Wireless flow regulating valve control method and system based on Internet of Things

Abstract

The invention discloses a wireless flow regulating valve control method and system based on the Internet of Things, and relates to the technical field of wireless flow regulating valves, and the method comprises the following steps: 1, collecting the flow, pressure, temperature and vibration data of a fluid in real time based on a wireless sensor node, and calculating the flow, pressure, temperature and vibration data of the fluid based on a data processing unit and a machine learning reasoning unit; 2, sensor data, local data summarization, preprocessing and partial control decision data are received through a gateway; 3, uploading the data to a cloud server based on a distributed gateway network module; step 4, based on the cloud server cluster module, adopting a deep reinforcement learning algorithm to carry out adaptive learning and optimization, generating a flow regulation instruction and sending the flow regulation instruction to a flow regulation valve; 5, based on the intelligent flow regulating valve module, the control model is adjusted in real time through a multi-dimensional feedback system; and step 6, performing global optimization and maintenance on the overall control strategy based on a multi-dimensional user interface module.

Description

Technical Field

[0001] The present invention relates to the technical field of wireless flow control valves, and in particular to a wireless flow control valve control method and system based on the Internet of Things. Background Art

[0002] The wireless flow control valve is a flow control device that integrates wireless communication technology. It can realize remote monitoring and regulation of flow without physical connection. This valve collects flow data in real time through built-in sensors and transmits the data to the monitoring center or user terminal through wireless modules, thereby realizing precise control of flow. Wireless flow control valves have the advantages of easy installation, high flexibility and low maintenance cost. They are widely used in water supply, heating and industrial fluid control. With the progress of industrial automation and the popularization of Internet of Things technology, the flow control system has gradually changed from traditional manual operation to intelligent and adaptive control.

[0003] However, the existing technologies often have the following problems when facing complex and changeable working conditions:

[0004] 1. Traditional flow regulation uses a simple algorithm and requires manual adjustment of the control valve control system. It cannot be controlled autonomously and has no automatic feedback mechanism, which further reduces the use effect and work efficiency of the system;

[0005] 2. Traditional flow regulation systems have poor accuracy and require regular calibration and maintenance. In addition, the implementation and debugging of precision control algorithms may require professional technicians, which increases the complexity and maintenance difficulty of the system. These factors may lead to a significant increase in the overall system cost.

[0006] 3. It is impossible to predict and maintain the system. When problems occur in the system, it is easy to cause system downtime. Once software or hardware failure occurs, professionals may be required to perform complex repairs and debugging, which increases the complexity and potential cost of maintenance.

[0007] Therefore, there is an urgent need for an intelligent supply chain management method and system that can dynamically adapt to market changes and optimize resource allocation to achieve efficient operation and sustainable development of the supply chain. Summary of the invention

[0008] The purpose of the present invention is to provide a wireless flow control valve control method and system based on the Internet of Things to solve the problems raised in the above background technology.

[0009] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0010] A wireless flow control valve control method based on the Internet of Things comprises the following steps:

[0011] The following steps are involved:

[0012] Step 1, wireless intelligent sensor node module: the wireless intelligent sensor node module includes a variety of sensors, data processing units and machine learning reasoning units. The wireless sensor nodes collect the flow, pressure and temperature data of the fluid in real time. Based on the data processing unit and the machine learning reasoning unit, data fusion and edge computing are performed, and preliminary prediction is performed based on the local linear regression model.

[0013] Step 2, distributed gateway network module: the distributed gateway network module includes multiple gateways, which receive sensor data, local data aggregation, preprocessing and some control decisions;

[0014] Step 3, cloud server cluster module: the cloud server cluster module includes a cloud server, which uploads data to the cloud server based on the distributed gateway network module to perform multi-objective optimization and global control decision-making;

[0015] Step 4, intelligent flow control valve module: the intelligent flow control valve adopts a deep reinforcement learning algorithm to perform adaptive learning and optimization based on the cloud server cluster module, generates a flow control instruction and sends it to the flow control valve;

[0016] Step 5, multi-dimensional user interface module: the multi-dimensional user interface is based on the intelligent flow control valve module and uses a multi-dimensional feedback system to adjust the control model in real time;

[0017] Step 6: System optimization and maintenance module, wherein the system optimization and maintenance module performs global optimization and maintenance on the overall control strategy based on the multi-dimensional user interface module.

[0018] A further improvement of the technical solution of the present invention is that the wireless intelligent sensor node module includes multiple sensors, temperature sensors, flow sensors, pressure sensors and vibration sensors, and the collected temperature is marked as T = {T1, T2..., T r}, where T r represents the rth data, r is a positive integer, and the flow mark is M = {M1, M2..., M l}, where M l Indicates the lth data, l is a positive integer, and the pressure is expressed by Y={Y1, Y2..., Y m}, where Y m represents the mth data, where m is a positive integer; the vibration information is represented by Z = {Z1, Z2..., Z a}, where Z a Represents the a-th piece of data, where a is a positive integer.

[0019] A further improvement of the technical solution of the present invention is that the wireless intelligent sensor node module performs data fusion and edge computing based on the collected data, and the data fusion formula is:

[0020]

[0021] Where D i represents the multidimensional data vector of the i-th sensor node, Δt represents the time interval or sampling time, V i1 、V i2 ...V in represents the measurement value of each sensor on the i-th sensor node, V ref1 、V ref2 ...V refn Indicates the reference value or expected value corresponding to each sensor measurement;

[0022] The data edge calculation formula is as follows:

[0023]

[0024] Where Q i (t+Δt) means the i-th node will be at the future time t+Δt i The predicted flow value within , Δt represents the time increment, k represents the past time step, K represents the total number of past time steps used in the regression analysis, and β k represents the regression coefficient at the kth time step, D i (tk) represents the input data of the i-th node at time tk, ∈ i represents the noise term.

[0025] A further improvement of the technical solution of the present invention is that the overall efficiency function formula of the supply chain is further expressed as:

[0026] The cloud server cluster module includes a multi-objective optimization model unit, which balances multiple control objectives based on a multi-objective optimization model. The multi-objective optimization model is specifically as follows:

[0027]

[0028] R(t)=-(Q target -Q actual (t)) 2 -λ7·E(t)

[0029] Where Q target , Q actualThey represent the target flow and actual flow respectively, R(t) and E(t) represent the response speed and energy consumption respectively, J(x) represents the optimization objective function, λ1, λ2, and λ3 represent weight coefficients, and X represents all adjustable control parameters in the system.

[0030] A further improvement of the technical solution of the present invention is that the intelligent flow control valve module includes an adaptive learning algorithm unit, and the adaptive learning algorithm unit system is based on an integrated deep reinforcement learning algorithm, and the specific formula is as follows:

[0031] Q(s,α)←Q(s,α)+α·[R(t)+γ·maxQ(s',α')-Q(s,α)]

[0032] Where Q(s,α) represents the expected reward value obtained by taking action a in state s, R(t) represents the feedback or reward obtained by the system for taking action a(t) at time t, α, γ, s′, α′ represent the learning rate, discount factor, state at the next moment, and action at the next moment, respectively.

[0033] A further improvement of the technical solution of the present invention is that: the multi-dimensional user interface module includes a dynamic adjustment mechanism unit and a multi-dimensional feedback system unit; the dynamic adjustment mechanism is dynamically adjusted in real time based on the intelligent flow control valve module, and the objective function of the dynamic adjustment mechanism is:

[0034]

[0035] Where Q predicted (t) represents the predicted flow value, u represents the input amount, λ4 represents the weight coefficient, and Δu(t) represents the change in the control input.

[0036] A further improvement of the technical solution of the present invention is that the multi-dimensional feedback system unit continuously modifies the control model based on the dynamic adjustment mechanism unit and the actual operation effect, and uses an online adaptive filter to revise it. The specific formula is as follows:

[0037] F(t)=F(t-1)+η·(L(t)-H(t)·F(t-1))

[0038] Where F(t) represents the filter parameters, η is the learning rate, which indicates the speed of updating the filter parameters, L(t) represents the prediction error of the system at time t, and H(t) represents the input data of the system at time t.

[0039] A further improvement of the technical solution of the present invention is that the system optimization and maintenance module includes a global optimization algorithm unit and a predictive maintenance unit. The global optimization algorithm unit performs global optimization on the overall control strategy by combining genetic algorithm and particle swarm optimization. The fitness function is:

[0040]

[0041] Where P is the particle position vector, which represents the current position of each particle (solution) in the particle swarm optimization algorithm, F(p) represents the fitness function, λ5 and λ6 represent the weights of the fitness function, and Q target , Q actual They represent the target flow and actual flow respectively, and E(t) represents the energy consumption of the system at time t.

[0042] A further improvement of the technical solution of the present invention is that the predictive maintenance unit uses a long short-term memory network to predict system failures based on historical operation data and real-time sensor data, and performs maintenance in advance. The prediction model is as follows:

[0043]

[0044] in represents the predicted state of the system at time t+1, X(t) represents the historical and current input data of the system at time t, W represents the connection weight of the neural network, and b represents the bias value of each neuron

[0045] A further improvement of the technical solution of the present invention is: a wireless flow control valve control system based on the Internet of Things, used to implement a wireless flow control valve control method based on the Internet of Things as described in any one of claims 1-9, characterized in that it includes a wireless intelligent sensor node module, a distributed gateway network module, a cloud server cluster module, an intelligent flow control valve module, a multi-dimensional user interface module and a system optimization and maintenance module and is electrically connected.

[0046] Due to the adoption of the above technical solution, the present invention has the following technical advances compared with the prior art:

[0047] 1. The present invention provides an innovative wireless flow control valve control method and system based on the Internet of Things. By introducing advanced multi-objective optimization and machine learning algorithms, the adaptive adjustment capability of the system can be improved. The algorithm enables the system to automatically adjust the control strategy when facing complex and changeable working conditions to ensure continuous high-performance output. The improvement of intelligence not only improves the response speed of the system, but also significantly enhances its adaptability to different working environments.

[0048] 2. The wireless flow control valve control method and system based on the Internet of Things provided by the present invention integrates a dynamic adjustment mechanism and multi-dimensional feedback control, so that the system can quickly sense and respond to changes in the external environment. Whether it is temperature, pressure or flow fluctuations, it can be adjusted promptly and effectively. The enhanced robustness ensures that the system can maintain a stable operating state under various extreme or uncertain conditions;

[0049] 3. The wireless flow control valve control method and system based on the Internet of Things provided by the present invention, by combining a multi-objective optimization algorithm and a genetic algorithm, our system achieves a significant reduction in energy consumption while ensuring control accuracy;

[0050] 4. The wireless flow control valve control method and system based on the Internet of Things provided by the present invention can accurately identify early signs of failure through the system optimization and maintenance module, thereby providing sufficient time for maintenance and avoiding the occurrence of failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0051] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0052] Figure 1 This is a flow chart of a wireless flow control valve control method and system based on the Internet of Things of the present invention; DETAILED DESCRIPTION

[0053] In order to make the purpose, technical solution and advantages of the embodiments of the present invention clearer, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0054] Embodiment 1, as Figure 1 As shown, the present invention provides a wireless flow regulating valve control system based on the Internet of Things, comprising the following steps:

[0055] Step 1, wireless intelligent sensor node module: The wireless intelligent sensor node module includes a variety of sensors, data processing units and machine learning reasoning units. The wireless sensor nodes collect the flow, pressure and temperature data of the fluid in real time. Based on the data processing unit and machine learning reasoning unit, data fusion and edge computing are performed, and preliminary predictions are made based on the local linear regression model.

[0056] Step 2, distributed gateway network module: The distributed gateway network module includes multiple gateways, which receive sensor data, local data aggregation, preprocessing and some control decisions;

[0057] The distributed gateway network achieves more efficient data processing and system management through the collaborative work of multiple gateways. The collaborative algorithm between gateways improves the speed and accuracy of data processing and enhances the robustness of the system, so that when some gateways fail, other gateways can continue to work, thereby ensuring the overall stability and reliability of the system. This design improves the scalability and processing capabilities of the system and adapts to application requirements in large-scale and complex environments.

[0058] Step 3, cloud server cluster module: The cloud server cluster module includes a cloud server, which uploads data to the cloud server based on the distributed gateway network module to perform multi-objective optimization and global control decision-making;

[0059] Step 4, intelligent flow control valve module: The intelligent flow control valve uses a deep reinforcement learning algorithm to perform adaptive learning and optimization based on the cloud server cluster module, generates a flow control instruction and sends it to the flow control valve;

[0060] The intelligent flow control valve has self-learning and self-diagnosis functions, and can adjust its own control strategy according to historical data. The self-learning and self-diagnosis functions of the intelligent flow control valve enable the system to dynamically adjust the control strategy according to the data in actual operation. This ability can improve the adaptability and accuracy of the valve under different working conditions, avoid the fixed parameter setting limitations of traditional flow control valves, and thus improve the flexibility and reliability of the system.

[0061] Step 5, multi-dimensional user interface module: the multi-dimensional user interface is based on the intelligent flow control valve module and uses a multi-dimensional feedback system to adjust the control model in real time;

[0062] Step 6, system optimization and maintenance module, the system optimization and maintenance module performs global optimization and maintenance on the overall control strategy based on the multi-dimensional user interface module.

[0063] By integrating multiple steps and technologies, the system is ensured to be highly intelligent and adaptable. Wireless sensor nodes provide real-time data, enabling the system to respond quickly to environmental changes. Edge computing reduces data transmission delays and improves the speed and accuracy of preliminary predictions. Multi-objective optimization in the cloud combines control decisions from a global perspective. Deep reinforcement learning enables the system to optimize control strategies through continuous learning and adjustment, thereby maintaining traffic stability and efficient use of energy. The multi-dimensional feedback system ensures the system's dynamic adjustment capabilities in actual operation and improves the system's robustness and flexibility.

[0064] The wireless intelligent sensor node module includes a variety of sensors, temperature sensors, flow sensors, pressure sensors and vibration sensors, and the collected temperature is marked as T = {T1, T2..., T r}, where T r represents the rth data, r is a positive integer, and the flow mark is M = {M1, M2..., M l}, where M l Indicates the lth data, l is a positive integer, and the pressure is expressed by Y={Y1, Y2..., Y m}, where Y m represents the mth data, where m is a positive integer; the vibration information is represented by Z = {Z1, Z2..., Z a}, where Z a Represents the a-th piece of data, where a is a positive integer.

[0065] The wireless intelligent sensor node module performs data fusion and edge computing based on the collected data. The data fusion formula is:

[0066]

[0067] Where D i represents the multidimensional data vector of the i-th sensor node, Δt represents the time interval or sampling time, V i1 、V i2 ...V in represents the measurement value of each sensor on the i-th sensor node, V ref1 、V ref2 ...V refn Indicates the reference value or expected value corresponding to each sensor measurement;

[0068] Where D i The data T, M, Y and Z measured by the temperature sensor, flow sensor, pressure sensor and vibration sensor are fused through the data fusion formula, which is conducive to the later processing of the data. The multi-dimensional data fusion algorithm provides more comprehensive and accurate fluid state information through comprehensive analysis of different sensor data. This algorithm can reduce the impact of sensor noise and data deviation and improve data reliability. This high-precision data fusion capability enables the system to perform more accurate flow prediction and control, thereby improving the overall performance and stability of the system.

[0069] The data edge calculation formula is as follows:

[0070]

[0071] Where Q i(t+Δt) means the i-th node will be at the future time t+Δt i The predicted flow value within , Δt represents the time increment, k represents the past time step, K represents the total number of past time steps used in the regression analysis, and β k represents the regression coefficient at the kth time step, D i (tk) represents the input data of the i-th node at time tk, ∈ i represents the noise term.

[0072] The cloud server cluster module includes a multi-objective optimization model unit. The multi-objective optimization model unit balances multiple control objectives based on a multi-objective optimization model. The multi-objective optimization model is as follows:

[0073]

[0074] Where Q target , Q actual They represent the target flow and actual flow respectively, R(t) and E(t) represent the response speed and energy consumption respectively, J(x) represents the optimization objective function, λ1, λ2, and λ3 represent weight coefficients, and X represents all adjustable control parameters in the system.

[0075] The optimization model comprehensively considers three key factors: flow target, response speed and energy consumption. By weighted combination of these targets, the secondary target can be optimized while meeting the primary target. This method provides a solution to balance different needs, so that the system can control the response speed and reduce energy consumption while achieving efficient flow regulation, thereby improving the performance and economic benefits of the overall system.

[0076] The intelligent flow control valve module includes an adaptive learning algorithm unit. The adaptive learning algorithm unit system is based on an integrated deep reinforcement learning algorithm. The specific formula is as follows:

[0077] Q(s,α)←Q(s,α)+α·[R(t)+γ·maxQ(s',α')-Q(s,α)]

[0078] R(t)=-(Q target -Q actual (t)) 2 -λ7·E(t)

[0079] Where Q(s,α) represents the expected reward value obtained by taking action a in state s, R(t) represents the feedback or reward obtained by the system for taking action a(t) at time t, α, γ, s′, α′ represent the learning rate, discount factor, state at the next moment, and action at the next moment, respectively.

[0080] The deep reinforcement learning algorithm is based on the Q-learning algorithm through the reward function

[0081] R(t)=-(Q target -Q actual (t)) 2 -λ7·E(t) is used to guide the learning process, so that the system can optimize the control strategy to maximize the long-term benefits. In the reward function, both the flow deviation and energy consumption are taken into consideration. In this way, we will not only strive to maintain flow stability, but also focus on reducing energy consumption. At the same time, we will ensure the efficiency and effectiveness of the system in practical applications, and avoid the waste of resources or poor performance that may result from pursuing a single goal.

[0082] The multi-dimensional user interface module includes a dynamic adjustment mechanism unit and a multi-dimensional feedback system unit: the dynamic adjustment mechanism is based on the intelligent flow control valve module to perform real-time dynamic adjustment, and the objective function of the dynamic adjustment mechanism is:

[0083]

[0084] Where Q predicted (t) represents the predicted flow value, u represents the input amount, λ4 represents the weight coefficient, and Δu(t) represents the change in the control input.

[0085] The algorithm can perform real-time optimization based on the prediction model and generate optimal control instructions. By considering the flow prediction error and control input changes, the algorithm can effectively respond to dynamic changes in the system and maintain flow stability. The penalty term in the objective function helps to control the smoothness of the input, thereby avoiding frequent adjustments and unnecessary operations, and improving the stability and control accuracy of the system.

[0086] The multi-dimensional feedback system unit continuously modifies the control model based on the dynamic adjustment mechanism unit and the actual operation effect, and uses an online adaptive filter to revise it. The specific formula is as follows:

[0087] F(t)=F(t-1)+η·(L(t)-H(t)·F(t-1))

[0088] Where F(t) represents the filter parameters, η is the learning rate, which indicates the speed of updating the filter parameters, L(t) represents the prediction error of the system at time t, and H(t) represents the input data of the system at time t.

[0089] The system optimization and maintenance module includes a global optimization algorithm unit and a predictive maintenance unit. The global optimization algorithm unit performs global optimization on the overall control strategy by combining genetic algorithm and particle swarm optimization. The fitness function is:

[0090]

[0091] Where P is the particle position vector, which represents the current position of each particle (solution) in the particle swarm optimization algorithm, F(p) represents the fitness function, λ5 and λ6 represent the weights of the fitness function, and Q target , Q actual They represent the target flow and actual flow respectively, and E(t) represents the energy consumption of the system at time t.

[0092] The combination of genetic algorithm and particle swarm optimization algorithm can provide powerful global search capabilities and avoid the problem of local optimal solutions. The weighted combination of multiple objectives in the fitness function enables the optimization process to consider both control accuracy and energy efficiency. This global optimization strategy not only improves the overall performance of the system, but also enables the system to adapt to different operating environments and conditions, with higher flexibility and stability.

[0093] The predictive maintenance unit uses a long short-term memory network to predict system failures based on historical operating data and real-time sensor data, and performs maintenance in advance. The prediction model is as follows:

[0094]

[0095] in represents the predicted state of the system at time t+1, X(t) represents the historical and current input data of the system at time t, W represents the connection weight of the neural network, and b represents the bias value of each neuron;

[0096] Long short-term memory networks have strong capabilities in time series data prediction and can capture long-term dependencies of system states. By predicting the future state of the system, they can detect potential failures in advance, thereby achieving predictive maintenance. This maintenance strategy reduces unplanned downtime, improves system reliability and availability, and reduces maintenance costs.

[0097] A wireless flow regulating valve control system based on the Internet of Things, used to implement a wireless flow regulating valve control method based on the Internet of Things according to any one of claims 1 to 9, comprising a wireless intelligent sensor node module, a distributed gateway network module, a cloud server cluster module, an intelligent flow regulating valve module, a multi-dimensional user interface module and a system optimization and maintenance module and an electrical connection.

[0098] Working principle: The staff uses a variety of sensors in the wireless intelligent sensor node module, temperature sensors, flow sensors, pressure sensors and vibration sensors to collect data from the flow control valve, and the collected temperature is marked as T = {T1, T2..., T r}, where T rrepresents the rth data, r is a positive integer, and the flow mark is M = {M1, M2..., M l}, where M l Indicates the lth data, l is a positive integer, and the pressure is expressed by Y={Y1, Y2..., Y m}, where Y m represents the mth data, where m is a positive integer; the vibration information is represented by Z = {Z1, Z2..., Z a}, where Z a Represents the a-th data, where a is a positive integer. Then, the data processing unit and the machine learning reasoning unit are used to perform data fusion and edge computing based on the data processing unit and the machine learning reasoning unit, and a preliminary prediction is made based on the local linear regression model. The data fusion formula is:

[0099]

[0100] Where D i represents the multidimensional data vector of the i-th sensor node, Δt represents the time interval or sampling time, V i1 、V i2 ...V in represents the measurement value of each sensor on the i-th sensor node, V ref1 、V ref2 ...V refn Represents the reference value or expected value corresponding to the measurement value of each sensor. The data edge calculation formula is as follows:

[0101]

[0102] Where Q i (t+Δt) means the i-th node will be at the future time t+Δt i The predicted flow value within , Δt represents the time increment, k represents the past time step, K represents the total number of past time steps used in the regression analysis, and β k represents the regression coefficient at the kth time step, D i (tk) represents the input data of the i-th node at time tk, ∈ i Represents the noise term, and then the processed data is received by the gateway in the distributed gateway network module to receive sensor data, local data aggregation, preprocessing and some control decision data. Then, based on the distributed gateway network module, the data is uploaded to the cloud server in the cloud server cluster module for multi-objective optimization and global control decision. The multi-objective optimization model is as follows:

[0103]

[0104] Where Q target , Qactual They represent the target flow and actual flow respectively, R(t) and E(t) represent the response speed and energy consumption respectively, J(x) represents the optimization objective function, λ1, λ2, and λ3 represent weight coefficients, and X represents all adjustable control parameters in the system.

[0105] Then, based on the cloud server cluster module, the deep reinforcement learning algorithm is used for adaptive learning and optimization. The adaptive learning algorithm unit system is based on the integrated deep reinforcement learning algorithm. The specific formula is as follows:

[0106] Q(s,α)←Q(s,α)+α·[R(t)+γ·maxQ(s',α')-Q(s,α)]

[0107] Where Q(s,α) represents the expected reward value obtained by taking action a in state s, R(t) represents the feedback or reward obtained by the system for taking action a(t) at time t, α, γ, s′, α′ represent the learning rate, discount factor, state at the next moment, and action at the next moment, respectively. A flow control instruction is generated and sent to the flow control valve, thereby realizing intelligent dynamic adjustment of the flow control valve. The dynamic adjustment mechanism is dynamically adjusted in real time based on the intelligent flow control valve module. The objective function of the dynamic adjustment mechanism is:

[0108]

[0109] Where Q predicted (t) represents the predicted flow value, u represents the input, λ4 represents the weight coefficient, and Δu(t) represents the change in the control input. After adjustment, based on the intelligent flow control valve module, the control model is adjusted in real time using the multi-dimensional feedback system, which is conducive to making the flow control valve more intelligent. The online adaptive filter is used to revise it. The specific formula is as follows:

[0110] F(t)=F(t-1)+η·(L(t)-H(t)·F(t-1))

[0111] Where F(t) represents the filter parameters, η is the learning rate, which represents the speed of updating the filter parameters, L(t) represents the prediction error of the system at time t, and H(t) represents the input data of the system at time t. At the same time, the system optimization and maintenance module is used to globally optimize and maintain the overall control strategy based on the multidimensional user interface module. The global optimization algorithm unit is used to globally optimize the overall control strategy by combining genetic algorithm and particle swarm optimization. The fitness function is:

[0112]

[0113] Where P is the particle position vector, which represents the current position of each particle (solution) in the particle swarm optimization algorithm, F(p) represents the fitness function, λ5 and λ6 represent the weights of the fitness function, and Q target , Q actual They represent the target flow and the actual flow respectively, and E(t) represents the energy consumption of the system at time t. The predictive maintenance unit uses a long short-term memory network to predict system failures based on historical operation data and real-time sensor data, and performs maintenance in advance. The prediction model is as follows:

[0114]

[0115] in represents the predicted state of the system at time t+1, X(t) represents the historical and current input data of the system at time t, W represents the connection weights of the neural network, and b represents the bias value of each neuron.

[0116] The above are only specific implementations of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A wireless flow control valve control method based on the Internet of Things, characterized in that: The following steps are involved: Step 1, wireless intelligent sensor node module: the wireless intelligent sensor node module includes a variety of sensors, a data processing unit and a machine learning reasoning unit. The wireless sensor node collects the flow, pressure and temperature data of the fluid in real time, performs data fusion and edge computing based on the data processing unit and the machine learning reasoning unit, and makes a preliminary prediction based on the local linear regression model; Step 2, distributed gateway network module: the distributed gateway network module includes multiple gateways, which receive sensor data, local data aggregation, preprocessing and some control decisions; Step 3, cloud server cluster module: the cloud server cluster module includes a cloud server, which uploads data to the cloud server based on the distributed gateway network module to perform multi-objective optimization and global control decision-making; Step 4, intelligent flow control valve module: the intelligent flow control valve adopts a deep reinforcement learning algorithm to perform adaptive learning and optimization based on the cloud server cluster module, generates a flow control instruction and sends it to the flow control valve; Step 5, multi-dimensional user interface module: the multi-dimensional user interface is based on the intelligent flow control valve module and uses a multi-dimensional feedback system to adjust the control model in real time; Step 6: System optimization and maintenance module, wherein the system optimization and maintenance module performs global optimization and maintenance on the overall control strategy based on the multi-dimensional user interface module.

2. According to the method for controlling a wireless flow control valve based on the Internet of Things in claim 1, it is characterized in that: The wireless intelligent sensor node module includes multiple sensors, temperature sensor, flow sensor, pressure sensor and vibration sensor, and the collected temperature is marked as T={T1, T2..., T r }, where T r represents the rth data, r is a positive integer, and the flow mark is M = {M1, M2..., M l }, where M l Indicates the lth data, l is a positive integer, and the pressure is expressed by Y={Y1, Y2..., Y m }, where Y m represents the mth data, where m is a positive integer; the vibration information is represented by Z = {Z1, Z2..., Z a }, where Z a Represents the a-th piece of data, where a is a positive integer.

3. According to the method for controlling a wireless flow regulating valve based on the Internet of Things in claim 2, it is characterized in that: The wireless intelligent sensor node module performs data fusion and edge computing based on the collected data. The data fusion formula is: Where D i represents the multidimensional data vector of the i-th sensor node, Δt represents the time interval or sampling time, V i1 、V i2 ...V in represents the measurement value of each sensor on the i-th sensor node, V ref1 、V ref2 ...V refn Indicates the reference value or expected value corresponding to each sensor measurement; The data edge calculation formula is as follows: Where Q i (t+Δt) means the i-th node will be at the future time t+Δt i The predicted flow value within , Δt represents the time increment, k represents the past time step, K represents the total number of past time steps used in the regression analysis, and β k represents the regression coefficient at the kth time step, D i (tk) represents the input data of the i-th node at time tk, and ∈i represents the noise term.

4. The method for controlling a wireless flow control valve based on the Internet of Things according to claim 1 is characterized in that: The cloud server cluster module includes a multi-objective optimization model unit, which balances multiple control objectives based on a multi-objective optimization model. The multi-objective optimization model is specifically as follows: R(t)=-(Q target -Q actual (t)) 2 -λ7·E(t) Where Q target , Q actual They represent the target flow and actual flow respectively, R(t) and E(t) represent the response speed and energy consumption respectively, J(x) represents the optimization objective function, λ1, λ2, and λ3 represent weight coefficients, and X represents all adjustable control parameters in the system.

5. The method for controlling a wireless flow regulating valve based on the Internet of Things according to claim 1 is characterized in that: The intelligent flow control valve module includes an adaptive learning algorithm unit. The adaptive learning algorithm unit system is based on an integrated deep reinforcement learning algorithm. The specific formula is as follows: Q(s,α)←Q(s,α)+α·[R(t)+γ·maxQ(s',α')-Q(s,α)] Where Q(s,α) represents the expected reward value obtained by taking action a in state s, R(t) represents the feedback or reward obtained by the system for taking action a(t) at time t, α, γ, s', α' represent the learning rate, discount factor, state at the next moment, and action at the next moment, respectively.

6. The method for controlling a wireless flow regulating valve based on the Internet of Things according to claim 1 is characterized in that: The multi-dimensional user interface module includes a dynamic adjustment mechanism unit and a multi-dimensional feedback system unit: the dynamic adjustment mechanism is dynamically adjusted in real time based on the intelligent flow control valve module, and the objective function of the dynamic adjustment mechanism is: Where Q predicted (t) represents the predicted flow value, u represents the input amount, λ4 represents the weight coefficient, and Δu(t) represents the change in the control input.

7. The method for controlling a wireless flow regulating valve based on the Internet of Things according to claim 1 is characterized in that: The multi-dimensional feedback system unit continuously modifies the control model based on the dynamic adjustment mechanism unit and the actual operation effect, and uses an online adaptive filter to revise it. The specific formula is as follows: F(t)=F(t-1)+η·(L(t)-H(t)·F(t-1)) Where F(t) represents the filter parameters, η is the learning rate, which indicates the speed of updating the filter parameters, L(t) represents the prediction error of the system at time t, and H(t) represents the input data of the system at time t.

8. The method for controlling a wireless flow regulating valve based on the Internet of Things according to claim 1 is characterized in that: The system optimization and maintenance module includes a global optimization algorithm unit and a predictive maintenance unit. The global optimization algorithm unit performs global optimization on the overall control strategy by combining genetic algorithm and particle swarm optimization. The fitness function is: Where P is the particle position vector, which represents the current position of each particle (solution) in the particle swarm optimization algorithm, F(p) represents the fitness function, λ5 and λ6 represent the weights of the fitness function, and Q target , Q actual They represent the target flow and actual flow respectively, and E(t) represents the energy consumption of the system at time t.

9. The method for controlling a wireless flow regulating valve based on the Internet of Things according to claim 8 is characterized in that: The predictive maintenance unit uses a long short-term memory network to predict system failures based on historical operation data and real-time sensor data, and performs maintenance in advance. The prediction model is as follows: in represents the predicted state of the system at time t+1, X(t) represents the historical and current input data of the system at time t, W represents the connection weights of the neural network, and b represents the bias value of each neuron.

10. A wireless flow control valve control system based on the Internet of Things, used to implement a wireless flow control valve control method based on the Internet of Things as claimed in any one of claims 1 to 9, characterized in that: It includes a wireless intelligent sensor node module, a distributed gateway network module, a cloud server cluster module, an intelligent flow control valve module, a multi-dimensional user interface module and a system optimization and maintenance module and electrical connections.

Images