Permanent magnet motor prediction control method based on self-learning hyper-local model

Abstract

The invention provides a permanent magnet synchronous motor predictive control method based on a self-learning hyper-local model, and relates to the field of permanent magnet motor control. The method comprises the steps that a historical data queue of input and output states of a controlled motor is constructed, the queue meets a first-in first-out rule, and output of the queue is used for parameter self-learning and model prediction; according to a self-learning algorithm, lumped disturbance terms and parameters for controlling input gains in the rotating speed super-local model and the current super-local model are updated online; using the updated hyper-local model and the current state to realize robust predictive control of the controlled motor; wherein the lumped perturbation term and the control input gain are approximated by the neural network, and related stability conditions and analyses are provided at the same time. According to the method, only historical data is used for approaching the actual dynamic state of the system, and the negative influence on the control performance due to parameter change and unmodeled disturbance is relieved; the input gain and the observer bandwidth do not need to be designed and controlled according to prior physical parameters, and the application range of the robust algorithm is further widened.

Description

Technical Field

[0001] The present invention relates to the field of permanent magnet motor predictive control, and specifically provides a permanent magnet motor predictive control method and device based on a self-learning super-local model. Background Art

[0002] Finite set model predictive control (FCS-MPC) of permanent magnet motors has been widely used because of its simple logic, easy implementation in digital systems, fast dynamic response, and ability to handle multi-variable and multi-constrained systems. However, its performance depends on an accurate system model, and parameter changes and unmodeled dynamics can greatly affect control performance.

[0003] In the related art, the hyperlocal model combined with the observer is currently the main way to improve the robustness of the predictive control of traditional permanent magnet motors. However, the design value of the control input gain in the hyperlocal model is a fixed value and depends on the prior physical parameters. When the parameters change, the fixed control input gain will restrict the robustness of this type of control method. In addition, the bandwidth of the observer in the hyperlocal model needs to be manually adjusted by the designer and kept fixed. The setting of the bandwidth requires a trade-off between the robustness and sensitivity of the control method. Therefore, the selected fixed bandwidth is difficult to adapt to all working conditions, which limits the scope of application of this type of method and may lead to a decrease in control accuracy or system oscillation under complex working conditions. Summary of the invention

[0004] In order to address the deficiencies in the prior art, the purpose of the present application is to provide a permanent magnet motor predictive control method and device based on a self-learning super-local model. This method can effectively solve the problem of insufficient adaptability of the permanent magnet motor finite set predictive control method in the related technology to working conditions.

[0005] In a first aspect, the present application provides a permanent magnet motor predictive control method based on a self-learning hyperlocal model, the method comprising: A self-learning hyperlocal model is constructed for the permanent magnet motor. The self-learning hyperlocal model includes a speed hyperlocal model and a current hyperlocal model. The lumped disturbance terms and control input gains in the speed hyperlocal model and the current hyperlocal model are all self-learned online. The self-learning parts in the speed hyperlocal model and the current hyperlocal model are approximated by online neural networks or fuzzy logic. Construct a historical data queue of the input and output status of the permanent magnet motor. The historical data queue meets the first-in-first-out update rule. The input and output status includes speed, current and voltage. Collect the input and output status of the permanent magnet motor at the current moment and update the historical data queue; Using the data in the updated historical data queue and the first self-learning formula, updating the lumped disturbance term and the control input gain in the speed hyperlocal model, and using the data in the updated historical data queue and the second self-learning formula, updating the lumped disturbance term and the control input gain in the current hyperlocal model; Substituting the set reference speed into the updated speed super-local model to obtain the optimal reference current, or substituting the set reference speed into the updated speed super-local model to obtain the optimal reference current, and substituting the optimal reference current into the updated current super-local model to obtain the optimal input voltage; An optimal switch combination is selected according to an optimal reference current and its corresponding current cost function or according to an optimal input voltage and its corresponding voltage cost function to drive the permanent magnet motor to operate.

[0006] In one embodiment, the speed superlocal model is: ; in, represents the sampling period of the permanent magnet motor speed, express The predicted speed of the permanent magnet motor at time, express The speed of the permanent magnet motor collected at the moment, express The q-axis current collected at the moment, and denote the lumped disturbance term and control input gain of the speed hyperlocal model respectively; Specifically, and All are approximated by online neural network in the speed superlocal model. express In the local model of the moment speed The corresponding neural network weights, express In the local model of the moment speed The corresponding neural network weights, express The input vector of the neural network in the hyperlocal model of the speed at the moment; The current hyperlocal model is: ; in, represents the sampling period of the current, express The predicted current on the q axis at time, express The predicted current of the d-axis at time, express The d / q axis current collected at the moment, express The d / q axis voltage collected at the moment, and denote the lumped disturbance term and control input gain of the current hyperlocal model respectively; Specifically, and All are approximated by online neural networks in the current hyperlocal model. express In the hyperlocal model of the current The corresponding neural network weights, express In the hyperlocal model of the current The corresponding neural network weights, express The input vector of the neural network in the hyperlocal model of the moment current; Each neural network has one hidden layer. The input of the neural network is a vector composed of data in the historical data queue. The queue length of the historical data queue is selected according to the required control accuracy and hardware computing speed. Neural networks can all be replaced by fuzzy logic.

[0007] In one embodiment, the first self-learning formula is obtained by minimizing a first error, where the first error is an error between a predicted rotational speed and an actual rotational speed; The first self-learning formula is: ; in, express The predicted speed of the permanent magnet motor at time express The actual speed of the permanent magnet motor at the moment, Represents the learning rate of the first self-learning formula.

[0008] In one embodiment, the second self-learning formula is obtained by minimizing a second error, where the second error is an error between the predicted current and the actual current; The second self-learning formula is: ; in, express The predicted current of the permanent magnet motor at time, express The actual current of the permanent magnet motor at the moment, Represents the learning rate of the second self-learning formula.

[0009] In one embodiment, the stability condition and parameter selection condition of the first self-learning formula are as follows: For the speed hyperlocal model, the following first Lyapunov function is constructed: ; in, , , , and They are and The corresponding optimal weight; Substitute the first self-learning formula into the first Lyapunov function, and When satisfied, we get: ; in, represents the upper bound of the activation function in the neural network, is the upper bound of the fitting error of the neural network under the optimal weights, ; According to the approximation theorem, the upper bound of the fitting error is As the number of neurons in the hidden layer of the neural network increases, Satisfy and the first learning rate Design meets When , that is, it satisfies the final consistency stability.

[0010] In one embodiment, the stability condition and parameter selection condition of the second self-learning formula are as follows: For the current hyperlocal model, the following second Lyapunov function is constructed: ; in, , ; and Don't and The corresponding optimal weight; Substitute the second self-learning formula into the second Lyapunov function, and When satisfied, we get: ; in, represents the upper bound of the activation function in the neural network, is the upper bound of the fitting error of the neural network under the optimal weights, ; According to the approximation theorem, the upper bound of the fitting error is As the number of neurons in the hidden layer of the neural network increases, Satisfies and the design of the second learning rate satisfies When , that is, it satisfies the final consistency stability.

[0011] In one embodiment, the optimal reference current is calculated by: And the reference speed Substitute the updated speed superlocal model to obtain the optimal reference current; The calculation formula for the optimal reference current is: ; in, , , represents the optimal reference current, represents the current amplitude constraint, express The speed of the permanent magnet motor collected at the moment, express The input vector of the neural network in the hyperlocal model of speed at time instant, Indicates the sampling period of the permanent magnet motor speed; The current cost function is: ; in, , , express The input vector of the neural network in the hyperlocal model of current at time instant, represents the sampling period of the current, Indicates i The d / q axis voltage corresponding to each switch combination is: express The d / q axis current collected at the moment.

[0012] In one embodiment, the optimal input voltage is calculated as: ; in, , , represents the sampling period of the current, express The d / q axis current collected at the moment, express The input vector of the neural network in the hyperlocal model of current at time instant, express The optimal reference current at the moment, express The optimal input voltage of the d / q axis at the moment; The voltage cost function is: ; in, Indicates i The d / q axis voltage corresponding to each switch combination.

[0013] In a second aspect, the present application also provides a permanent magnet motor predictive control device based on a self-learning hyperlocal model, the device comprising: A construction module is used to construct a self-learning hyperlocal model for a permanent magnet motor. The self-learning hyperlocal model includes a speed hyperlocal model and a current hyperlocal model. The lumped disturbance term and the control input gain in the speed hyperlocal model and the current hyperlocal model are both self-learned online. The self-learning parts in the speed hyperlocal model and the current hyperlocal model are approximated by an online neural network or fuzzy logic. The construction module is also used to construct a historical data queue of the input and output states of the permanent magnet motor. The historical data queue satisfies a first-in-first-out update rule. The input and output states include speed, current, and voltage. The acquisition module is used to collect the input and output status of the permanent magnet motor at the current moment and update the historical data queue; An updating module, used to update the lumped disturbance term and the control input gain in the speed hyperlocal model by using the data in the updated historical data queue and the first self-learning formula, and to update the lumped disturbance term and the control input gain in the current hyperlocal model by using the data in the updated historical data queue and the second self-learning formula; The selection module is used to substitute the set reference speed into the updated speed super-local model to obtain the optimal reference current, or substitute the set reference speed into the updated speed super-local model to obtain the optimal reference current, and substitute the optimal reference current into the updated current super-local model to obtain the optimal input voltage; the selection module is also used to select the optimal switch combination according to the optimal reference current and its corresponding current cost function or according to the optimal input voltage and its corresponding voltage cost function to drive the permanent magnet motor to operate.

[0014] In a third aspect, the present application also provides a computer device including a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the permanent magnet motor predictive control method based on the self-learning hyperlocal model of the first aspect is implemented.

[0015] The above-mentioned permanent magnet motor predictive control method based on the self-learning super-local model includes: constructing a speed super-local model and a current super-local model, and the lumped disturbance terms and control input gains in the speed super-local model and the current super-local model are dynamically updated through an online self-learning mechanism; constructing a historical data queue that follows the first-in-first-out principle to store the input and output state data (including speed, current and voltage) of the permanent magnet motor, and updating it in real time; using the data in the updated historical data queue and the first self-learning formula to update the lumped disturbance terms and control input gains in the speed super-local model, and using the data in the updated historical data queue and the second self-learning formula to update the lumped disturbance terms and control input gains in the current super-local model; according to the optimal reference current and its corresponding current cost function or the optimal input voltage and its corresponding voltage cost function, selecting the optimal inverter switch combination to drive the motor to operate.

[0016] This scheme introduces the online self-learning mechanism into the hyperlocal model, uses only historical data to approximate the actual dynamics of the system, and realizes the online update of the control input gain and the lumped disturbance term, alleviating the negative impact of parameter changes and unmodeled disturbances on the control performance. This scheme does not need to design the control input gain and observer bandwidth based on physical information, reduces the dependence on prior motor parameters and manual experience, and can further improve the robustness and applicability of predictive control. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 is a flow chart of a permanent magnet motor predictive control method based on a self-learning super-local model in one embodiment; Figure 2 is a structural diagram of a permanent magnet motor predictive control method based on a self-learning super-local model in one embodiment; FIG3 (a) and FIG3 (b) are respectively a control effect diagram of the speed and current of a permanent magnet motor in one embodiment; Figure 4 A device diagram of a permanent magnet motor predictive control device based on a self-learning super-local model in one embodiment; Figure 5 FIG. 4 is a diagram showing the internal structure of a computer device in one embodiment. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. Unless otherwise defined, the technical terms or scientific terms involved in the present application should have the general meaning understood by people with general skills in the technical field to which the present application belongs.

[0019] In one embodiment, Figure 1 As shown, a permanent magnet motor predictive control method based on a self-learning super-local model is provided, and the method comprises the following steps: Step 101: construct a self-learning hyperlocal model for the permanent magnet motor, the self-learning hyperlocal model includes a speed hyperlocal model and a current hyperlocal model, the lumped disturbance terms and control input gains in the speed hyperlocal model and the current hyperlocal model are all self-learned online, and the self-learning parts in the speed hyperlocal model and the current hyperlocal model are approximated by an online neural network or fuzzy logic.

[0020] A self-learning hyperlocal model is constructed for permanent magnet motors. The self-learning hyperlocal model includes a speed hyperlocal model and a current hyperlocal model. The speed hyperlocal model includes a speed equation. The speed hyperlocal model is used to describe the relationship between the speed of the permanent magnet motor and the q-axis current. The speed hyperlocal model includes a lumped disturbance term and a control input gain. The current hyperlocal model includes a current equation. The current hyperlocal model is used to describe the relationship between current and voltage. The current hyperlocal model includes a lumped disturbance term and a control input gain. Among them, the control input gain is used to describe the gain of the system input quantity, and the lumped disturbance term is used to describe the combined influence of internal and external disturbances. Among them, the internal disturbance includes the change of the permanent magnet motor parameters, and the external disturbance includes the sudden change or fluctuation of the load, the change of environmental conditions, etc.

[0021] The speed hyperlocal model is used to predict the change of motor speed. Through the online self-learning mechanism, the lumped disturbance term and control input gain can be updated in real time to adapt to the actual operating status. The online self-learning mechanism can use historical state data and real-time sampled state data to minimize the difference between the predicted value and the actual value, and continuously adjust the lumped disturbance term and control input gain, so that the speed hyperlocal model can more accurately reflect the dynamic characteristics of the permanent magnet motor.

[0022] Similarly, the current hyperlocal model is used to predict the change of motor current. Through the online self-learning mechanism, the lumped disturbance term and control input gain can be updated in real time to adapt to the actual operating state. The online self-learning mechanism adjusts the lumped disturbance term and control input gain in real time according to the actual operating state of the permanent magnet motor to ensure that the current hyperlocal model output is consistent with the actual current change of the permanent magnet motor.

[0023] It should be noted that the self-learning part of the speed hyperlocal model and the current hyperlocal model is realized through online neural network or fuzzy logic approximation. Neural network approximates complex nonlinear dynamic characteristics by learning the input-output relationship in the historical data queue, while fuzzy logic handles uncertainty and ambiguity through preset rules and real-time data adjustment. Both can effectively capture dynamic changes and ensure the prediction accuracy and control performance of the hyperlocal model.

[0024] Step 102: construct a historical data queue of the input and output states of the permanent magnet motor. The historical data queue satisfies a first-in-first-out update rule. The input and output states include speed, current, and voltage.

[0025] The historical data queue is used to store the input and output status data of the permanent magnet motor, including speed, current and voltage. The historical data queue follows the first-in-first-out (FIFO) update rule to ensure the timeliness and relevance of the input and output status data.

[0026] When new input and output status data is collected, the new input and output status data will be added to the end of the queue, and the oldest input and output status data in the queue will be removed. The first-in-first-out update rule can ensure that the historical data queue always contains the latest input and output status data that is most valuable to control.

[0027] Step 103: Collect the input and output status of the permanent magnet motor at the current moment and update the historical data queue.

[0028] Each time data sampling is performed, the current speed, current and voltage values ​​are recorded as the latest input and output status data. The latest input and output status data collected will be added to the end of the historical data queue, and the oldest input and output status data in the queue will be removed, thereby keeping the queue length unchanged.

[0029] Step 104: Using the data in the updated historical data queue and the first self-learning formula, update the lumped disturbance term and the control input gain in the speed super-local model, and using the data in the updated historical data queue and the second self-learning formula, update the lumped disturbance term and the control input gain in the current super-local model.

[0030] The first self-learning formula and the second self-learning formula are used to update the lumped disturbance terms and control input gains in the speed hyperlocal model and the current hyperlocal model, respectively. The core of the first self-learning formula and the second self-learning formula is to dynamically adjust the lumped disturbance terms and control input gains in the speed hyperlocal model and the lumped disturbance terms and control input gains in the current hyperlocal model by minimizing the difference between the predicted value and the actual value.

[0031] For the speed hyperlocal model, the first self-learning formula can update the lumped disturbance term and control input gain of the speed hyperlocal model by calculating the difference between the predicted speed and the actual speed. The second self-learning formula can update the lumped disturbance term and control input gain of the current hyperlocal model by calculating the difference between the predicted current and the actual current.

[0032] Step 105: Substitute the set reference speed into the updated speed superlocal model to obtain the optimal reference current, or substitute the set reference speed into the updated speed superlocal model to obtain the optimal reference current, and substitute the optimal reference current into the updated current superlocal model to obtain the optimal input voltage.

[0033] The updated speed superlocal model can more accurately reflect the actual operating state of the permanent magnet motor. The reference speed is input into the updated speed superlocal model. Further, the speed superlocal model can obtain the optimal reference current according to the current input and output state data of the permanent magnet motor and the reference speed.

[0034] It should be noted that the reference speed is usually determined according to the speed requirements and load conditions of the permanent magnet motor. For example, in semiconductor manufacturing equipment, in order to ensure processing accuracy, an accurate reference speed needs to be set. The reference current is the output of the outer loop and the input of the inner loop in the permanent magnet motor control, which is used to guide the actual current adjustment of the permanent magnet motor to achieve the desired torque output.

[0035] The current hyperlocal model is used to describe the relationship between current and voltage. Through the current hyperlocal model, the optimal reference current can be converted into the optimal input voltage.

[0036] Step 106: Select an optimal switch combination according to the optimal reference current and its corresponding current cost function or according to the optimal input voltage and its corresponding voltage cost function to drive the permanent magnet motor to operate.

[0037] There are two ways to select the optimal switch combination to drive the permanent magnet motor: One of the methods is: for each possible switch combination of the inverter, the corresponding predicted current is calculated using the current hyperlocal model. The optimal reference current and its corresponding current cost function are used to calculate the difference between the predicted current and the optimal reference current under each switch combination. The smaller the difference, the closer the predicted current under the switch combination is to the optimal reference current, and thus the switch combination is better.

[0038] By traversing all inverter switch combinations and calculating the corresponding current cost function values, the switch combination with the minimum current cost function value can be selected as the optimal switch combination. The optimal switch combination will be used to control the on and off of the inverter, thereby driving the permanent magnet motor to run at the desired speed.

[0039] Another way: Substitute the optimal reference current into the current hyperlocal model to obtain the optimal input voltage. Each switch combination of the inverter corresponds to an input voltage. The difference between the input voltage corresponding to each switch combination and the optimal input voltage is calculated through the optimal input voltage and its corresponding voltage cost function. The smaller the difference, the closer the input voltage under the switch combination is to the optimal input voltage, and thus the switch combination is better.

[0040] By traversing all inverter switch combinations and calculating the corresponding voltage cost function values, the switch combination with the minimum voltage cost function value is selected as the optimal switch combination. This optimal switch combination will be used to control the on and off of the inverter, thereby driving the permanent magnet motor to run at the desired speed.

[0041] It should be noted that the inverter, as a key component in the permanent magnet motor, has the main function of converting DC power into AC power to provide the required voltage and current for the permanent magnet motor. The inverter contains multiple power electronic switching devices, and the combination state of these switching devices determines the voltage waveform and amplitude of the inverter output.

[0042] The optimal switch combination has been determined through the evaluation of the self-learning hyperlocal model and the cost function. The optimal switch combination is obtained by evaluating all possible switch state combinations and selecting the combination that can make the actual current or voltage of the permanent magnet motor closest to the optimal reference value. After the optimal switch combination is determined, the corresponding control signal is sent to the inverter, and the control signal is used to drive the switch devices inside the inverter to perform on-off operations according to the optimal combination.

[0043] In this embodiment, the method includes: constructing a speed superlocal model and a current superlocal model, wherein the lumped disturbance terms and control input gains in the speed superlocal model and the current superlocal model are dynamically updated through an online self-learning mechanism; constructing a historical data queue that follows the first-in-first-out principle, which is used to store the input and output state data (including speed, current and voltage) of the permanent magnet motor, and update it in real time; using the data in the updated historical data queue and the first self-learning formula to update the lumped disturbance terms and control input gains in the speed superlocal model, and using the data in the updated historical data queue and the second self-learning formula to update the lumped disturbance terms and control input gains in the current superlocal model; and selecting the optimal inverter switch combination to drive the motor to operate according to the optimal reference current and its corresponding current cost function or the optimal input voltage and its corresponding voltage cost function.

[0044] This method introduces the online self-learning mechanism into the hyperlocal model, uses only historical data to approximate the actual dynamics of the system, and realizes the online update of the control input gain and the lumped disturbance term, alleviating the negative impact of parameter changes and unmodeled disturbances on the control performance. At the same time, this method does not need to design the control input gain and observer bandwidth based on physical information, reduces the dependence on a priori predicted motor parameters and manual experience, and can further improve the robustness and applicability of predictive control.

[0045] In one embodiment, Figure 2 As shown in the figure, the structure of the permanent magnet motor predictive control method based on the self-learning hyperlocal model is shown. The current input and output state data of the permanent magnet motor (PMSM) are collected in real time. The input and output state data include: Time speed ω ( n )as well as Current at all times i dq ( k ) and voltage u dq ( k ). Use the real-time collected input and output status data to update the historical data queue according to the first-in-first-out update rule. The historical data queue can output Input vector of neural network in the hyperlocal model of instantaneous current as well as The input vector of the neural network in the hyperlocal model of speed at the moment . Historical data series can also be output Input vector of neural network in the hyperlocal model of instantaneous current as well as The input vector of the neural network in the hyperlocal model of speed at the moment .based on , , Time speed ω ( n )as well as Current at all times i dq ( k ) and voltage u dq ( k ), using the first and second self-learning formulas, update the lumped disturbance term weights and control input gain weights in the speed super-local model and the current super-local model. Using the updated lumped disturbance term weights and control input gain weights in the speed super-local model and the current super-local model, Input vector of neural network in the hyperlocal model of instantaneous current as well as The input vector of the neural network in the hyperlocal model of speed at the moment , the updated speed hyperlocal model and current hyperlocal model can be obtained.

[0046] The reference speed ω * , actual speed And the updated lumped disturbance term and input gain Input to system constraints and speed control module to generate optimal reference current i dq * According to the actual current i dq ( k ), optimal reference current i dq * And the candidate voltages corresponding to different switching states And the set disturbance term of the updated current hyperlocal model and control input gain , calculate the current cost function, and select the switch state with the minimum current cost function value , driving the permanent magnet synchronous motor (PMSM) to operate.

[0047] In one embodiment, the initial speed super-local model is: ; in, and are the speed of the permanent magnet motor and the current of the q axis, and are the lumped disturbance term in the speed equation and the control input gain respectively.

[0048] Discretize the initial speed hyperlocal model in the first order and select a suitable speed sampling period T s1 , dividing the continuous time into discrete time steps; further, at each sampling moment ,use ω ( n )express The speed value at the moment; according to the Euler method, the initial speed hyperlocal model is approximated as a difference equation, and the speed hyperlocal model is obtained as: ; Further, and Respectively The lumped disturbance term and control input gain in the speed superlocal model at time instant, = , = It should be noted that and All are approximated by online neural network in the speed superlocal model. express In the local model of the moment speed The corresponding neural network weights, express In the local model of the moment speed The corresponding neural network weights, express Input vector of the neural network in the hyperlocal model of speed at time instant.

[0049] Therefore, the speed hyperlocal model can be as follows: ; in, represents the sampling period of the permanent magnet motor speed, express The predicted speed of the permanent magnet motor at time, express The speed of the permanent magnet motor collected at the moment, express The q-axis current collected at the moment, and denote the lumped disturbance term and control input gain of the speed hyperlocal model respectively; and All are approximated by online neural network in the speed superlocal model. express In the local model of the moment speed The corresponding neural network weights, express In the local model of the moment speed The corresponding neural network weights, express Input vector of the neural network in the hyperlocal model of speed at time instant.

[0050] In one embodiment, the initial current hyperlocal model is: ; in, is the stator d-axis current; , are the q-axis voltage and d-axis voltage at the stator port; , is the lumped disturbance term of the q-axis and d-axis in the current equation; , are the control input gains of the q-axis and d-axis in the current equation.

[0051] For the initial current hyperlocal model, select an appropriate current sampling period T s2 , which divides continuous time into discrete time steps. At each sampling moment ,use i q ( k )and i d ( k ) respectively represent The q-axis and d-axis current values ​​at the moment. According to the Euler method, the initial current hyperlocal model is approximated as a differential equation, that is, the current hyperlocal model is:

[0052] ; Further, and Respectively The q-axis and d-axis lumped disturbance terms and control input gains in the current hyperlocal model at time instant, = , = , = as well as = .in, and All are approximated by online neural networks in the current hyperlocal model. express In the hyperlocal model of the current The corresponding neural network weights, express In the hyperlocal model of the current The corresponding neural network weights, express Input vector of the neural network in the hyperlocal model of moment current.

[0053] Therefore, the current hyperlocal model can be as follows: ; in, represents the sampling period of the current, express The predicted current on the q axis at time, express The predicted current of the d-axis at time, express The d / q axis current collected at the moment, express The d / q axis voltage collected at the moment, and denote the lumped disturbance term and control input gain of the current hyperlocal model, respectively. express In the hyperlocal model of the current The corresponding neural network weights, express In the hyperlocal model of the current The corresponding neural network weights, express Input vector of the neural network in the hyperlocal model of moment current.

[0054] In one embodiment, the neural network has one hidden layer, the input of the neural network is a vector composed of data in the historical data queue, and the queue length of the historical data queue is selected according to the required control accuracy and the computing speed of the hardware. It should be noted that the neural network in the current hyperlocal model and the neural network in the speed hyperlocal model can be replaced by fuzzy logic.

[0055] In one embodiment, the speed sampling period is the outer loop sampling period, the current sampling period is the inner loop sampling period. Due to the different time constants of the inner and outer loops, the current sampling period Usually the speed sampling period 1 / 5 or 1 / 10 of.

[0056] In one embodiment, the speed hyperlocal model is substituted into the first error and the first error is minimized to obtain a first self-learning formula of the speed hyperlocal model; based on the first self-learning formula of the speed hyperlocal model, a lumped disturbance term of the updated speed hyperlocal model is obtained. The weights and control input gain The weight of .

[0057] Specifically, the first error is: .

[0058] Substituting the speed super-local model into the first error and minimizing the first error, the first self-learning formula of the speed super-local model can be obtained: ; in, represents the sampling period of the permanent magnet motor speed, express Lumped disturbance term of superlocal model of speed at time The weight of express Lumped disturbance term of superlocal model of speed at time The weight of express The input vector of the neural network in the hyperlocal model of the speed at the moment; express Control input gain of superlocal model of speed at moment The weight of express Control input gain of the superlocal model of speed at the moment The weight of express The predicted speed of the permanent magnet motor at time express The actual speed of the permanent magnet motor at the moment (i.e. the collected speed), Represents the learning rate of the first self-learning formula.

[0059] In one embodiment, the current hyperlocal model is substituted into the second error and the second error is minimized to obtain a second self-learning formula of the current hyperlocal model; based on the second self-learning formula of the current hyperlocal model, an updated lumped disturbance term of the current hyperlocal model is obtained. The weight and input gain The weight of .

[0060] Specifically, the second error is: .

[0061] Substituting the current hyperlocal model into the second error and minimizing the second error, the second self-learning formula of the current hyperlocal model is obtained: ; in, represents the sampling period of permanent magnet motor current, The hyperlocal model of the current at the moment The lumped disturbance term at time The weight of express Lumped disturbance term of the hyperlocal model of current at time The weight of express Control input gain of the hyperlocal model of instantaneous current The weight of express The predicted current of the permanent magnet motor at time, express The actual current of the permanent magnet motor collected at the moment, represents the learning rate of the second self-learning formula, express The input vector of the neural network in the hyperlocal model of current at time instant, express The actual voltage of the permanent magnet motor at this moment.

[0062] In one embodiment, the stability condition and parameter selection condition of the first self-learning formula are as follows: For the speed hyperlocal model, the following first Lyapunov function is constructed: ; in, , , , and They are and The corresponding optimal weight; Substitute the first self-learning formula into the first Lyapunov function, and When satisfied, we get: ; in, represents the upper bound of the activation function in the neural network, is the upper bound of the fitting error of the neural network under the optimal weights, ; According to the approximation theorem, the upper bound of the fitting error is As the number of neurons in the hidden layer of the neural network increases, Satisfy and the first learning rate Design meets When , that is, it satisfies the final consistency stability.

[0063] In one embodiment, the stability condition and parameter selection condition of the second self-learning formula are as follows: For the current hyperlocal model, the following second Lyapunov function is constructed: ; in, , ; and Don't and The corresponding optimal weight; Substitute the second self-learning formula into the second Lyapunov function, and When satisfied, we get: ; in, represents the upper bound of the activation function in the neural network, is the upper bound of the fitting error of the neural network under the optimal weights, ; According to the approximation theorem, the upper bound of the fitting error is As the number of neurons in the hidden layer of the neural network increases, Satisfies and the design of the second learning rate satisfies When , that is, it satisfies the final consistency stability.

[0064] It should be noted that is used to constrain the fitting error and the first learning rate Parameters, is used to constrain the fitting error , learning rate Parameters. and The value range is between 0 and 1. Fitting error of constrained speed hyperlocal model The relationship between the speed error and the interference term is: Fitting Error of the Hyperlocal Model of Constrained Current Relationship with current error and interference terms.

[0065] In this embodiment, by constructing the Lyapunov function, the final consistency is ensured to be stable. By dynamically adjusting the learning rate and neural network parameters, it is possible to adapt to internal and external disturbances, reduce errors, and improve control accuracy and robustness. As the number of neurons in the neural network increases, the fitting error gradually decreases, further improving the control performance.

[0066] In one embodiment, the optimal reference current is calculated by: And the reference speed Substitute the updated speed superlocal model to obtain the optimal reference current; The calculation formula for the optimal reference current is: ; in, , , represents the sampling period of the permanent magnet motor speed, and denote the lumped disturbance term and control input gain of the speed hyperlocal model, respectively. express In the local model of the moment speed The corresponding neural network weights, express In the local model of the moment speed The corresponding neural network weights. represents the optimal reference current, represents the current amplitude constraint, express The speed of the permanent magnet motor collected at the moment, express The input vector of the neural network in the hyperlocal model of speed at time instant, i q ∗ represents the optimal reference current of q axis, i d ∗ Indicates the optimal reference current of the d-axis.

[0067] It should be noted that the d-axis current is set to zero ( ), that is, under the maximum torque control strategy, the d-axis current is usually set to zero to reduce torque fluctuations and improve control performance. ω ∗ Substitute the speed super-local model, which can predict the speed at the next moment based on the current speed, lumped disturbance parameters, control input gain and q-axis current. The optimal reference current is obtained through the speed super-local model calculation i q ∗ and i d ∗ .

[0068] The current cost function is: ; in, , , represents the sampling period of the current, and Respectively The lumped disturbance term and control input gain of the current hyperlocal model at time instant, express In the hyperlocal model of the current The corresponding neural network weights, express In the hyperlocal model of the current The corresponding neural network weights. express The input vector of the neural network in the hyperlocal model of current at time instant, Indicates i The d / q axis voltage corresponding to each switch combination is: express The d / q axis current collected at the moment, i q ∗ Indicates the optimal reference current.

[0069] It should be noted that the current cost function can be used to evaluate the difference between the predicted current and the optimal reference current. By minimizing the current cost function, the predicted current can be made as close to the optimal reference current as possible, thereby achieving high-precision current control. By traversing all possible inverter switch combinations and calculating the corresponding current cost function value, the switch combination with the minimum current cost function value can be selected as the optimal switch combination. The optimal switch combination will be used to control the on and off of the inverter, thereby driving the permanent magnet motor to run at the desired speed.

[0070] In one embodiment, the optimal input voltage is calculated as follows: ; in, , , represents the sampling period of the current, express The d / q axis current collected at the moment, express The input vector of the neural network in the hyperlocal model of current at time instant, express The optimal reference current at the moment, express The optimal input voltage of the d / q axis at the moment.

[0071] It should be noted that The optimal reference current at the moment and Substituting into the current hyperlocal model, we can get The optimal input voltage of the d / q axis at the moment .

[0072] Furthermore, the voltage cost function is: ; in, express The optimal input voltage of the d / q axis at the moment, Indicates i The d / q axis voltage corresponding to each switch combination.

[0073] Further, The optimal input voltage of the d / q axis at the moment Substitute into the voltage cost function and calculate the voltage cost function value.

[0074] By traversing all possible inverter switch combinations, calculating the input voltage corresponding to each switch combination, and calculating the corresponding voltage cost function value, the switch combination with the minimum voltage cost function value can be selected as the optimal switch combination. This optimal switch combination will be used to control the on and off of the inverter, thereby driving the permanent magnet motor to run at the desired speed.

[0075] In one embodiment, the control effect of the proposed predictive control method based on the self-learning hyperlocal model on the permanent magnet motor is tested, including starting, constant speed, speed mutation, torque mutation and other working conditions. Among them, Figure 3 (a) is the speed control effect diagram of the permanent magnet motor, and Figure 3 (b) is the current control effect diagram of the permanent magnet motor. In the startup stage, the permanent magnet motor outputs the maximum torque to accelerate to the given 800RPM, and the q-axis current is saturated at the same time; entering the next stage, the permanent magnet motor maintains constant speed operation, the q-axis current is reduced, and only load torque is provided; entering the next stage, the given speed suddenly changes to 1200RPM, the q-axis current of the permanent magnet motor is immediately saturated, and the maximum torque is provided to accelerate the permanent magnet motor to the given value, and enter the constant speed operation again; entering the next stage, the load of the permanent magnet motor suddenly changes, and the q-axis current responds immediately, providing a larger torque to maintain the constant speed operation of the permanent magnet motor.

[0076] Based on the same concept as the predictive control method of permanent magnet motor based on self-learning hyperlocal model, such as Figure 4 As shown, the present application also provides a permanent magnet motor predictive control device based on a self-learning super-local model, the device comprising: A construction module 401 is used to construct a self-learning hyperlocal model for a permanent magnet motor, wherein the self-learning hyperlocal model includes a speed hyperlocal model and a current hyperlocal model, wherein the lumped disturbance term and the control input gain in the speed hyperlocal model and the current hyperlocal model are both self-learned online, and the self-learning parts in the speed hyperlocal model and the current hyperlocal model are approximated by an online neural network or fuzzy logic; the construction module is also used to construct a historical data queue of the input and output states of the permanent magnet motor, wherein the historical data queue satisfies a first-in-first-out update rule, and the input and output states include speed, current, and voltage; The acquisition module 402 is used to acquire the input and output status of the permanent magnet motor at the current moment and update the historical data queue; An updating module 403 is used to update the lumped disturbance term and the control input gain in the speed hyperlocal model by using the data in the updated historical data queue and the first self-learning formula, and to update the lumped disturbance term and the control input gain in the current hyperlocal model by using the data in the updated historical data queue and the second self-learning formula; The selection module 404 is used to substitute the set reference speed into the updated speed super-local model to obtain the optimal reference current, or substitute the set reference speed into the updated speed super-local model to obtain the optimal reference current, and substitute the optimal reference current into the updated current super-local model to obtain the optimal input voltage; the selection module is also used to select the optimal switch combination according to the optimal reference current and its corresponding current cost function or according to the optimal input voltage and its corresponding voltage cost function to drive the permanent magnet motor to operate.

[0077] In one embodiment, the speed super local model obtained by the acquisition module 401 is: ; in, represents the sampling period of the permanent magnet motor speed, express The predicted speed of the permanent magnet motor at time, express The speed of the permanent magnet motor collected at the moment, express The q-axis current collected at the moment, and denote the lumped disturbance term and control input gain of the speed hyperlocal model respectively; and All are approximated by online neural network in the speed superlocal model. express In the local model of the moment speed The corresponding neural network weights, express In the local model of the moment speed The corresponding neural network weights, express The input vector of the neural network in the hyperlocal model of the speed at the moment; The current super-local model obtained by the acquisition module 401 is: ; in, represents the sampling period of the current, express The predicted current on the q axis at time, express The predicted current of the d-axis at time, express The d / q axis current collected at the moment, express The d / q axis voltage collected at the moment, and denote the lumped disturbance term and control input gain of the current hyperlocal model respectively; Specifically, and All are approximated by online neural networks in the current hyperlocal model. express In the hyperlocal model of the current The corresponding neural network weights, express In the hyperlocal model of the current The corresponding neural network weights, express The input vector of the neural network in the hyperlocal model of the moment current; Each neural network has one hidden layer. The input of the neural network is a vector composed of data in the historical data queue. The queue length of the historical data queue is selected according to the required control accuracy and hardware computing speed. Neural networks can all be replaced by fuzzy logic.

[0078] In one embodiment, the updating module 403 calculates a first self-learning formula, where the first self-learning formula is obtained by minimizing a first error, where the first error is an error between a predicted rotation speed and an actual rotation speed; The first self-learning formula is: ; in, express The predicted speed of the permanent magnet motor at time express The actual speed of the permanent magnet motor at the moment, Represents the learning rate of the first self-learning formula.

[0079] In one embodiment, the updating module 403 calculates a second self-learning formula, where the second self-learning formula is obtained by minimizing a second error, where the second error is an error between the predicted current and the actual current; The second self-learning formula is: ; in, express The predicted current of the permanent magnet motor at time, express The actual current of the permanent magnet motor at the moment, Represents the learning rate of the second self-learning formula.

[0080] In one embodiment, the updating module 403 constructs the following first Lyapunov function for the speed superlocal model: ; in, , , , and They are and The corresponding optimal weight; Substitute the first self-learning formula into the first Lyapunov function, and When satisfied, we get: ; in, represents the upper bound of the activation function in the neural network, is the upper bound of the fitting error of the neural network under the optimal weights, ; According to the approximation theorem, the upper bound of the fitting error is As the number of neurons in the hidden layer of the neural network increases, Satisfy and the first learning rate Design meets When , that is, it satisfies the final consistency stability.

[0081] In one embodiment, the updating module 403 constructs the following second Lyapunov function for the current hyperlocal model: ; in, , ; and Don't and The corresponding optimal weight; Substitute the second self-learning formula into the second Lyapunov function, and When satisfied, we get: ; in, represents the upper bound of the activation function in the neural network, is the upper bound of the fitting error of the neural network under the optimal weights, ; According to the approximation theorem, the upper bound of the fitting error is As the number of neurons in the hidden layer of the neural network increases, Satisfies and the design of the second learning rate satisfies When , that is, it satisfies the final consistency stability.

[0082] In one embodiment, the selection module 404 will And the reference speed Substitute the updated speed superlocal model to obtain the optimal reference current; The calculation formula for the optimal reference current is: ; in, , , represents the optimal reference current, represents the current amplitude constraint, express The speed of the permanent magnet motor collected at the moment, express The input vector of the neural network in the hyperlocal model of speed at time instant, Indicates the sampling period of the permanent magnet motor speed; The current cost function is: ; in, , , express The input vector of the neural network in the hyperlocal model of current at time instant, represents the sampling period of the current, Indicates i The d / q axis voltage corresponding to each switch combination is: express The d / q axis current collected at the moment.

[0083] In one embodiment, the formula used by the selection module 404 to calculate the optimal input voltage is: ; in, , , represents the sampling period of the current, express The d / q axis current collected at the moment, express The input vector of the neural network in the hyperlocal model of current at time instant, express The optimal reference current at the moment, express The optimal input voltage of the d / q axis at the moment; The voltage cost function is: ; in, Indicates i The d / q axis voltage corresponding to each switch combination.

[0084] Based on the same concept as the permanent magnet motor predictive control method based on the self-learning super-local model, the present application also provides a computer device, the computer device includes a processor and a memory, the memory stores a computer program, and the processor implements the permanent magnet motor predictive control method based on the self-learning super-local model when executing the computer program.

[0085] In one embodiment, a computer device is provided, whose internal structure diagram can be as follows: Figure 5 As shown. The computer device includes a processor, a memory, a communication interface, a display screen and an input device connected through a system bus. Among them, the processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system and a computer program. The internal memory provides an environment for the operation of the operating system and the computer program in the non-volatile storage medium. The communication interface of the computer device is used to communicate with an external terminal in a wired or wireless manner, and the wireless manner can be achieved through WIFI, a mobile cellular network, NFC (near field communication) or other technologies. When the computer program is executed by the processor, a permanent magnet motor predictive control method based on a self-learning super-local model is implemented. The display screen of the computer device can be a liquid crystal display screen or an electronic ink display screen, and the input device of the computer device can be a touch layer covered on the display screen, or a key, trackball or touchpad set on the computer device shell, or an external keyboard, touchpad or mouse.

[0086] Those skilled in the art will understand that Figure 5 The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0087] The technical features of the above embodiments may be combined arbitrarily. To make the description concise, 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.

[0088] The above embodiments only express several implementation methods of the present application, and the descriptions thereof are relatively specific and detailed, but they cannot be understood as limiting the scope of the present application. It should be pointed out that, for a person of ordinary skill in the art, several modifications and improvements can be made without departing from the concept of the present application, and these all belong to the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the attached claims.

Claims

1. A predictive control method for a permanent magnet motor based on a self-learning hyperlocal model, characterized in that: The method comprises the following steps: A self-learning hyperlocal model is constructed for the permanent magnet motor, wherein the self-learning hyperlocal model includes a speed hyperlocal model and a current hyperlocal model, wherein the lumped disturbance term and the control input gain in the speed hyperlocal model and the current hyperlocal model are both self-learned online, and the self-learning parts in the speed hyperlocal model and the current hyperlocal model are approximated by an online neural network or fuzzy logic; Constructing a historical data queue of the input and output states of the permanent magnet motor, wherein the historical data queue satisfies a first-in-first-out update rule, and the input and output states include rotation speed, current and voltage; Collect the input and output status of the permanent magnet motor at the current moment and update the historical data queue; Using the updated data in the historical data queue and the first self-learning formula, the lumped disturbance term and the control input gain in the speed super-local model are updated, and using the updated data in the historical data queue and the second self-learning formula, the lumped disturbance term and the control input gain in the current super-local model are updated; Substituting the set reference speed into the updated speed super-local model to obtain the optimal reference current, or substituting the set reference speed into the updated speed super-local model to obtain the optimal reference current, and substituting the optimal reference current into the updated current super-local model to obtain the optimal input voltage; An optimal switch combination is selected according to the optimal reference current and its corresponding current cost function or according to the optimal input voltage and its corresponding voltage cost function to drive the permanent magnet motor to operate.

2. The method for predictive control of a permanent magnet motor based on a self-learning super-local model according to claim 1, characterized in that: The speed super-local model is: ； in, represents the sampling period of the permanent magnet motor speed, express The predicted speed of the permanent magnet motor at the time, express The speed of the permanent magnet motor collected at the time, express The q-axis current collected at the moment, and denote the lumped disturbance term and control input gain of the speed super-local model respectively; Specifically, and are all approximated by the online neural network in the speed superlocal model. express The speed at the moment is super local model The corresponding neural network weights, express The speed at the moment is super local model The corresponding neural network weights, express The input vector of the neural network in the speed hyperlocal model at the time; The current hyperlocal model is: ； in, represents the sampling period of the current, express The predicted current on the q axis at time, express The predicted current of the d-axis at time, express The d / q axis current collected at the moment, express The d / q axis voltage collected at the moment, and denote the lumped disturbance term and the control input gain of the current hyperlocal model respectively; Specifically, and are all approximated by an online neural network in the current hyperlocal model, express The current hyperlocal model at the moment The corresponding neural network weights, express The current hyperlocal model at the moment The corresponding neural network weights, express The input vector of the neural network in the current hyperlocal model at the time instant; The neural network has one hidden layer, the input of the neural network is a vector composed of data in the historical data queue, and the queue length of the historical data queue is selected according to the required control accuracy and hardware operation speed; The neural networks can all be replaced by fuzzy logic.

3. The method for predictive control of a permanent magnet motor based on a self-learning super-local model according to claim 2, characterized in that: The first self-learning formula is obtained by minimizing a first error, where the first error is an error between a predicted rotation speed and an actual rotation speed; The first self-learning formula is: ； in, express The predicted speed of the permanent magnet motor at time express The actual speed of the permanent magnet motor at the time, Represents the learning rate of the first self-learning formula.

4. The method for predictive control of a permanent magnet motor based on a self-learning super-local model according to claim 2, characterized in that: The second self-learning formula is obtained by minimizing a second error, where the second error is an error between the predicted current and the actual current; The second self-learning formula is: ； in, express The predicted current of the permanent magnet motor at time, express The actual current of the permanent magnet motor at the moment is, Represents the learning rate of the second self-learning formula.

5. The method for predictive control of a permanent magnet motor based on a self-learning super-local model according to claim 3, characterized in that: The stability conditions and parameter selection conditions of the first self-learning formula are as follows: For the speed hyperlocal model, the following first Lyapunov function is constructed: ； in, , , , and They are and The corresponding optimal weight; Substituting the first self-learning formula into the first Lyapunov function, and When satisfied, we get: ; in, represents the upper bound of the activation function in the neural network, is the upper bound of the fitting error of the neural network under the optimal weights, ; According to the approximation theorem, the upper bound of the fitting error is As the number of neurons in the hidden layer of the neural network increases, Satisfy and the first learning rate Design meets When , that is, it satisfies the final consistency stability.

6. The method for predictive control of a permanent magnet motor based on a self-learning super-local model according to claim 4, characterized in that: The stability conditions and parameter selection conditions of the second self-learning formula are as follows: For the current hyperlocal model, the following second Lyapunov function is constructed: ； in, , ; and Don't and The corresponding optimal weight; Substituting the second self-learning formula into the second Lyapunov function, and When satisfied, we get: ; in, represents the upper bound of the activation function in the neural network, is the upper bound of the fitting error of the neural network under the optimal weights, ; According to the approximation theorem, the upper bound of the fitting error is As the number of neurons in the hidden layer of the neural network increases, Satisfies and the design of the second learning rate satisfies When , that is, it satisfies the final consistency stability.

7. The method for predictive control of a permanent magnet motor based on a self-learning super-local model according to claim 2, characterized in that: The optimal reference current is calculated as follows: And the reference speed Substitute the updated speed superlocal model to obtain the optimal reference current; The calculation formula of the optimal reference current is: ； in, , , represents the optimal reference current, represents the current amplitude constraint, express The speed of the permanent magnet motor collected at the time, express The input vector of the neural network in the speed hyperlocal model at time , Represents the sampling period of the permanent magnet motor speed; The current cost function is: ； in, , , express The input vector of the neural network in the current hyperlocal model at time instant, represents the sampling period of the current, Indicates i The d / q axis voltage corresponding to each switch combination is: express The d / q axis current collected at the moment.

8. The method for predictive control of a permanent magnet motor based on a self-learning super-local model according to claim 2, characterized in that: The calculation formula of the optimal input voltage is: ； in, , , represents the sampling period of the current, express The d / q axis current collected at the moment, express The input vector of the neural network in the current hyperlocal model at time instant, express The optimal reference current at the moment, express The optimal input voltage of the d / q axis at the moment; The voltage cost function is: ; in, Indicates i The d / q axis voltage corresponding to each switch combination.

9. A permanent magnet motor predictive control device based on a self-learning super-local model, characterized in that: The device comprises: A construction module is used to construct a self-learning hyperlocal model for the permanent magnet motor, the self-learning hyperlocal model includes a speed hyperlocal model and a current hyperlocal model, the lumped disturbance term and the control input gain in the speed hyperlocal model and the current hyperlocal model are both online self-learned, and the self-learning parts in the speed hyperlocal model and the current hyperlocal model are approximated by an online neural network or fuzzy logic; the construction module is also used to construct a historical data queue of the input and output states of the permanent magnet motor, the historical data queue satisfies a first-in-first-out update rule, and the input and output states include speed, current and voltage; An acquisition module, used for acquiring the input and output status of the permanent magnet motor at the current moment and updating the historical data queue; An updating module, used to update the lumped disturbance term and the control input gain in the speed super-local model by using the updated data in the historical data queue and the first self-learning formula, and to update the lumped disturbance term and the control input gain in the current super-local model by using the updated data in the historical data queue and the second self-learning formula; A selection module is used to substitute the set reference speed into the updated speed super-local model to obtain the optimal reference current, or substitute the set reference speed into the updated speed super-local model to obtain the optimal reference current, and substitute the optimal reference current into the updated current super-local model to obtain the optimal input voltage; the selection module is also used to select the optimal switch combination according to the optimal reference current and its corresponding current cost function or according to the optimal input voltage and its corresponding voltage cost function to drive the permanent magnet motor to operate.

10. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the permanent magnet motor predictive control method based on the self-learning super local model according to any one of claims 1 to 8 are implemented.

Images