Motor speed regulation and temperature control method based on MPC algorithm

Abstract

The application discloses a motor speed regulation and temperature control method based on an MPC algorithm, and comprises the following steps: step 1, collecting real-time data of a motor as a sample space, and establishing a neural network model related to real-time temperature of the motor as an environment model; step 2, the environment model predicts a future temperature trend based on a current temperature, and the future temperature trend is interacted with a real environment by using an MPC algorithm to obtain an interaction track; and step 3, a speed regulation and temperature control action is made on the motor according to the interaction track. The motor is subjected to speed regulation and temperature control, so that the heat generation of the motor can be reduced, the normal work of the motor is ensured, and the service life of the motor is prolonged.

Description

Technical Field

[0001] This invention relates to the field of motor control technology, and in particular to a motor speed regulation and temperature control method based on the MPC algorithm. Background Technology

[0002] In the industrial field, electric motors are the power source for equipment. While converting power into mechanical energy, motors also convert some energy into heat energy. When this heat energy exceeds a certain range, it weakens the motor's performance, and may even burn it out if it exceeds its load-bearing limit, severely impacting the overall performance of the equipment and affecting production. Therefore, it is necessary to control the motor's temperature. Existing patent document CN114337098A discloses a motor enhanced heat dissipation structure. By setting a special enhanced heat dissipation geometry on the surface through which the cooling medium flows, this structure can create a longitudinal vortex phenomenon in its wake region, breaking or reducing the velocity boundary and temperature boundary layer, and disturbing the fluid velocity, thereby improving the heat exchange capacity of the cooling medium. This method has a certain auxiliary heat dissipation effect, but it is not decisive and is not suitable for motor cooling systems under high load conditions. Furthermore, the above method achieves motor heat dissipation through changes in the physical structure. With the rapid development of computer technology, it has also been gradually applied to the control field. Therefore, how to combine computer technology and machine learning algorithms to achieve motor speed regulation and temperature control without changing the motor's physical structure has become a technical problem that needs to be solved. Summary of the Invention

[0003] The purpose of this invention is to provide a motor speed regulation and temperature control method based on the MPC algorithm. This invention regulates the speed and temperature of the motor, reducing heat generation, ensuring normal motor operation, and extending the motor's service life.

[0004] The technical solution of this invention: A motor speed regulation and temperature control method based on MPC algorithm, comprising the following steps:

[0005] Step 1: Collect real-time data of the motor as the sample space, and establish a neural network model related to the real-time temperature of the motor as the environment model;

[0006] Step 2: The environmental model predicts future temperature trends based on the current temperature, and uses the MPC algorithm to interact with the future temperature trends and the real environment to obtain the interaction trajectory.

[0007] Step 3: Adjust the speed and temperature of the motor according to the interaction trajectory.

[0008] The aforementioned motor speed regulation and temperature control method based on the MPC algorithm includes motor torque T as an input to the neural network. n Motor speed v, motor power loss P tGiven an ambient temperature T0, the output of the neural network is the motor temperature T; the neural network includes an input layer, a hidden layer, and an output layer; the input layer has 4 nodes, the hidden layer has 7 nodes, and the output layer has 1 node.

[0009] The aforementioned motor speed regulation and temperature control method based on the MPC algorithm employs the controlled variable method. When the motor temperature stabilizes after a long period of operation, the computer collects the motor torque T at different speeds during no-load operation. n Motor speed v, motor temperature T, and ambient temperature T0.

[0010] The aforementioned motor speed regulation and temperature control method based on the MPC algorithm is based on the principle of energy conservation. When the motor temperature stabilizes after a long period of operation, the total motor loss P is calculated by measuring the air velocity, flow rate, and temperature difference between the motor and the inlet and outlet at different speeds during no-load operation. t The formula is:

[0011] P t =cmΔT

[0012] Where c is the specific heat capacity of air, m is the mass of air flowing through the motor per unit time, and ΔT is the change in air temperature at the air inlet and outlet.

[0013] The aforementioned motor speed regulation and temperature control method based on the MPC algorithm includes the following steps for obtaining the interaction trajectory:

[0014] Step 2.1: Taking the current motor temperature as state s and the motor speed as action a, construct a state transition function based on the current state and action. and reward value Simulation data were collected based on future temperature trends predicted by environmental models.

[0015] Step 2.2: Starting from the current state s0, for each action 'a' in the future temperature trend, append an action sequence of length T to obtain [a0, a1, a2, ..., a T ];

[0016] Step 2.3: Use the cross-entropy method to obtain the optimal action sequence, and select the first action 'a' from the optimal sequence each time. k To interact with the environment model,

[0017] Step 2.4: Interact with the environment model using this action sequence to generate a trajectory.

[0018] The aforementioned motor speed regulation and temperature control method based on the MPC algorithm involves obtaining the optimal action sequence using the cross-entropy method by first setting the number of iterations e, and then selecting N action sequences A1, ..., A2 from the action sequence distribution P(N). N For each action sequence A1, ..., A N The cumulative reward is evaluated using the environment module, and then M optimal action sequences A are retained based on the evaluation results. i1 ,···,A iN Using M optimal action sequences A i1 ,···,A iN Update the action sequence distribution P(N), repeat the loop e times, and calculate the mean of the first action of all optimal action sequences. This mean is taken as the action at the current time.

[0019] The aforementioned motor speed regulation and temperature control method based on the MPC algorithm includes the following steps:

[0020] Step 3.1: Based on the interaction trajectory, calculate a cumulative reward value for (s0, a0) with γ as the decay factor, i.e.

[0021] Step 3.2: For each action 'a', perform the calculation process of Step 3.1, and select the action with the highest cumulative value to execute.

[0022] Step 3.3: Establish constraint relationships:

[0023]

[0024] In actual operation, the reward value is positively correlated with temperature changes, and the temperature is not higher than the preset upper limit and the speed is not lower than the preset lower limit.

[0025] Compared with existing technologies, this invention collects real-time motor data as a sample space, then establishes a neural network model related to the motor's real-time temperature as an environmental model, and then uses the MPC algorithm to predict future temperature trends based on the current temperature according to the environmental model and interacts with the real environment, finally making speed adjustment and temperature control actions. This invention can automatically modify motor parameters when the motor temperature is about to reach the upper limit threshold, so as to reduce heat generation and maintain it within a reasonable temperature range without stopping the machine. Attached Figure Description

[0026] Figure 1 This is a schematic diagram illustrating the principle of the neural network of this invention;

[0027] Figure 2 This is a schematic diagram illustrating the process of obtaining the optimal action sequence using the cross-entropy method;

[0028] Figure 3 This is a schematic diagram illustrating the principle of the MPC algorithm. Detailed Implementation

[0029] The present invention will be further described below with reference to embodiments, but these embodiments are not intended to limit the scope of the invention.

[0030] Example: A motor speed regulation and temperature control method based on MPC algorithm, comprising the following steps:

[0031] Step 1: Collect real-time data of the motor as the sample space, and establish a neural network model related to the real-time temperature of the motor as the environment model;

[0032] The input to the neural network includes the motor torque T. n Motor speed v, motor power loss P t Given an ambient temperature T0, the output of the neural network is the motor temperature T. As shown in the figure, the neural network includes an input layer, a hidden layer, and an output layer. The input layer has 4 nodes, the hidden layer has 7 nodes, and the output layer has 1 node.

[0033] The real-time data acquisition of the motor adopts the controlled variable method. When the motor temperature stabilizes after a long period of operation, the computer collects the motor torque T at different speeds during no-load operation. n Motor speed v, motor temperature T, and ambient temperature T0.

[0034] Based on the principle of energy conservation, when the motor reaches a stable temperature after a long period of operation, the total loss P of the motor is calculated by measuring the air velocity, flow rate, and temperature difference between the motor and the inlet and outlet at different speeds during no-load operation. t The formula is:

[0035] P t =cmΔT

[0036] Where c is the specific heat capacity of air, m is the mass of air flowing through the motor per unit time, and ΔT is the change in air temperature at the air inlet and outlet.

[0037] Step 2: The environmental model predicts future temperature trends based on the current temperature, and uses the MPC algorithm to interact with the future temperature trends and the real environment to obtain the interaction trajectory.

[0038] The acquisition of the interaction trajectory includes the following steps:

[0039] Step 2.1: Taking the current motor temperature as state s and the motor speed as action a, construct a state transition function based on the current state and action. and reward value Simulation data were collected based on future temperature trends predicted by environmental models.

[0040] Step 2.2: Starting from the current state s0, for each action 'a' in the future temperature trend, append an action sequence of length T to obtain [a0, a1, a2, ..., a T ];

[0041] Step 2.3: Use the cross-entropy method to obtain the optimal action sequence, and select the first action 'a' from the optimal sequence each time. k To interact with the environment model; such as Figure 2 As shown, the process of obtaining the optimal action sequence using the cross-entropy method involves first setting the number of iterations e, and then selecting N action sequences A1, ..., A2 from the action sequence distribution P(N). N For each action sequence A1, ..., A N The cumulative reward is evaluated using the environment module, and then M optimal action sequences A are retained based on the evaluation results. i1 ,···,A iN Using M optimal action sequences A i1 ,···,A iN Update the action sequence distribution P(N), repeat the loop e times, and calculate the mean of the first action of all optimal action sequences. This mean is taken as the action at the current time.

[0042] Step 2.4: Interact with the environment model using this action sequence to generate a trajectory.

[0043] Step 3: Adjust the speed and temperature of the motor according to the interaction trajectory.

[0044] like Figure 3 The speed regulation and temperature control action includes the following steps:

[0045] Step 3.1: Based on the interaction trajectory, calculate a cumulative reward value for (s0, a0) with γ as the decay factor, i.e.

[0046] Step 3.2: For each action 'a', perform the calculation process of Step 3.1, and select the action with the highest cumulative value to execute.

[0047] Step 3.3: Establish constraint relationships:

[0048]

[0049] In actual operation, the reward value is positively correlated with temperature changes, and the temperature is not higher than the preset upper limit and the speed is not lower than the preset lower limit.

[0050] In summary, this invention uses real-time motor data as a sample space, then establishes a neural network model related to the motor's real-time temperature as an environmental model, and employs an MPC algorithm to predict future temperature trends based on the current temperature using the environmental model, interacting with the real environment to finally implement speed and temperature control actions. This invention can automatically modify motor parameters when the motor temperature is about to reach a threshold upper limit, reducing heat generation and maintaining it within a reasonable temperature range without stopping the machine.

Claims

1. A motor speed regulation and temperature control method based on MPC algorithm, characterized in that: Includes the following steps: Step 1: Collect real-time data of the motor as the sample space, and establish a neural network model related to the real-time temperature of the motor as the environment model; Step 2: The environmental model predicts the future temperature trend based on the current temperature, and uses the MPC algorithm to interact with the future temperature trend and the real environment to obtain the interaction trajectory. Step 3: Adjust the speed and temperature of the motor according to the interaction trajectory; The acquisition of the interaction trajectory includes the following steps: Step 2.1: Based on the current motor temperature. The motor's speed is the operating speed. Construct a state transition function based on the current state and actions. and reward value Simulation data is collected based on future temperature trends predicted by environmental models. ; Step 2.2, from the current state Starting with every action in anticipation of future temperature trends. Attached is a length of The action sequence is obtained ; Step 2.3: Use the cross-entropy method to obtain the optimal action sequence, and select the first action from the optimal sequence each time. To interact with the environment model; Step 2.4: Interact with the environment model using this action sequence to generate a trajectory. ; The process of obtaining the optimal action sequence using the cross-entropy method involves first setting the number of iterations. From the action sequence distribution Select Action sequence For each action sequence The cumulative rewards are evaluated using the environment module, and then retained based on the evaluation results. Optimal action sequence Using the optimal action sequence Update action sequence distribution Repeated loop After that, the mean of the first action in all optimal action sequences is calculated, and this mean is taken as the action to be taken at the current moment. The speed regulation and temperature control action includes the following steps: Step 3.1: Based on the interaction trajectory, Calculate to get a The cumulative reward value of the decay factor, i.e. ; Step 3.2, for each action All actions follow the calculation process in step 3.1, and the action with the highest cumulative value is selected for execution. ; Step 3.3: Establish constraint relationships: ； In actual operation, the reward value is positively correlated with temperature changes, and the temperature is not higher than the preset upper limit and the speed is not lower than the preset lower limit.

2. The motor speed regulation and temperature control method based on MPC algorithm according to claim 1, characterized in that: The input to the neural network includes motor torque. Motor speed Motor power loss and ambient temperature The output of the neural network is the motor temperature. The neural network includes an input layer, a hidden layer, and an output layer; the input layer has 4 nodes, the hidden layer has 7 nodes, and the output layer has 1 node.

3. The motor speed regulation and temperature control method based on the MPC algorithm according to claim 2, characterized in that: Using the controlled variable method, after the motor has run for a long time and its temperature has stabilized, the computer collects the motor torque at different speeds during no-load operation. Motor speed Motor temperature and ambient temperature .

4. The motor speed regulation and temperature control method based on MPC algorithm according to claim 2, characterized in that: Based on the principle of energy conservation, when the motor reaches a stable temperature after a long period of operation, the total loss of the motor is calculated by measuring the air velocity, flow rate, and temperature difference between the motor and the inlet and outlet at different speeds during no-load operation. The formula is: in, The specific heat capacity of air, The mass of air flowing through the motor per unit time. This represents the change in air temperature at the air inlet and outlet.

Images