An Internet of Things-based hydraulic valve linkage control system

Abstract

The present invention relates to the technical field of hydraulic control, and specifically discloses an Internet-of-Things-based hydraulic valve linkage control system, including sensors, actuators, and controllers. This solution uses a closed-loop feedback control system to ensure the stable operation of the hydraulic system. Through the interconnection of sensors, controllers, and actuators, the linkage between hydraulic valves and hydraulic cylinders in the hydraulic system is achieved. By real-time controlling the parameters of the hydraulic valves, the position of the hydraulic cylinders is accurately adjusted, improving the robustness and accuracy of the hydraulic system; a control strategy is designed based on the MRAC algorithm and the DRL algorithm, enabling the hydraulic valves to increase their learning ability and self-adaptability while approaching the desired control state, improving the response speed, efficiency, and stability of the control system, and being applicable to various complex hydraulic scenarios.

Description

Technical Field

[0001] The present invention relates to the technical field of hydraulic control, and specifically refers to a hydraulic valve linkage control system based on the Internet of Things. Background Art

[0002] Hydraulic valves are key components in hydraulic systems, used to control the flow direction, pressure, and flow rate of liquids. Modern hydraulic systems often use electro-hydraulic valves to achieve complex control functions and automated operations of hydraulic valves through electrical signals. Traditional hydraulic systems have a slow response speed when facing rapidly changing working loads, making it difficult to adjust hydraulic valves in a timely manner, and requiring a large amount of manual intervention and debugging, resulting in high maintenance costs and operational complexity; general hydraulic valve control systems rely on controllers with fixed parameters, and have poor control performance under the conditions of parameter changes and uncertain dynamic characteristics, resulting in energy waste and low resource utilization efficiency. Summary of the Invention

[0003] In view of the above situation, to overcome the defects of the prior art, the present invention provides a hydraulic valve linkage control system based on the Internet of Things. Aiming at the problems that traditional hydraulic systems have a slow response speed when facing rapidly changing working loads, making it difficult to adjust hydraulic valves in a timely manner, and requiring a large amount of manual intervention and debugging, resulting in high maintenance costs and operational complexity, this solution uses a closed-loop feedback control system to ensure the stable operation of the hydraulic system. Through the interconnection of sensors, controllers, and actuators, the linkage between hydraulic valves and hydraulic cylinders in the hydraulic system is realized. By real-time controlling the parameters of the hydraulic valves, the position of the hydraulic cylinders is accurately adjusted, improving the robustness and accuracy of the hydraulic system; aiming at the problems that general hydraulic valve control systems rely on controllers with fixed parameters, and have poor control performance under the conditions of parameter changes and uncertain dynamic characteristics, resulting in energy waste and low resource utilization efficiency, this solution designs a control strategy based on the MRAC algorithm and the DRL algorithm, enabling the hydraulic valves to increase learning ability and self-adaptability while approaching the desired control state, improving the response speed, efficiency, and stability of the control system, and being applicable to various complex hydraulic scenarios.

[0004] A hydraulic valve linkage control system based on the Internet of Things provided by the present invention includes sensors, actuators, and controllers;

[0005] The sensors collect real-time data of pressure, flow rate, and temperature, and send the obtained real-time data to the controller;

[0006] The controller designs a control strategy based on the MRAC algorithm and the DRL algorithm, feeds the real-time data back into the control strategy to generate a control signal, and sends the control signal to the actuator;

[0007] The actuator drives the hydraulic valve to perform corresponding operations according to the control signal;

[0008] The control strategy is designed based on the MRAC algorithm and the DRL algorithm, and includes the following steps:

[0009] Step S1: Reference model design. Design a reference model based on the MRAC algorithm to describe the desired position of the hydraulic cylinder. Take the desired hydraulic valve parameters as the input signal, and the output signal is obtained by calculating the input signal through the reference model. The formula used is as follows:

[0010] Y m (s) = G m (s)·R(s);

[0011] In the formula, m represents the reference model, s represents the complex frequency variable, G m (s) represents the transfer function of the reference model, Y m (s) represents the frequency-domain representation of the output signal, and R(s) represents the frequency-domain representation of the input signal;

[0012] Step S2: Error calculation. Convert the output signal to the time domain as the reference value, control the operation of the hydraulic valve based on the reference value, and take the difference between the reference value and the output value as the control error to describe the difference between the actual hydraulic valve parameters and the desired hydraulic valve parameters. The formula used is as follows:

[0013] e(t) = y m (t) - y(t);

[0014] In the formula, t represents the time, e(t) represents the control error, y m (t) represents the reference value, and y(t) represents the output value;

[0015] Step S3: Design the adaptation law. Adjust the hydraulic valve parameters according to the adaptation law to make the output value gradually approach the reference value and generate the basic control signal. The formula used is as follows:

[0016] u MRAC (t) = θ(t)·r(t);

[0017] In the formula, θ(t) represents the hydraulic valve parameters, u MRAC (t) represents the basic control signal, and r(t) represents the time-domain representation of the input signal;

[0018] Step S4: Strategy optimization. Use the DRL algorithm to construct a deep neural network as the optimization strategy for the adaptation law to generate the compensation signal. The formula used is as follows:

[0019]

[0020] In the formula, u DRL (t) represents the compensation signal, represents the optimization function, and s(t) represents the real-time data;

[0021] Step S5: Generate a control signal by combining the basic control signal generated using the MRAC algorithm with the compensation signal adjusted by the DRL algorithm. The formula used is as follows:

[0022] u(t) = u MRAC (t) + u DRL (t);

[0023] In the formula, u(t) represents the control signal.

[0024] Furthermore, in step S4, the policy optimization includes the following steps:

[0025] Step S41: State and action representation. The state of the optimized policy is the real-time data collected by the sensor, and the action is the hydraulic valve parameter;

[0026] Step S42: Q-value estimation. Construct a deep neural network as the Q function to calculate the Q-value of the corresponding action in each state, representing the possibility that the output value reaches the reference value;

[0027] Step S43: Action selection. Use the ε-greedy policy to select a random action with probability ε and select the action with the highest current Q-value with probability 1 - ε;

[0028] Step S44: Reward function design. Use the response speed of the control system as the reward signal, giving a positive reward when the response speed is fast and a negative reward when the response speed is slow;

[0029] Step S45: Parameter optimization. Use the stochastic gradient descent algorithm to adjust the weights of the deep neural network;

[0030] Step S46: Compensation signal generation. Steps S42 to S45 are the construction process of the deep neural network. Input the state into the deep neural network and generate the corresponding compensation signal according to the action selected each time.

[0031] The beneficial effects achieved by the present invention using the above solution are as follows:

[0032] (1) Aiming at the problems that the traditional hydraulic system has a slow response speed when facing rapidly changing working loads, it is difficult to adjust the hydraulic valve in time, and requires a large amount of manual intervention and debugging, with high maintenance costs and operation complexity. This solution uses a closed-loop feedback control system to ensure the stable operation of the hydraulic system. Through the interconnection of sensors, controllers, and actuators, the linkage between the hydraulic valve and the hydraulic cylinder in the hydraulic system is realized. By precisely controlling the position of the hydraulic cylinder through real-time control of the hydraulic valve parameters, the robustness and accuracy of the hydraulic system are improved.

[0033] (2) Aiming at the problem that the general hydraulic valve control system depends on a controller with fixed parameters, has poor control performance under the conditions of parameter changes and uncertain dynamic characteristics, and has problems of energy waste and low resource utilization efficiency, this solution designs a control strategy based on the MRAC algorithm and the DRL algorithm, enabling the hydraulic valve to increase its learning ability and adaptability while approaching the desired control state, improving the response speed, efficiency, and stability of the control system, and being applicable to various complex hydraulic scenarios. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 FIG. is a schematic diagram of a hydraulic valve linkage control system based on the Internet of Things proposed by the present invention;

[0035] Figure 2 FIG. is a schematic diagram of a deep neural network constructed for Example VI.

[0036] The drawings are used to provide a further understanding of the present invention, and constitute a part of the specification. They are used together with the embodiments of the present invention to explain the present invention, but do not constitute a limitation to the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0037] The following will clearly and completely describe the technical solutions in the embodiments of the present invention with reference to the drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments; based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the scope of protection of the present invention.

[0038] Example 1, referring to Figure 1 , a hydraulic valve linkage control system based on the Internet of Things provided by the present invention includes a sensor, an actuator, and a controller;

[0039] The sensor collects real-time data of pressure, flow rate, and temperature, and sends the obtained real-time data to the controller;

[0040] The controller designs a control strategy based on the MRAC algorithm and the DRL algorithm, feeds the real-time data back into the control strategy to generate a control signal, and sends the control signal to the actuator;

[0041] The actuator drives the hydraulic valve to perform corresponding operations according to the control signal;

[0042] The control strategy is designed based on the MRAC algorithm and the DRL algorithm, and includes the following steps:

[0043] Step S1: Reference model design. A reference model is designed based on the MRAC algorithm to describe the desired position of the hydraulic cylinder. The desired hydraulic valve parameters are used as input signals, and the input signals are calculated by the reference model to obtain output signals. The formula used is as follows:

[0044] Y m (s) = G m (s)·R(s);

[0045] Wherein, m represents the reference model, s represents the complex frequency variable, G m (s) represents the transfer function of the reference model, Y m (s) represents the frequency-domain representation of the output signal, and R(s) represents the frequency-domain representation of the input signal;

[0046] Step S2: Error calculation. Convert the output signal to the time domain as the reference value, control the operation of the hydraulic valve based on the reference value, and use the difference between the reference value and the output value as the control error to describe the difference between the actual hydraulic valve parameters and the desired hydraulic valve parameters. The formula used is as follows:

[0047] e(t) = y m (t) - y(t);

[0048] Wherein, t represents the moment, e(t) represents the control error, y m (t) represents the reference value, and y(t) represents the output value;

[0049] Step S3: Design the adaptation law. Adjust the hydraulic valve parameters according to the adaptation law to make the output value gradually approach the reference value and generate the basic control signal. The formula used is as follows:

[0050] u MRAC (t) = θ(t)·r(t);

[0051] Wherein, θ(t) represents the hydraulic valve parameters, u MRAC (t) represents the basic control signal, and r(t) represents the time-domain representation of the input signal;

[0052] Step S4: Strategy optimization. Use the DRL algorithm to construct a deep neural network as the optimization strategy for the adaptation law and generate the compensation signal. The formula used is as follows:

[0053]

[0054] Wherein, u DRL (t) represents the compensation signal, represents the optimization function, and s(t) represents the real-time data;

[0055] Step S5: Combine the basic control signal generated by the MRAC algorithm with the compensation signal adjusted by the DRL algorithm to generate the control signal. The formula used is as follows:

[0056] u(t) = u MRAC (t) + u DRL (t);

[0057] In the formula, u(t) represents the control signal.

[0058] By performing the above operations, for a controller of a general hydraulic valve control system that relies on fixed parameters, the control performance is poor under the conditions of parameter changes and uncertain dynamic characteristics, and there are problems of energy waste and low resource utilization efficiency. This solution designs a control strategy based on the MRAC algorithm and the DRL algorithm, enabling the hydraulic valve to approach the desired control state while increasing the learning ability and adaptability, improving the response speed, efficiency, and stability of the control system, and being applicable to various complex hydraulic scenarios.

[0059] Example 2, refer to Figure 1 , this example is based on the above example. Initially, the reference value and the real-time data monitored by the sensor are input into the hydraulic valve linkage control system, and the measurement error caused by the dynamic change of the data is transmitted to the controller. The controller generates a control signal through the control strategy, and the control signal is sent to the actuator in combination with the control error to achieve the control of the hydraulic valve. At the same time, there are modeling errors in the dynamic characteristics and nonlinear behavior generated by the complexity in the hydraulic system. The execution result of the actuator and the modeling error are combined as the output value and fed back to the hydraulic valve linkage control system through the sensor to form a closed-loop control structure.

[0060] By performing the above operations, for the traditional hydraulic system when facing rapidly changing working loads, the response speed is slow, it is difficult to adjust the hydraulic valve in a timely manner, and a large amount of manual intervention and debugging are required, resulting in high maintenance costs and high operation complexity. This solution uses a closed-loop feedback control system to ensure the stable operation of the hydraulic system. Through the interconnection of the sensor, controller, and actuator, the linkage between the hydraulic valve and the hydraulic cylinder in the hydraulic system is realized. By accurately controlling the parameters of the hydraulic valve in real time, the position of the hydraulic cylinder is precisely adjusted, improving the robustness and accuracy of the hydraulic system.

[0061] Example 3, refer to Figure 1 , this example is based on the above example. In step S1, a first-order system model is used as the reference model, and the formula used is as follows:

[0062]

[0063] In the formula, K m represents the gain of the reference model, and τ m represents the time constant that determines the system response speed;

[0064] When s = 0, t → ∞, the first-order system is in a steady state, and the value of the transfer function is K m , and at this time, the output signal is K m times the input signal.

[0065] Example 4, refer to Figure 1 , based on the above example, in step S2, the output signal is converted to the time domain using Laplace transform.

[0066] Example 5, refer to Figure 1 , based on the above example, in step S3, the adaptive law is designed using Lyapunov stability theory, and the formula used is as follows:

[0067]

[0068] In the formula, represents the rate of change of the hydraulic valve parameters with time, γ represents the adaptive gain, represents the regression vector.

[0069] Example 6, refer to Figure 1 and 2 , based on the above example, in step S4, the strategy optimization includes the following steps:

[0070] Step S41: State and action representation. The state of the optimized strategy is the real-time data collected by the sensor, and the action is the hydraulic valve parameter;

[0071] Step S42: Q-value estimation. A deep neural network is constructed as the Q function to calculate the Q value of the corresponding action in each state, representing the possibility that the output value reaches the reference value;

[0072] Step S43: Action selection. The ε-greedy strategy is used to select a random action with probability ε and the action with the highest current Q value with probability 1 - ε;

[0073] Step S44: Reward function design. The response speed of the control system is used as the reward signal, and a positive reward is given when the response speed is fast, and a negative reward is given when the response speed is slow;

[0074] Step S45: Parameter optimization. The weights of the deep neural network are adjusted using the stochastic gradient descent algorithm;

[0075] Step S46: Compensation signal generation. Steps S42 to S45 are the construction process of the deep neural network. The state is input into the deep neural network, and the corresponding compensation signal is generated according to the action selected each time;

[0076] In this example, the input layer of the deep neural network is set with 3 nodes, which respectively input three states of pressure, flow rate, and temperature; the hidden layer has 2 layers, and each hidden layer is set with 50 nodes; the output layer is set with 1 node, and the Q value of each action is output.

[0077] It should be noted that, in this document, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or further includes elements inherent to such process, method, article or device.

[0078] Although the embodiments of the present invention have been shown and described, those of ordinary skill in the art can understand that various changes, modifications, substitutions and variations can be made to these embodiments without departing from the principles and spirit of the present invention. The scope of the present invention is defined by the appended claims and their equivalents.

[0079] The above description of the present invention and its embodiments is not restrictive. What is shown in the drawings is only one of the embodiments of the present invention, and the actual structure is not limited thereto. In general, if those of ordinary skill in the art are inspired by it and, without departing from the purpose of the present invention, design similar structural modes and embodiments to this technical solution without creative efforts, they should all fall within the protection scope of the present invention.

Claims

1. A hydraulic valve linkage control system based on the Internet of Things, characterized in that: Including sensors, actuators, and controllers; The sensor collects real-time data of pressure, flow and temperature, and sends the obtained real-time data to the controller; The controller designs a control strategy based on the MRAC algorithm and the DRL algorithm, feeds back real-time data into the control strategy to generate a control signal, and sends the control signal to the actuator; The actuator drives the hydraulic valve to perform corresponding operations according to the control signal; The control strategy is designed based on the MRAC algorithm and the DRL algorithm, and includes the following steps: Step S1: Reference model design. A reference model is designed based on the MRAC algorithm to describe the desired position of the hydraulic cylinder. The desired hydraulic valve parameters are used as input signals. The input signals are calculated through the reference model to obtain output signals. The formula used is as follows: ； In the formula, represents the reference model, represents a complex frequency variable, represents the transfer function of the reference model, represents the frequency domain representation of the output signal, represents the frequency domain representation of the input signal; Step S2: Error calculation, convert the output signal into the time domain as a reference value, and control the operation of the hydraulic valve based on the reference value. The difference between the reference value and the output value is used as the control error to describe the difference between the actual hydraulic valve parameters and the expected hydraulic valve parameters. The formula used is as follows: ； In the formula, Indicates the time, represents the control error, Indicates the reference value, Indicates the output value; Step S3: Design an adaptive law, adjust the hydraulic valve parameters according to the adaptive law, make the output value gradually approach the reference value, and generate a basic control signal. The formula used is as follows: ； In the formula, Indicates hydraulic valve parameters, Represents the basic control signal, represents the time domain representation of the input signal; Step S4: Strategy optimization, using the DRL algorithm to construct a deep neural network as an optimization strategy for the adaptive law to generate a compensation signal. The formula used is as follows: ； In the formula, represents the compensation signal, represents the optimization function, Represents real-time data; Step S5: Generate a control signal by combining the basic control signal generated by the MRAC algorithm with the compensation signal adjusted by the DRL algorithm. The formula used is as follows: ； In the formula, Indicates control signal; In step S4, the strategy optimization includes the following steps: Step S41: state and action representation, the state of the optimization strategy is the real-time data collected by the sensor, and the action is the hydraulic valve parameter; Step S42: Q value estimation, constructing a deep neural network as a Q function, calculating the Q value of the corresponding action under each state, indicating the possibility of the output value reaching the reference value; Step S43: Action selection, using the ε-greedy strategy, with probability Choose a random action with probability Select the action with the highest current Q value; Step S44: reward function design, taking the response speed of the control system as a reward signal, giving a positive reward when the response speed is fast, and giving a negative reward when the response speed is slow; Step S45: parameter optimization, using a stochastic gradient descent algorithm to adjust the weights of the deep neural network; Step S46: compensation signal generation. Steps S42 to S45 are the process of building a deep neural network. The state is input into the deep neural network, and a corresponding compensation signal is generated according to each selected action.

Images