Water supply control method and device and water quality detection and water supply control internet-of-things system
By constructing a water supply control model combining multi-layer neural network model with PID controller, the existing water supply control methods are solved in the water level control accuracy, and more accurate water supply control and intelligent water quality monitoring are achieved, ensuring water supply safety and management efficiency.
Patent Information
- Application Number
- CN202411580926.2
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-07
- Publication Date
- 2025-02-18
AI Technical Summary
The existing water supply control methods have insufficient water level control accuracy, making it difficult to ensure water supply safety. At the same time, water quality detection and water supply control lack intelligent and networked means.
采用多层神经网络模型(如SSA的NARX神经网络模型、SSA的BiGRU神经网络模型和Elman神经网络模型)结合PID控制器,构建供水控制模型,通过电流、转速和水位传感器数据进行实时检测和控制,实现更精确的水位调节和水质监控。
It improves the accuracy and stability of water supply control, ensures water supply safety, and at the same time, intelligent and networked management of water quality detection and water supply control is realized through the Internet of Things system.
Smart Images
Figure CN119472411A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to water quality detection and water supply control, and in particular to a water supply control method, a device and a water quality detection and water supply control Internet of Things system. Background Art
[0002] The existing water supply control methods mainly use PID, simple neural network models and other means to adjust the water level. There is a problem of low control accuracy, and it is difficult to accurately control the water level at the target value, which will affect the safety of water supply. On the other hand, water quality safety is also crucial when supplying water. At the same time, with the development of big data and Internet of Things technology, it has become an inevitable trend for water quality detection and water supply control to develop in an intensive, networked and intelligent direction. Summary of the invention
[0003] Purpose of the invention: The first purpose of the present invention is to provide a water supply control method with more precise water level control to ensure water supply safety; the second purpose of the present invention is to provide a water supply control device; the third purpose of the present invention is to provide a water quality detection and water supply control Internet of Things system.
[0004] Technical solution: The water supply control method of the present invention comprises:
[0005] (1) The current sensor, the speed sensor and the water level sensor are used to detect the current and the speed of the water pump driving device and the water level of the container respectively;
[0006] (2) Construct a water supply control model;
[0007] The water supply control model includes SSA's NARX neural network model-SSA's BiGRU neural network model 1-2, PID controller-SSA's BiGRU neural network model 1-6, water supply parameter processing module 1-3, SSA's Elman neural network model-SSA's BiGRU neural network model 1-2 and SSA's BiGRU neural network model-AdaBoost model 3-4;
[0008] The water level target value is used as the input of SSA's NARX neural network model - SSA's BiGRU neural network model 1. The difference and the difference change rate between the water level target value and the output of SSA's NARX neural network model - SSA's BiGRU neural network model 2 are used as the input of PID controller - SSA's BiGRU neural network model 1. The difference and the difference change rate between the output of PID controller - SSA's BiGRU neural network model 1 and the current decoupling value output by SSA's Elman neural network model - SSA's BiGRU neural network model 1 are used as the input of PID controller - SSA's BiGRU neural network model 2. The speed target value is 0 and the difference between SSA's Elman neural network model - SSA's BiGRU neural network model The difference and the rate of change of the speed decoupling value output by model 1 are used as the input of the BiGRU neural network model 3 of the PID controller-SSA, the output of the NARX neural network model-SSA BiGRU neural network model 1, the output of the PID controller-SSA BiGRU neural network model 2, the output of the PID controller-SSA BiGRU neural network model 3 and the output of the BiGRU neural network model-AdaBoost model 4 of SSA are used as the input of the BiGRU neural network model-AdaBoost model 3 of SSA respectively, the output of the BiGRU neural network model-AdaBoost model 3 of SSA is used as the input of the water pump driving device, and the output of the water pump driving device adjusts the water level of the container;
[0009] The current sensor output is used as the input of water supply parameter processing module 1 and SSA's Elman neural network model-SSA's BiGRU neural network model 1 respectively, the speed sensor output is used as the input of water supply parameter processing module 2 and SSA's Elman neural network model-SSA's BiGRU neural network model 1 respectively, the outputs of water supply parameter processing module 1 and water supply parameter processing module 2 are used as the input of SSA's Elman neural network model-SSA's BiGRU neural network model 2 respectively, the difference between the current decoupling prediction value output by SSA's Elman neural network model-SSA's BiGRU neural network model 2 and the current sensor output and the rate of change of the difference are used as the input of PID controller-SSA's BiGRU neural network model 4, the speed decoupling prediction value output by SSA's Elman neural network model-SSA's BiGRU neural network model 2 is used as the input of PID controller-SSA's BiGRU neural network model 4. The difference of the speed sensor output and the rate of change of the difference are used as the input of the BiGRU neural network model 5 of the PID controller-SSA, the output of the water level sensor is used as the input of the water supply parameter processing module 3, the difference between the output of the water supply parameter processing module 3 and the output of the water level sensor and the rate of change of the difference are used as the input of the BiGRU neural network model 6 of the PID controller-SSA, the BiGRU neural network model 4 of the PID controller-SSA, the BiGRU neural network model 5 of the PID controller-SSA, the BiGRU neural network model 6 of the PID controller-SSA and the output of the water level sensor are used as the input of the BiGRU neural network model-AdaBoost model 4 of the SSA respectively, and the output of the BiGRU neural network model-AdaBoost model 4 of the SSA is also used as the input of the NARX neural network model of the SSA-BiGRU neural network model 2 of the SSA;
[0010] The water supply parameter processing module is used to obtain the predicted value of the parameter sensor output.
[0011] Furthermore, the water supply parameter processing module includes SSA's BiGRU neural network model-AdaBoost model 1-2, SSA's NARX neural network model and SSA's Elman neural network model; the parameter sensor outputs are respectively used as inputs of SSA's BiGRU neural network model-AdaBoost model 1 and SSA's BiGRU neural network model-AdaBoost model 2, the outputs of SSA's BiGRU neural network model-AdaBoost model 1 are respectively used as inputs of SSA's NARX neural network model, SSA's Elman neural network model and SSA's BiGRU neural network model-AdaBoost model 2, the output of SSA's NARX neural network model is used as input of SSA's Elman neural network model, the output of SSA's Elman neural network model is used as input of SSA's BiGRU neural network model-AdaBoost model 2, and the output of SSA's BiGRU neural network model-AdaBoost model 2 is used as the predicted value of the parameter sensor output.
[0012] The water supply control device of the present invention comprises a memory and a processor. The memory stores a computer program. When the computer program is executed by the processor, the steps of the water supply control method are implemented.
[0013] The water quality detection and water supply control Internet of Things system described in the present invention is used to implement the steps of the water supply control method, including a perception control layer, a network layer, a water supply monitoring cloud platform and an application layer. The network layer is used to realize data interaction among the perception control layer, the water supply monitoring cloud platform and the application layer. The application layer is used to access the water supply monitoring cloud platform and send instructions to the perception control layer. The perception control layer is used to detect water quality and water level parameters and perform water level control. The detection data is uploaded to the water supply monitoring cloud platform.
[0014] Furthermore, the perception control layer is also used to detect pump room environmental parameters.
[0015] Furthermore, the perception control layer includes multiple water supply measurement and control units, which include an STM32 microprocessor, an environmental parameter sensor, a water quality parameter sensor, a water level sensor, a water pump drive device sensor, a wireless communication module and a control device. The water pump drive device sensor includes a current sensor and a speed sensor, and each sensor is connected to the STM32 microprocessor through a corresponding conditioning circuit; the wireless communication module and the control device are respectively connected to the STM32 microprocessor, and the wireless communication module is used to communicate with the water supply monitoring cloud platform. The control device includes a valve drive device, a sewage pump and a water pump drive device; the software of the water supply measurement and control unit mainly realizes wireless communication, data acquisition and preprocessing, and a water supply control model.
[0016] Furthermore, the environmental parameter sensor includes a camera, a temperature and humidity sensor, a flame sensor and a carbon dioxide sensor.
[0017] Furthermore, the water quality parameter sensors include a pH sensor, a water temperature sensor, a dissolved oxygen sensor and a residual chlorine sensor.
[0018] Furthermore, the application layer includes a water supply APP mobile terminal and a water supply monitoring terminal, which communicate with the water supply monitoring cloud platform through the 5G network; the main functions of the water supply monitoring terminal are communication parameter setting, collection data management and monitoring; the water supply APP mobile terminal mainly realizes the collection, processing, management and monitoring of parameters of the environment, water quality and water pump drive device.
[0019] Furthermore, the water supply monitoring terminal adopts an industrial control computer.
[0020] Beneficial effects: Compared with the prior art, the present invention has the following significant advantages:
[0021] (1) The SSA BiGRU neural network model-AdaBoost model is an Adaboost (adaptive boosting algorithm) ensemble learning method for the SSA BiGRU neural network model. A strong prediction model is formed by combining multiple SSA BiGRU neural network model weak classifiers. In each iteration, Adaboost makes the SSA BiGRU neural network model-AdaBoost model more focused on samples with larger prediction errors based on the weights of previous prediction samples, thereby improving the overall SSA BiGRU neural network model-AdaBoost model prediction performance and improving the SSA BiGRU neural network model-AdaBoost model prediction accuracy.
[0022] (2) The BiGRU neural network model is used to better capture the long-term dependencies in the time series data of water supply information. Considering both past and future water supply information at the same time helps to accurately analyze water supply information and improve the accuracy and stability of extracting long-term dependencies between water supply information.
[0023] (3) The NARX neural network model of SSA introduces the concept of input water supply information time series. During the calculation process, its own output will be fed back to the input end. The input end adds input delay and output feedback to the input end mechanism, which enhances the memory ability of the output water supply information historical data. Therefore, the NARX neural network model is suitable for water supply information time series prediction.
[0024] (4) The Elman neural network model adds a receiving layer to feed back the past water supply information of the hidden layer to the input layer, which can make full use of the water supply information at past moments to participate in the Elman neural network model prediction of the current water supply information, and to a certain extent maintain the dynamics, stability and prediction accuracy of the Elman neural network model. BRIEF DESCRIPTION OF THE DRAWINGS
[0025] Figure 1 is a structural schematic diagram of a water supply control model in an embodiment of the present invention;
[0026] Figure 2 is a structural schematic diagram of a water supply parameter processing module in an embodiment of the present invention;
[0027] Figure 3 Schematic diagram of the structure of the water quality detection and water supply control Internet of Things system in an embodiment of the present invention;
[0028] Figure 4 Schematic diagram of the structure of the water supply measurement and control unit in the embodiment of the present invention. DETAILED DESCRIPTION
[0029] The present invention will be further described below in conjunction with the accompanying drawings.
[0030] The embodiment of the present invention provides a water supply control method, which specifically includes the following steps:
[0031] (1) The current sensor, the speed sensor and the water level sensor are used to detect the current and the speed of the water pump driving device and the water level of the container respectively;
[0032] (2) Construct a water supply control model;
[0033] like Figure 1 As shown, the water supply control model includes SSA's NARX neural network model-SSA's BiGRU neural network model 1-2, PID controller-SSA's BiGRU neural network model 1-6, water supply parameter processing module 1-3, SSA's Elman neural network model-SSA's BiGRU neural network model 1-2 and SSA's BiGRU neural network model-AdaBoost model 3-4;
[0034] The water level target value is used as the input of SSA's NARX neural network model - SSA's BiGRU neural network model 1. The difference and the difference change rate between the water level target value and the output of SSA's NARX neural network model - SSA's BiGRU neural network model 2 are used as the input of PID controller - SSA's BiGRU neural network model 1. The difference and the difference change rate between the output of PID controller - SSA's BiGRU neural network model 1 and the current decoupling value output by SSA's Elman neural network model - SSA's BiGRU neural network model 1 are used as the input of PID controller - SSA's BiGRU neural network model 2. The speed target value is 0 and the difference between SSA's Elman neural network model - SSA's BiGRU neural network model The difference and the rate of change of the speed decoupling value output by model 1 are used as the input of the BiGRU neural network model 3 of the PID controller-SSA, the output of the NARX neural network model-SSA BiGRU neural network model 1, the output of the PID controller-SSA BiGRU neural network model 2, the output of the PID controller-SSA BiGRU neural network model 3 and the output of the BiGRU neural network model-AdaBoost model 4 of SSA are used as the input of the BiGRU neural network model-AdaBoost model 3 of SSA respectively, the output of the BiGRU neural network model-AdaBoost model 3 of SSA is used as the input of the water pump driving device, and the output of the water pump driving device adjusts the water level of the container;
[0035] The current sensor output is used as the input of water supply parameter processing module 1 and SSA's Elman neural network model-SSA's BiGRU neural network model 1 respectively, the speed sensor output is used as the input of water supply parameter processing module 2 and SSA's Elman neural network model-SSA's BiGRU neural network model 1 respectively, the output of water supply parameter processing module 1 and water supply parameter processing module 2 are used as the input of SSA's Elman neural network model-SSA's BiGRU neural network model 2 respectively, the difference between the current decoupling prediction value output by SSA's Elman neural network model-SSA's BiGRU neural network model 2 and the current sensor output and the rate of change of the difference are used as the input of PID controller-SSA's BiGRU neural network model 4, the difference between the speed decoupling prediction value output by SSA's Elman neural network model-SSA's BiGRU neural network model 2 and the speed sensor output and the difference The rate of change is used as the input of the BiGRU neural network model 5 of the PID controller-SSA, the output of the water level sensor is used as the input of the water supply parameter processing module 3, the difference between the output of the water supply parameter processing module 3 and the output of the water level sensor and the rate of change of the difference are used as the input of the BiGRU neural network model 6 of the PID controller-SSA, the outputs of the BiGRU neural network model 4 of the PID controller-SSA, the BiGRU neural network model 5 of the PID controller-SSA, the BiGRU neural network model 6 of the PID controller-SSA and the water level sensor are used as the input of the BiGRU neural network model-AdaBoost model 4 of the SSA respectively, and the output of the BiGRU neural network model-AdaBoost model 4 of the SSA is also used as the input of the NARX neural network model of the SSA-BiGRU neural network model 2 of the SSA; the water supply parameter processing module is used to obtain the predicted value of the parameter sensor output.
[0036] like Figure 2As shown, the water supply parameter processing module includes SSA's BiGRU neural network model-AdaBoost model 1-2, SSA's NARX neural network model and SSA's Elman neural network model; the parameter sensor outputs are respectively used as inputs of SSA's BiGRU neural network model-AdaBoost model 1 and SSA's BiGRU neural network model-AdaBoost model 2, the outputs of SSA's BiGRU neural network model-AdaBoost model 1 are respectively used as inputs of SSA's NARX neural network model, SSA's Elman neural network model and SSA's BiGRU neural network model-AdaBoost model 2, the output of SSA's NARX neural network model is used as input of SSA's Elman neural network model, the output of SSA's Elman neural network model is used as input of SSA's BiGRU neural network model-AdaBoost model 2, and the output of SSA's BiGRU neural network model-AdaBoost model 2 is used as the predicted value of the parameter sensor output.
[0037] 1. Design of water supply parameter processing module
[0038] The water supply parameter processing module includes SSA's BiGRU neural network model-AdaBoost model 1-2, SSA's NARX neural network model and SSA's Elman neural network model.
[0039] 1) BiGRU neural network model of SSA-AdaBoost model design
[0040] The SSA BiGRU neural network model-AdaBoost model is a SSA BiGRU neural network model that integrates weak learners through Adaboost. The Adaboost algorithm is used to integrate multiple SSA BiGRU neural network models. In each round of iteration, Adaboost adjusts the weights of samples based on previous prediction results, so that the SSA BiGRU neural network model pays more attention to samples with large prediction errors, thereby improving the prediction performance of the SSA BiGRU neural network model-AdaBoost model.
[0041] The steps of the BiGRU neural network model of Adaboost integrated SSA are as follows:
[0042] (1) Initialize sample weights. Randomly select m groups of training data in the sample space and initialize the weights of all samples to equal values.
[0043] (2) Iteratively train the BiGRU neural network model of SSA;
[0044] (3) Calculate the BiGRU neural network model weights for each SSA;
[0045] (4) updating the weight of the BiGRU neural network model of each SSA according to the prediction error and weight of each sample on the BiGRU neural network model of each SSA;
[0046] (5) The BiGRU neural network model of SSA is constructed based on the weights - the AdaBoost model.
[0047] The SSA BiGRU neural network model is a method for processing input water supply information time series data, learning the long-term dependencies in the water supply time series information, and can simultaneously process the forward and backward information of the input water supply time series. It involves two SSA GRU neural network models running in opposite directions along the input water supply information time series. This bidirectional input water supply information time series enables the BiGRU neural network model to simultaneously consider past and future water supply information, thereby more comprehensively understanding the dependencies between water supply information time series data.
[0048] The BiGRU neural network model of SSA (Sparrow Search Algorithm) is the SSA optimized BiGRU neural network model. The steps of the SSA optimized BiGRU neural network model are as follows:
[0049] (1) Initialize the BiGRU neural network model structure parameters and the individual fitness of the sparrow population;
[0050] (2) Update the location of the discoverer based on the warning value;
[0051] (3) Update the location of the joiner;
[0052] (4) The sparrows that are aware of danger update their positions. Sparrows at the periphery of the population move toward safe areas, while sparrows at the center of the population walk randomly to get closer to other sparrows.
[0053] (5) Calculate the fitness value of the new position of the individual sparrow, compare the updated fitness value with the original optimal fitness, and replace the inferior sparrow position with the optimal one to update the global optimal information;
[0054] (6) If the iteration number requirement is met, proceed to the next step, otherwise return to step (3);
[0055] (7) Output the optimal individual sparrow position, that is, the optimal parameter set for this BiGRU neural network model training.
[0056] 2) Design of NARX neural network model for SSA
[0057] The NARX neural network model of SSA is an SSA optimized NARX neural network model, and the method of SSA optimizing the NARX neural network model refers to the steps of SSA optimizing the BiGRU neural network model in the present invention.
[0058] The NARX neural network model introduces the concept of input water supply information time series. During the calculation process, it will feed back its own output to the input. Its input adds input delay and output feedback to the input mechanism, which enhances the memory ability of the output water supply information historical data. Therefore, the NARX neural network model is suitable for water supply information time series prediction.
[0059] The NARX neural network model consists of three parts: an input layer, a hidden layer and an output layer. The neurons between two adjacent layers are connected to each other through different weights. Considering the real-time and feasibility of the algorithm, the present invention only considers the NARX neural network model containing one hidden layer. The NARX neural network model includes a three-layer neural network model of one input layer, one hidden layer and one output layer, wherein the time delay layer (Time Delay Layer, TDL) is the water supply information input layer, which means that the data of the input water supply information and the output of the NARX neural network model are fed back to the input end data after a delay of several sampling cycles before participating in the NARX neural network model operation.
[0060] 3) Design of Elman neural network model for SSA
[0061] The Elman neural network model of SSA is an SSA optimized Elman neural network model, and the method of SSA optimizing the Elman neural network model refers to the steps of SSA optimizing the BiGRU neural network model in the present invention.
[0062] The Elman neural network model is a typical local regression neural network, which consists of four parts: input layer, hidden layer, output layer and receiving layer. The Elman neural network model adds a receiving layer to feed back the water supply information of the hidden layer to the input layer, which can make full use of the water supply information of the past moments to the Elman neural network model prediction of the current water supply information, and to a certain extent maintains the dynamics, stability and prediction accuracy of the Elman neural network model.
[0063] 2. Water supply control model design
[0064] The water supply control model includes SSA's NARX neural network model-SSA's BiGRU neural network model 1-2, PID controller-SSA's BiGRU neural network model 1-6, water supply parameter processing module 1-3, SSA's Elman neural network model-SSA's BiGRU neural network model 1-2 and SSA's BiGRU neural network model-AdaBoost model 3-4. The design method of the water supply parameter processing module 1-3 and the SSA's BiGRU neural network model-AdaBoost model 3-4 refers to the relevant design steps in the present invention.
[0065] 1) Design of NARX neural network model of SSA - BiGRU neural network model of SSA
[0066] The NARX neural network model of SSA - the BiGRU neural network model of SSA is a series connection of the NARX neural network model of SSA and the BiGRU neural network model of SSA. The design methods of the NARX neural network model of SSA and the BiGRU neural network model of SSA refer to the relevant design steps in the present invention.
[0067] 2) BiGRU neural network model design for PID controller-SSA
[0068] PID controller-SSA's BiGRU neural network model is a PID controller connected in series with SSA's BiGRU neural network model. The PID controller is mainly composed of proportion P, integral I and differential D. The controlled quantity is effectively controlled through proportion, integral and differential calculations. The basis of the PID controller is the proportional control of the controlled quantity. Integral control can effectively reduce the steady-state error of the controlled quantity, which is very likely to cause an increase in the overshoot of the controlled quantity. Differential control promotes the response speed of the large inertia system of the controlled object and effectively reduces the overshoot of the controlled quantity. The relationship between the input and output u(t) of the PID controller is:
[0069]
[0070] Where, e(t) represents the error input; u(t) represents the control output; K P Represents the proportionality coefficient; K I Represents the integral coefficient; K D represents the differential coefficient.
[0071] The design method of the BiGRU neural network model of SSA refers to the relevant design steps in the present invention.
[0072] 3) Elman neural network model of SSA-BiGRU neural network model design of SSA
[0073] The Elman neural network model of SSA-the BiGRU neural network model of SSA is the Elman neural network model of SSA and the BiGRU neural network model of SSA connected in series. The design method of the Elman neural network model of SSA and the BiGRU neural network model of SSA refers to the relevant design steps in the present invention.
[0074] An embodiment of the present invention further provides a water supply control device, including a memory and a processor, wherein a computer program is stored in the memory, and when the computer program is executed by the processor, the steps of the water supply control method described in the embodiment of the present invention are implemented.
[0075] like Figure 3 As shown, an embodiment of the present invention further provides a water quality detection and water supply control Internet of Things system, which is used to implement the steps of the water supply control method described in the embodiment of the present invention, including a perception control layer, a network layer, a water supply monitoring cloud platform and an application layer. The network layer is used to realize data interaction between the perception control layer, the water supply monitoring cloud platform and the application layer. The application layer is used to access the water supply monitoring cloud platform and issue instructions to the perception control layer. The perception control layer detects the pump room environmental parameters and water quality and water level parameters through multiple water supply measurement and control units, and performs water level control. The detection data is uploaded to the water supply monitoring cloud platform.
[0076] Specifically, Figure 4 As shown, the water supply measurement and control unit includes an STM32 microprocessor, an environmental parameter sensor, a water quality parameter sensor, a water level sensor, a water pump drive sensor, a wireless communication module and a control device, wherein the water pump drive sensor includes a current sensor and a speed sensor, the environmental parameter sensor includes a camera, a temperature and humidity sensor, a flame sensor and a carbon dioxide sensor, and the water quality parameter sensor includes a PH sensor, a water temperature sensor, a dissolved oxygen sensor and a residual chlorine sensor, and each sensor is connected to the STM32 microprocessor through a corresponding conditioning circuit. The wireless communication module and the control device are respectively connected to the STM32 microprocessor, and the network layer realizes data communication between the water supply measurement and control unit and the water supply monitoring cloud platform through the wireless communication module. The control device includes a valve drive device, a sewage pump and a water pump drive device. The software of the water supply measurement and control unit mainly realizes wireless communication, data acquisition and preprocessing, and a water supply control model.
[0077] The application layer includes the water supply APP mobile terminal and the water supply monitoring terminal. The water supply monitoring terminal uses an industrial control computer to communicate with the water supply monitoring cloud platform through the 5G network. It mainly realizes the collection, processing, management and monitoring of the parameters of the environment, water quality and water pump drive device. The main functions of the water supply monitoring terminal are communication parameter setting, collection data management and monitoring. The water supply APP mobile terminal realizes information exchange with the water supply monitoring cloud platform through the 5G network. The water supply APP mobile terminal mainly realizes the collection, processing, management and monitoring of the parameters of the environment, water quality and water pump drive device. Managers can access the water supply monitoring cloud platform through the water supply monitoring terminal or the water supply APP mobile terminal to realize environmental monitoring, water quality monitoring, water supply monitoring and environmental and water quality big data storage and management to ensure water supply safety.
Claims
1. A water supply control method, characterized in that: include: (1) The current sensor, the speed sensor and the water level sensor are used to detect the current and the speed of the water pump driving device and the water level of the container respectively; (2) Construct a water supply control model; The water supply control model includes SSA's NARX neural network model-SSA's BiGRU neural network model 1-2, PID controller-SSA's BiGRU neural network model 1-6, water supply parameter processing module 1-3, SSA's Elman neural network model-SSA's BiGRU neural network model 1-2 and SSA's BiGRU neural network model-AdaBoost model 3-4; The water level target value is used as the input of SSA's NARX neural network model - SSA's BiGRU neural network model 1. The difference and the difference change rate between the water level target value and the output of SSA's NARX neural network model - SSA's BiGRU neural network model 2 are used as the input of PID controller - SSA's BiGRU neural network model 1. The difference and the difference change rate between the output of PID controller - SSA's BiGRU neural network model 1 and the current decoupling value output by SSA's Elman neural network model - SSA's BiGRU neural network model 1 are used as the input of PID controller - SSA's BiGRU neural network model 2. The speed target value is 0 and the difference between SSA's Elman neural network model - SSA's BiGRU neural network model The difference and the rate of change of the speed decoupling value output by model 1 are used as the input of the BiGRU neural network model 3 of the PID controller-SSA, the output of the NARX neural network model-SSA BiGRU neural network model 1, the output of the PID controller-SSA BiGRU neural network model 2, the output of the PID controller-SSA BiGRU neural network model 3 and the output of the BiGRU neural network model-AdaBoost model 4 of SSA are used as the input of the BiGRU neural network model-AdaBoost model 3 of SSA respectively, the output of the BiGRU neural network model-AdaBoost model 3 of SSA is used as the input of the water pump driving device, and the output of the water pump driving device adjusts the water level of the container; The current sensor output is used as the input of water supply parameter processing module 1 and SSA's Elman neural network model-SSA's BiGRU neural network model 1 respectively, the speed sensor output is used as the input of water supply parameter processing module 2 and SSA's Elman neural network model-SSA's BiGRU neural network model 1 respectively, the outputs of water supply parameter processing module 1 and water supply parameter processing module 2 are used as the input of SSA's Elman neural network model-SSA's BiGRU neural network model 2 respectively, the difference between the current decoupling prediction value output by SSA's Elman neural network model-SSA's BiGRU neural network model 2 and the current sensor output and the rate of change of the difference are used as the input of PID controller-SSA's BiGRU neural network model 4, the speed decoupling prediction value output by SSA's Elman neural network model-SSA's BiGRU neural network model 2 is used as the input of PID controller-SSA's BiGRU neural network model 4. The difference of the speed sensor output and the rate of change of the difference are used as the input of the BiGRU neural network model 5 of the PID controller-SSA, the output of the water level sensor is used as the input of the water supply parameter processing module 3, the difference between the output of the water supply parameter processing module 3 and the output of the water level sensor and the rate of change of the difference are used as the input of the BiGRU neural network model 6 of the PID controller-SSA, the BiGRU neural network model 4 of the PID controller-SSA, the BiGRU neural network model 5 of the PID controller-SSA, the BiGRU neural network model 6 of the PID controller-SSA and the output of the water level sensor are used as the input of the BiGRU neural network model-AdaBoost model 4 of the SSA respectively, and the output of the BiGRU neural network model-AdaBoost model 4 of the SSA is also used as the input of the NARX neural network model of the SSA-BiGRU neural network model 2 of the SSA; The water supply parameter processing module is used to obtain the predicted value of the parameter sensor output.
2. The water supply control method according to claim 1, characterized in that: The water supply parameter processing module includes SSA's BiGRU neural network model-AdaBoost model 1-2, SSA's NARX neural network model and SSA's Elman neural network model; the parameter sensor outputs are respectively used as inputs of SSA's BiGRU neural network model-AdaBoost model 1 and SSA's BiGRU neural network model-AdaBoost model 2, the outputs of SSA's BiGRU neural network model-AdaBoost model 1 are respectively used as inputs of SSA's NARX neural network model, SSA's Elman neural network model and SSA's BiGRU neural network model-AdaBoost model 2, the output of SSA's NARX neural network model is used as input of SSA's Elman neural network model, the output of SSA's Elman neural network model is used as input of SSA's BiGRU neural network model-AdaBoost model 2, and the output of SSA's BiGRU neural network model-AdaBoost model 2 is used as the predicted value of the parameter sensor output.
3. A water supply control device, comprising a memory and a processor, wherein a computer program is stored in the memory, characterized in that: When the computer program is executed by a processor, the steps of the water supply control method according to claim 1 or 2 are implemented.
4. A water quality detection and water supply control Internet of Things system, used to implement the steps of the water supply control method according to claim 1 or 2, characterized in that: It includes perception control layer, network layer, water supply monitoring cloud platform and application layer. The network layer is used to realize data interaction among the perception control layer, water supply monitoring cloud platform and application layer. The application layer is used to access the water supply monitoring cloud platform and send instructions to the perception control layer. The perception control layer is used to detect water quality and water level parameters and perform water level control. The detection data is uploaded to the water supply monitoring cloud platform.
5. The water quality detection and water supply control Internet of Things system according to claim 4 is characterized in that: The perception control layer is also used to detect pump room environmental parameters.
6. The water quality detection and water supply control Internet of Things system according to claim 5 is characterized in that: The perception control layer includes multiple water supply measurement and control units, which include an STM32 microprocessor, an environmental parameter sensor, a water quality parameter sensor, a water level sensor, a water pump drive device sensor, a wireless communication module and a control device. The water pump drive device sensor includes a current sensor and a speed sensor, and each sensor is connected to the STM32 microprocessor through a corresponding conditioning circuit; the wireless communication module and the control device are respectively connected to the STM32 microprocessor, and the wireless communication module is used to communicate with the water supply monitoring cloud platform. The control device includes a valve drive device, a sewage pump and a water pump drive device; the software of the water supply measurement and control unit mainly realizes wireless communication, data acquisition and preprocessing, and a water supply control model.
7. The water quality detection and water supply control Internet of Things system according to claim 6 is characterized in that: The environmental parameter sensors include a camera, a temperature and humidity sensor, a flame sensor and a carbon dioxide sensor.
8. The water quality detection and water supply control Internet of Things system according to claim 6 is characterized in that: The water quality parameter sensors include a pH sensor, a water temperature sensor, a dissolved oxygen sensor and a residual chlorine sensor.
9. The water quality detection and water supply control Internet of Things system according to claim 6, characterized in that: The application layer includes a water supply APP mobile terminal and a water supply monitoring terminal, which communicate with the water supply monitoring cloud platform through a 5G network; the main functions of the water supply monitoring terminal are communication parameter setting, collection data management and monitoring; the water supply APP mobile terminal mainly realizes the collection, processing, management and monitoring of parameters of the environment, water quality and water pump drive device.
10. The water quality detection and water supply control Internet of Things system according to claim 9, characterized in that: The water supply monitoring terminal adopts an industrial control computer.