such as... Figure 1 As shown, the method includes the following steps S10-S60.
[0022] S10: From the perspective of signal and energy, the input-saturated uncertain electromechanical system is treated as a multi-port signal processing and energy conversion device. The first expression of the dynamic model is established. Based on the actuator input saturation problem in the actual system, a smoothing function is introduced to update the first expression, resulting in the second expression. Based on the defined filtering error variable, an auxiliary system input saturation compensation mechanism is constructed to handle the input saturation problem.
[0023] This embodiment considers the impact of input saturation constraints and unknown external disturbances on the system. From the perspective of signal and energy, the input-saturated uncertain electromechanical system is treated as a multi-port signal processing and energy conversion device, and a dynamic model is established as follows: (1) In the formula, For state vectors, For having perturbation A smooth vector field, Represents the input saturation vector. To control the input gain matrix, Due to unknown external interference, For the set of real numbers, n and m It is a positive integer. State vector x The first derivative.
[0024] Property 1: Control input gain matrix Its generalized inverse matrix is described as , , and .
[0025] Property 2: Using constants The described parameter perturbation has a nominal value of zero, and the following properties always hold true. .
[0026] The actuator input saturation problem in a real system is described as... (2) In the formula, For practical design control input signals, and These are the upper and lower bounds of the physical limitations of the actuator, respectively. It is a saturation function. j It is an integer.
[0027] When control input signal satisfy or Frequent signal switching can easily lead to jitter, increase wear on mechanical transmission components, and even disrupt the stable operation of the system.
[0028] The above discussion addressed model parameter perturbation, unknown external disturbances, and input saturation issues in electromechanical systems. In the control system design section below, a collaborative optimization control scheme based on a smooth convex combination mechanism is proposed to solve these problems.
[0029] To weaken equation (2) describing the input signal in the input saturation phenomenon To address the wear effects on the actuator and mechanical transmission mechanism caused by frequent switching when reaching the actuator's critical limit, a smoothing function is introduced to approximate the input saturation phenomenon in equation (2) as follows. (3) In the formula, , Let `sign` be the smooth approximation function, and `sign` be the sign function. u M The boundary equilibrium function, erf This is the error function.
[0030] Furthermore, the non-smooth saturation phenomenon (2) is re-expressed as (4) In the formula, Let be the approximation error of the saturated model, and . Let be an unknown positive constant, and satisfy the following conditions: .
[0031] Consider the smooth saturation function (4) and the resulting approximation error. Equation (1) can be rewritten as (5) In the formula, , , . f d For external lumped interference, , The approximation error generated by the smoothing saturation function This is a smooth approximation function.
[0032] To address the impact of noise in the system's state signal measurement, the following filter error variable is defined. (6) In the formula, For trajectory tracking error, For the desired trajectory, For design parameters, It is a time variable. As an auxiliary vector, we construct the following auxiliary system input saturation compensation mechanism to handle the input saturation problem: (7) In the formula, As an auxiliary variable, For saturation error, This is the gain matrix; , , , It is a design parameter, and .
[0033] S20: Establish a collaborative optimization controller; wherein the collaborative optimization controller is composed of a convex combination mechanism of a first controller based on signal processing and a second controller based on energy transformation.
[0034] like Figure 2 As shown, the control strategy based on signal control processing and energy transformation designed in this embodiment aims to achieve: ① rapid adjustment of dynamic process trajectory signals; ② optimized control of steady-state process energy dissipation and accurate trajectory tracking; ③ continuous and stable change of control variables without vibration during full-condition operation.
[0035] Collaborative Optimization Controller Can be controlled by a signal processing-based controller and energy conversion-based controller The convex combination mechanism is constructed as follows: The smooth convex combination mechanism is designed as follows: (8) In the formula, It is the identity matrix. , , and The designed smooth switching function satisfies... , The input-saturated uncertain PMSM servo system employs a closed-loop structure with a collaborative optimization control strategy, such as... Figure 2 As shown.
[0036] When designing switching functions using the direct switching method, it's easy for the system to experience surges during operation, reducing component lifespan and even compromising overall system stability, especially under rapidly changing conditions. To address this issue, we consider modifying the switching function... and The design is intended to be smooth and continuous to avoid negative impacts such as oscillations and shocks caused by non-smoothness. Furthermore, trajectory tracking error is introduced into the smooth switching function. and its rate of change It can respond more effectively to changes in system state, track the desired trajectory more accurately, and reduce lag effects. More importantly, it can achieve a refined collaborative optimization process, allowing collaborative optimization control to fully leverage its significant advantages at different stages of system operation based on changes in trajectory error.
[0037] Based on trajectory error and error change rate Consider the types of exponential and non-exponential functions, and the smooth continuous switching function. The selection of [the function] can be seen from equation (8). It is only necessary to determine the smooth, continuous switching function based on the relationship between error and the rate of change of error. That's fine. Consider... Designed to be and The relevant exponential and non-exponential functions, i.e., Gaussian functions Normal distribution function Hyperbolic secant function and Laplace distribution function Cauchy distribution function arctangent function Its specific expression and curve trajectory are shown in Table 1 and Table 2, respectively. Figure 3 The information is provided in the text.
[0038] from Figure 3 As can be seen, compared with other continuous functions in Table 1, the Gaussian function... It possesses advantages such as high switching speed, high precision, high smoothness, and good overall performance, as shown in Table 2. Therefore, the smooth continuous switching function... and The formal design is (9) In the formula, and As a weighting factor, and , .
[0039] Table 1 Switching functions for different types
[0040] Table 2 Performance Comparison Results of Different Functions
[0041] Choosing a Gaussian function based on trajectory error and its rate of change as the smooth convex combination mechanism can more effectively achieve the synergistic effect of signal processing and energy transformation control. From the smooth convex combination mechanism (8) and the selected smooth switching function (9), it can be seen that when the trajectory error is large or the error rate of change is large, , , Signal processing-based control plays a dominant role, enabling the control objective to converge rapidly. When the error and its rate of change are small, , , Energy transformation-based control plays a dominant role, optimizing and regulating the system's input energy, output energy, and energy dissipation to improve control accuracy and reduce energy consumption. This synergistic mechanism, by combining signal processing and energy optimization strategies, can balance the system's accuracy and efficiency under different operating conditions, ensuring that the system exhibits good robustness and adaptability under various circumstances.
[0042] S30: Based on the second expression and the auxiliary system input saturation compensation mechanism, establish a variable gain interference observer; wherein, the variable gain interference observer is used to observe the lumped unknown external interference and compensate it into the system.
[0043] To improve the system's anti-interference performance, this embodiment designs an error-based variable-gain interference observer to observe lumped unknown external interference. And compensate it into the system.
[0044] Assumption 1: Interference in the system It is continuously differentiable, and its derivative is bounded, that is... , It is an unknown positive number.
[0045] Assumption 2: Parameter perturbation in the system The impact can be compensated for by the adaptive law designed in the following section, that is... .
[0046] Combining equations (5) and (7), the error-based variable gain disturbance observer is constructed as follows: (10) In the formula, It's interference. The observation vector. It is the intermediate state vector of the designed variable gain interference observer.
[0047] nonlinear functions and error-based time-varying gain matrix Designed for (11) In the formula, The gain matrix is determined The range of adjustment amplitude, This determines the minimum value of the gain matrix and satisfies... , , ; , and is the gain coefficient, and diag is the symbol for the diagonal matrix.
[0048] The proposed variable gain concept mainly includes two elements: constant elements and time-varying elements. The gain coefficient of the constant element ( , Throughout the control process, The gain remains constant. The time-varying unit automatically adjusts the gain based on the error and its rate of change. (See Table 2...) Figure 3 and Figure 4 It can be obtained , Therefore, when dynamic processes or disturbances occur, the error... and its rate of change As it increases, the gain follows. Increasing the tracking error promotes rapid convergence of interference observations, giving the system strong interference suppression capabilities and providing strong compensation during the system startup phase. and its rate of change When the steady-state process is reduced and there is no disturbance, the gain Consequently, the energy input is reduced automatically, thus lowering energy consumption.
[0049] Define the interference observation error as Therefore, the error dynamics of the disturbance observer can be described as follows: (12) In the formula, f ( x ) is a smooth vector field.
[0050] S40: Design the first controller based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external disturbances.
[0051] This embodiment uses an input-saturated PMSM servo system with parameter perturbations and unknown external interference as the input and output signal processing device. To achieve rapid adjustment of the trajectory tracking signal, an API controller (first controller) based on signal processing is designed. The API controller design has a simple structure and is easy to implement.
[0052] Based on the results of equations (5), (6), and (7), the error dynamics of system (5) can be described as follows: (13) In the formula, For having perturbation p A smooth vector field.
[0053] Assumption 3: Structural Information , and Can be an unknown constant ( ) and known functions Extraction, i.e. ,and .
[0054] Furthermore, the signal processing-based API controller is designed as (14) In the formula, and The constant gain matrix and time-varying gain parameters are respectively designed to satisfy... ,
[0055] To handle the effects of parameter perturbations, the variable gain adaptive update algorithm is implemented as follows: (15) In the formula, For virtual parameters, for The estimation, the estimation error is expressed as . For design parameters, , , For design parameters, The function is known.
[0056] S50: Design a second controller based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external disturbances.
[0057] This embodiment treats the system as an energy conversion device with multiple ports. An energy conversion-based control design is constructed using EPH control principles, adaptive and variable damping techniques to achieve optimized energy dissipation control and precise trajectory tracking.
[0058] Based on equations (5) and (7), equation (13) can be described as a dissipative EPH system model. (16) In the formula, It is a bounded and continuously differentiable Hamiltonian storage function. , Hamiltonian function The partial derivatives, Here is the damping matrix. The input signal for the energy control design.
[0059] By configuring the interconnect structure, damping injection, and energy shaping of the EPH system, it is possible to achieve (17) In the formula, The energy control feedback signal is perturbed by parameters. This is a parameter perturbation compensation adaptive control law.
[0060] Using property 2, the perturbation of system parameters can be represented by a constant. To describe it, its nominal value is usually zero, that is , , , .
[0061] Define the structure matrix of system formula (16) as follows: , . It is an internal interconnection matrix, and satisfies . It is the damping matrix in the system, and satisfies .
[0062] Property 1.3: Interconnection Matrix It is an antisymmetric matrix that satisfies The damping matrix has a positive semi-definite symmetric property, satisfying... .
[0063] (1) Interconnection structure and damping injection By introducing the required interconnect structure ,Right now This enables the allocation of interconnected structures. Damping injection technology involves injecting external damping. It can be adjusted to the desired damping structure. ,Right now , .
[0064] With increasingly stringent control requirements, traditional fixed-damping injection techniques are no longer sufficient. Therefore, this embodiment proposes a variable-damping injection mechanism in the following form. (18) In the formula, A virtual input signal is injected into the variable damping injection mechanism. and For state vectors, , , , and To design positive parameters. and These represent the dynamic initial value and steady-state value of the damped injection, respectively. For any bounded function... satisfy , .
[0065] Compared with constant damping injection, the proposed variable damping injection mechanism (18) can adjust the damping value during the startup process and accelerate the system response speed. During steady state, the injected damping is reduced to ensure steady state characteristics and reduce energy loss.
[0066] (2) Energy Formation Energy shaping is the process of adding energy to a system, i.e. This ensures the energy balance and stability of the closed-loop system.
[0067] The expected Hamiltonian function Constructed as (19) In the formula, , For design parameters. Desired Hamiltonian function. The first-order partial derivative can be calculated as .
[0068] Although the perturbation of system parameters can be achieved using a constant To describe this, its nominal value is usually zero, but the interconnect structure configuration matrix, desired damping structure, and energy shaping are all related to the perturbation. related.
[0069] Based on system formula (16), the structure of the dissipative EPH system that completes interconnect configuration, damping injection, and energy shaping is described as follows: (20) In the formula, For the desired interconnection matrix, Let be the desired damping matrix.
[0070] Decomposition and perturbation The relevant functions are obtained. (twenty one) In the formula, , , and The perturbation quantity is the decomposed component.
[0071] Consider equations (16) and (20), and assume that there exists a function satisfy (twenty two) In the formula, It is an adaptive parameter vector.
[0072] The adaptive compensation control and adaptive update law are designed as follows: (twenty three) In the formula, For adaptive compensation control law, For adaptive parameter vectors The estimate, For the designed gain parameters, , This is a function for smooth switching.
[0073] Furthermore, an adaptive EPH controller (second controller) based on energy transformation will be used. Designed for (twenty four) S60: Substitute the designed first controller and second controller into the convex combination mechanism to obtain the collaborative optimization control input, and input the collaborative optimization control input signal to the electromechanical system to realize closed-loop collaborative optimization control.
[0074] This embodiment uses a PMSM servo system to experimentally verify the proposed method. The dynamics of the PMSM servo system are described below. (25) In the formula, , , and They are respectively Stator current and stator voltage on the shaft; , and These are the motor's mechanical angular velocity, rotor flux, and stator phase resistance, respectively. and For stator inductance, The coefficient of viscous friction is... For extreme logarithms, It is the moment of inertia; , , , and Uncertainty in system model parameters; For electromagnetic torque, For unknown load torque, , and It is an unknown external disturbance.
[0075] The experimental platform for the PMSM servo system consists of a servo inverter, industrial computer, control box, motion control card transmission box, torque sensor, and 130MB150A surface-mount PMSM. The torque sensor is model YH502, with a range of [missing information]. Nm, power supply is DC 24 V, output is The control algorithm is implemented using Matlab modeling software, and its design is based on Simulink software. Code is generated through compilation, and the control algorithm is verified using the LINKS-RT real-time simulation software package. The structure of the PMSM servo system based on LINKS-RT is as follows: Figure 5 As shown. The experimental platform parameters are: , , , , , , The parameters of the drive motor and the load motor are the same.
[0076] To verify the effectiveness of the proposed method, PI control is used as a comparison method, and the speed loop controller and current loop controller are given below. (27) In the formula, .
[0077] The parameters of the proposed collaborative optimization control strategy are designed as follows: , , , , , , , , , , , , , , , The API control (14) and adaptive EPH control (24) use the same parameters. The parameters of the traditional PI controller (34) are... , .
[0078] This embodiment uses an unknown load disturbance variation as an example to verify the robustness and disturbance suppression performance of the proposed collaborative optimization control strategy. The forms of load disturbance considered in this embodiment are shown in Table 3. To effectively and qualitatively evaluate the performance of the proposed controller, several different performance indicators are used for verification, such as overshoot, settling time, steady-state error, q-axis peak current / voltage, disturbance recovery time, speed reduction / increase due to disturbance, and speed flutter amplitude under load disturbance.
[0079] Table 3 Different load interference forms
[0080] (1) Start-up fixed load interference suppression performance Considering the PMSM servo system during startup... Under fixed load disturbance operation, the speed, dq-axis current, and voltage response curves using PI-no-smooth saturation (PI-NSS), PI, API, EPH, and the proposed synergistic optimization control method are shown below. Figure 6As shown in the figure, the API method achieves steady-state stability faster than the PI-NSS and PI methods, with settling times of 0.15 s, 0.5 s, and 0.5 s, respectively. The PI-NSS method, in particular, exhibits significant pulse current and voltage during startup due to the lack of soft-start and smooth saturation processing. The steady-state speed tracking errors for the PI-NSS, PI, API, and EPH methods are -9 / 8 rpm, -9 / 8 rpm, -3 / 3 rpm, and -2 / 1 rpm, respectively. This demonstrates that the proposed collaborative optimization control strategy achieves faster dynamic process tracking trajectory adjustment compared to the energy-conversion-based adaptive EPH method, and higher steady-state process control accuracy compared to the signal processing-based API method. The response curves of the API, EPH, and proposed collaborative optimization control methods not only show no overshoot but also outperform the PI-NSS and PI methods in both dynamic and steady-state performance. Table 4 compares the performance indicators of the five different methods during the startup of the PMSM servo system under fixed load disturbance.
[0081] Table 4 Performance Indicators of Different Methods under Constant Load
[0082] (2) Sudden load interference suppression performance Considering PMSM servo systems in At time s, the load is 2 The situation of sudden disturbances. The speed, dq-axis current, and voltage response curves using PI-NSS, PI, API, EPH, and synergistic optimization control methods are shown below. Figure 7 As shown in the figure, when the load disturbance changes, the speed curves using PI-NSS and PI control both exhibit significant speed decreases and increases. The response curves using API based on signal processing, adaptive EPH based on energy conversion, and the collaborative optimization method show smaller speed decreases and increases, shorter recovery times, and smaller speed response curve fluctuations during load disturbances. The speed decreases and increases for PI-NSS, PI, API, EPH, and the collaborative optimization control method are -48 / 50 rpm, -35 / 38 rpm, -25 / 17 rpm, -25 / 16 rpm, and -25 / 17 rpm, respectively. Furthermore, compared to adaptive methods in other literature, the proposed collaborative optimization control strategy does not exhibit inrush current during startup. Compared to disturbance compensation methods, the system exhibits smaller speed response fluctuations and shorter recovery times when subjected to load disturbances. These results demonstrate that the proposed collaborative optimization control method significantly improves the system's robustness and anti-interference performance. A comparison of the performance indicators of different methods under sudden load disturbances is shown in Table 5.
[0083] (3) Time-varying load interference suppression performance Considering PMSM servo system uptime At time s, subjected to load The time-varying disturbance conditions are shown. The speed, dq-axis current, and voltage response curves using PI-NSS, PI, API, EPH, and the proposed synergistic optimization control method are as follows: Figure 8 As shown, after the occurrence of time-varying load disturbances, the speed response curves of PI-NSS and PI control exhibit obvious chattering. The speed response amplitudes of PI-NSS, PI, API, EPH, and the proposed collaborative optimization control method after being subjected to time-varying load disturbances are -15 / 15 rpm, -14 / 14 rpm, -5 / 4 rpm, -6 / 6 rpm, and -5 / 5 rpm, respectively. Furthermore, the speed response curves of API based on signal processing, adaptive EPH based on energy transformation, and the proposed collaborative optimization control method show smaller changes compared to the other two methods, reflecting better time-varying disturbance suppression performance. Compared to the disturbance observation methods in the literature, the proposed collaborative optimization control strategy exhibits smaller current fluctuations when the system is subjected to time-varying load disturbances. In addition, the speed response of the collaborative optimization control strategy shows faster trajectory signal dynamic adjustment speed compared to the API method based on signal processing, higher steady-state process control accuracy and better overall tracking performance compared to the adaptive EPH method based on energy transformation, and better time-varying load disturbance suppression performance than both PI-NSS and PI methods. Table 6 shows a comparison of the performance indicators of different methods under time-varying load disturbance conditions.
[0084] Table 6 Performance Indicators of Different Methods under Time-Varying Load Disturbances
[0085] The above analysis shows that, regardless of the system's operating conditions, the proposed collaborative optimization control strategy exhibits smaller speed tracking error in the steady-state process compared to the PI-NSS, PI, and API methods, and faster trajectory adjustment speed in the dynamic process compared to the EPH method. Furthermore, its current and voltage curves are more stable than those of the PI-NSS, PI, and API methods. In summary, the collaborative optimization control strategy achieves rapid trajectory signal adjustment during dynamic operation, optimized energy dissipation control and precise trajectory tracking during steady-state operation, and continuous and stable change of control variables without vibration throughout the entire operation, indirectly demonstrating certain energy-saving effects and possessing significant application prospects and value.
[0086] The above embodiments are only used to illustrate this application and are not intended to limit this application. Those skilled in the art can make various changes and modifications without departing from the spirit and scope of this application. Therefore, all equivalent technical solutions also fall within the scope of this application, and the patent protection scope of this application should be defined by the claims.
Claims
1. A method for signal processing and energy conversion in an electromechanical system, characterized in that, The method includes the following steps: From the perspective of signal and energy, the input-saturated uncertain electromechanical system is treated as a multi-port signal processing and energy conversion device. The first expression of the dynamic model is established. Based on the actuator input saturation problem in the actual system, a smoothing function is introduced to update the first expression, resulting in the second expression. Based on the defined filtering error variable, an auxiliary system input saturation compensation mechanism is constructed to handle the input saturation problem. A collaborative optimization controller is established; wherein the collaborative optimization controller is composed of a convex combination mechanism of a first controller based on signal processing and a second controller based on energy transformation; Based on the second expression and the auxiliary system input saturation compensation mechanism, a variable gain interference observer is established; wherein, the variable gain interference observer is used to observe the lumped unknown external interference and compensate it into the system; Based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external disturbances, a first controller is designed. Based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external disturbances, design a second controller; Substituting the designed first and second controllers into the convex combination mechanism, a collaborative optimization control input is obtained. This collaborative optimization control input is then input to the electromechanical system to achieve closed-loop collaborative optimization control.
2. The method according to claim 1, characterized in that, The first expression of the dynamic model is: （1） In the formula, For state vectors, For having perturbation A smooth vector field, Represents the input saturation vector. To control the input gain matrix, Due to unknown external interference, For the set of real numbers, n and m It is a positive integer. State vector x The first derivative.
3. The method according to claim 2, characterized in that, Based on the effective gradient and considering the actuator input saturation problem in the actual system, a smoothing function is introduced to update the first expression, resulting in a second expression, including: The actuator input saturation problem in a real system can be described as follows: （2） In the formula, For practical design control input signals, and These are the upper and lower bounds of the physical limitations of the actuator, respectively. It is a saturation function. j It is an integer; By introducing a smoothing function, the actuator input saturation problem in the actual system can be approximated as follows: （3） In the formula, , Let `sign` be the smooth approximation function, and `sign` be the sign function. u M The boundary equilibrium function, erf It is the error function; The actuator input saturation problem in a real system can be re-expressed as the following smooth saturation function: （4） In the formula, This represents the approximation error of the saturated model, and , Let be an unknown positive constant, and satisfy the following conditions: ; Based on the smoothing saturation function and its resulting approximation error, the first expression is updated to obtain the second expression: （5） In the formula, f d For external lumped interference, , The approximation error generated by the smoothing saturation function This is a smooth approximation function.
4. The method according to claim 3, characterized in that, The defined filtering error variable is: （6） In the formula, For filtering error variables, For trajectory tracking error, For the desired trajectory, For design parameters, Let t be an auxiliary vector, and t be time. It is a time variable; The input saturation compensation mechanism of the auxiliary system is expressed as follows: （7） In the formula, As an auxiliary variable, For saturation error, This is the gain matrix.
5. The method according to claim 1, characterized in that, The convex combination mechanism is represented as follows: （8） In the formula, It is the identity matrix. , , and The designed smooth switching function satisfies... , ; To collaboratively optimize the controller, As the first controller, As the second controller, g To control the input gain matrix; The smooth switching function and Represented as: （9） In the formula, For trajectory tracking error, The rate of change of the trajectory tracking error. and is the weighting factor, and e is the natural constant.
6. The method according to claim 4, characterized in that, The variable gain interference observer is represented as follows: （10） In the formula, It's interference. The observation vector, It is the intermediate state vector of the designed variable gain disturbance observer. It is a nonlinear function. The time-varying gain matrix is based on the error. The derivative of the desired trajectory signal; nonlinear functions and error-based time-varying gain matrix Designed as follows: （11） In the formula, and The gain matrix is a constant unit. The time-varying gain matrix is determined The range of adjustment amplitude, The minimum value of the time-varying gain matrix is determined, and satisfies... , , ; , and Here, is the gain coefficient, and diag is the symbol for the diagonal matrix; Define the interference observation error as The error dynamics of the variable gain disturbance observer can be described as follows: （12） In the formula, f ( x ) is a smooth vector field.
7. The method according to claim 6, characterized in that, Based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external disturbances, the design of the first controller can be achieved through the following methods: Based on the input saturation compensation mechanism of the auxiliary system and unknown external disturbances, the error dynamics of the second expression can be described as follows: （13） In the formula, For having perturbation p A smooth vector field; Let the structural information , and Can be an unknown constant and known functions Extraction, i.e. ,and The first controller is designed as follows: （14） In the formula, , and These are the designed constant gain matrix and time-varying gain parameters, respectively, satisfying... , ; The first controller performs variable gain adaptive updates using the following formula: （15） In the formula, For virtual parameters, for The estimate, the estimation error is expressed as , For design parameters, , T is the matrix transpose. For design parameters, The function is known.
8. The method according to claim 7, characterized in that, Based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external disturbances, the design methods for the second controller include: Based on the second expression, the auxiliary system input saturation compensation mechanism, and unknown external disturbances, the error dynamics of the second expression are described as a dissipative EPH system model: （16） In the formula, It is a bounded and continuously differentiable Hamiltonian storage function. , Hamiltonian function The partial derivatives, Here is the damping matrix; By configuring the interconnect structure, damping injection, and energy shaping of the dissipative EPH system, we obtain: （17） In the formula The energy control feedback signal is perturbed by parameters. For parameter perturbation compensation adaptive control law; Introducing an interconnection structure , This enables the allocation of interconnection structures; among which, For the desired interconnection matrix, For interconnection matrix, Damping injection is achieved through a variable damping injection mechanism, which is expressed as follows: （18） In the formula, A virtual input signal is injected into the variable damping injection mechanism. and For state vectors, , , , and To design positive parameters, and Let be the dynamic initial value and steady-state value of the damped injection, respectively, for any bounded function. satisfy , ; Energy shaping is the process of adding energy to a system, i.e. ;in, Let Hamiltonian be the desired Hamiltonian function; The expected Hamiltonian function The construction is as follows: （19） In the formula, , For design parameters, the desired Hamiltonian function The first-order partial derivative is calculated as follows: ; Based on the dissipative EPH system model, the structure of the dissipative EPH system that completes interconnection structure allocation, damping injection, and energy shaping is described as follows: （20） In the formula, For the desired interconnection matrix, Let be the desired damping matrix; Decomposition and perturbation The relevant functions yielded the following: （21） In the formula, , , and The perturbation quantity derived from the decomposition; Considering equations (16) and (20), we assume that there exists a function. satisfy: （22） In the formula, It is an adaptive parameter vector; The adaptive compensation control and adaptive update law are designed as follows: （23） In the formula, For adaptive compensation control law, For adaptive parameter vectors The estimate, For the designed gain parameters, , For the design of a smooth switching function; The expression for the second controller is determined based on adaptive compensation control and adaptive update law.
9. The method according to claim 8, characterized in that, The expression for the second controller is: （24）。 10. The method according to any one of claims 1 to 9, characterized in that, The method is applied to a PMSM servo system, the dynamics of which are described below: （25） In the formula, , , and They are respectively Stator current and stator voltage on the shaft; , and These are the motor's mechanical angular velocity, rotor flux, and stator phase resistance, respectively. and For stator inductance, The coefficient of viscous friction is... For extreme logarithms, It is the moment of inertia; , , , and Uncertainty in system model parameters; For electromagnetic torque, For unknown load torque, , and It is an unknown external disturbance.