Integrated sewage treatment equipment air blowing stirring sludge intelligent control system

Abstract

The application discloses an integrated sewage treatment equipment air blowing stirring sludge intelligent control system, and comprises a man-machine interactive interface, a sludge detection sensing module, an aeration prediction module, an air blower control module and a data storage module; the modules are connected through a 485 communication interface.The application adopts a DRN network algorithm based on a residual shrinkage module, adopts sludge parameters collected through a Doppler double-frequency ultrasonic wave measurement method, predicts aeration stirring frequency and direction required in an integrated sewage treatment equipment aerobic tank, and thus controls the working power of an air blower and the direction of a nozzle to accurately aerate in the process.The application solves the problem that the existing integrated sewage treatment equipment cannot accurately control air blowing aeration through underwater sludge content and distribution during use, which leads to insufficient aeration, uneven stirring or excessive aeration, and serious energy waste.

Description

TECHNICAL FIELD

[0001] The present application relates to the sewage treatment technical field, specifically to the integrated sewage treatment equipment's blast stirring sludge intelligent control system. BACKGROUND

[0002] With the development of the times, the progress of science and technology, the living standard is improving, and the domestic sewage is also increasing. Using integrated sewage treatment equipment can effectively treat domestic sewage, realize water circulation, reduce water resource waste and environmental pollution.

[0003] The integrated sewage treatment equipment currently used is mainly controlled by PLC, the aeration disc in the aerobic tank is dual-purpose, oxygen is added to the tank and the sludge at the bottom of the tank is floated, the power of the air blower aeration is fixed, too little aeration will lead to insufficient stirring, resulting in poor sewage treatment effect, and too much aeration will lead to resource waste and cost increase. SUMMARY

[0004] In view of the above problems, the present application provides an integrated sewage treatment equipment's blast stirring sludge intelligent control system. The present application designs a set of integrated sewage treatment equipment's blast stirring sludge intelligent control system, which can effectively predict the working frequency of the air blower and control the air blower to carry out precise aeration, so as to improve the sewage treatment effect and save resources.

[0005] In order to achieve the purpose, the scheme is as follows:

[0006] The present application provides an integrated sewage treatment equipment's blast stirring sludge intelligent control system, which comprises: a man-machine interaction interface, a sludge detection sensor acquisition module, an aeration prediction module, an air blower control module and a data storage module.

[0007] The man-machine interaction interface is connected with the sludge detection sensor acquisition module, the aeration prediction module and the air blower control module; the sludge detection sensor acquisition module is connected with the aeration prediction module; the aeration prediction module is connected with the air blower control module; the data storage module is connected with the sludge detection sensor acquisition module, the aeration prediction module and the air blower control module; the connection mode between the above-mentioned modules is 485 communication interface mode.

[0008] Further, the aeration prediction module adopts a model prediction control method, which comprehensively considers a series of parameters such as PH value, exposure rate, mass distribution rate, sludge depth and solid suspended substance concentration in the aerobic tank; calculates the aeration amount required to blow up the sludge at the bottom of the integrated sewage treatment equipment, obtains the aeration rate through calculation, and controls the air blower control module to control the aeration process.

[0009] Furthermore, the aeration prediction module uses the residual contraction module's DRN neural network to predict the blower's operating intensity during the aeration process. The model's inputs are a series of parameters, including the normalized pH value, exposure rate, mass distribution rate, sludge depth, and suspended solids concentration in the aerobic tank. These parameters are trained on the DRN neural network. After training, the model is used as an optimization function to output the blower power required for aeration. This process includes the following steps:

[0010] (1) Receive the silt parameters and water quality parameters collected by each sensor in the silt detection sensor acquisition module, and perform fast Fourier transform (FFT) on these parameters. Since most of the signal features in the time domain will be hidden, converting the signal from the time domain to the frequency domain through FFT is more conducive to training the network model;

[0011] The calculation formula of FFT is as follows:

[0012] ;

[0013] The amplitude spectrum and energy spectrum of Fourier transform are:

[0014] ;

[0015] ;

[0016] (2) Convolutional layer input calculation: convolution operation is performed on the local feature X extracted from the input layer, and then the ReLU activation function is selected to process the feature:

[0017] ;

[0018] Among them, the convolution operation is used express, It is m The convolution layer k Output, i represents the number of convolution kernels, Representative convolutional layer m Bias;

[0019] (3) Classification is performed through the fully connected layer, and the output value of the output layer is calculated:

[0020] ;

[0021] in: It means the k Tier j The output of a neuron; and is the first in the fully connected layer k-1 Tier ineurons and k Tier j The weights and biases in each neuron;

[0022] (4) Solve the overfitting and gradient explosion problems through the residual module, making the residual term easier to train;

[0023] The output of the previous layer of the network is used as the input of the convolutional layer, and then enters the next convolutional layer through the activation function, and so on. Through the identity mapping, the final input is added to the initial output as the output of the residual module;

[0024] (5) Signal noise reduction through soft thresholding, x Represents the signal characteristics, T Represents the threshold value. The specific formula is as follows:

[0025] ;

[0026] (6) Determine whether the expected effect can be achieved based on the output results and expected values.

[0027] 300 sets of historical data were selected and allocated into training, validation, and test sets in a ratio of 7:3:1. This allocation is simpler and easier to implement, effectively avoiding model overfitting and improving the model's adaptability to new data and prediction accuracy. After the model is connected to the system, in addition to recording historical blower parameters and water quality parameters, the collected parameters are also fed into the established aeration prediction module to update and iterate the model, thereby increasing the accuracy of the prediction model.

[0028] The aeration prediction module for the integrated sewage treatment plant uses a DRN model algorithm based on historical operational data from the integrated sewage treatment plant. The aeration prediction model uses historical pH values, exposure rates, mass distribution rates, sludge depths, and suspended solids concentrations in the aerobic tank as inputs, and uses calculated blower operating power as output.

[0029] The intelligent control system of blast-stirring sludge of integrated sewage treatment equipment includes the following steps:

[0030] (1) Collect data on pH value, exposure rate, mass distribution rate, sludge depth, and suspended solids concentration in the aerobic sewage treatment tank;

[0031] (2) Perform FFT preprocessing on the collected data to extract effective features;

[0032] (3) Determine the DRN network model structure and determine the connections and input and output between layers based on the collected data;

[0033] (4) Add residual shrinkage module for training and adjustment. Select part of the data to form a training set to continue the training and optimization of the model, thereby improving the accuracy of the model;

[0034] (5) The trained network model is used for aeration prediction in the sewage treatment process. The collected sludge and water quality parameter data in the aerobic tank are input into the model, and the working power prediction value of the air blower is obtained through calculation and analysis of the DRN network model;

[0035] (6) The prediction value is input into the air blower control module (X04), so as to control the real working power of the air blower and achieve the purpose of precise aeration. At the same time, the data is stored in the data storage module (X05) for data analysis.

[0036] Compared with the prior art, the present application has the following beneficial effects: the present application solves the problem that the existing integrated sewage treatment equipment cannot accurately control the air blowing aeration through the underwater sludge content and distribution during use, resulting in insufficient aeration, uneven stirring or excessive aeration, causing serious energy waste. BRIEF DESCRIPTION OF DRAWINGS

[0037] Figure 1 is the overall design drawing of the present application;

[0038] Figure 2 is the DRN network structure diagram;

[0039] Figure 3 is the residual module structure diagram;

[0040] Among them, X01 is a man-machine interaction interface, X02 is a sludge detection sensor collection module, X03 is an aeration prediction module, X04 is an air blower control module, and X05 is a data storage module. DETAILED EMBODIMENT

[0041] The application will be described in detail below in conjunction with the drawings.

[0042] Figure 1 is the overall design drawing of a kind of air blowing stirring sludge intelligent control system applied to integrated sewage treatment equipment, the system can control air blower working intensity, thereby precise aeration is carried out. This system includes five modules of man-machine interaction interface X01, sludge detection sensor collection module X02, aeration prediction module X03, air blower control module X04 and data storage module X05.

[0043] The human-computer interaction interface X01 transmits information with the other four modules through 485 communication technology. The interface has real-time monitoring, prediction and control functions, and can display the real-time state of the sludge in the aerobic tank and the running state of the air blower in the integrated sewage treatment equipment, including various parameter monitoring data, prediction values, alarm states, etc. The interface can display data in the form of charts and curves, provide remote access and control functions, and enable remote monitoring, prediction and management of the equipment, so that the equipment can be automatically and intelligently operated.

[0044] The aeration prediction module X03 in the system uses the DRN neural network of the residual shrinkage module to predict the working intensity of the air blower during the aeration process. The input of the model is a series of parameters such as pH value, exposure rate, mass distribution rate, sludge depth and suspended solid concentration in the aerobic tank after normalization processing. These parameters are put into the DRN neural network for training. After training, the model is used as an optimization function to output the required air blower power. This processing process includes the following steps:

[0045] (1) Accept the sludge parameters and water quality parameters collected by each sensor in the sludge detection sensor collection module, and perform fast Fourier transform FFT on these parameters. Since most of the signal characteristics in the time domain are hidden, it is more conducive to training the network model to convert the signal from the time domain to the frequency domain through FFT;

[0046] The calculation formula of FFT is as follows:

[0047] ;

[0048] The amplitude spectrum and energy spectrum of the Fourier transform are:

[0049] ;

[0050] ;

[0051] (2) Convolution layer input calculation: perform convolution operation on the local features X extracted in the input layer, and then select ReLU activation function to process the features:

[0052] ;

[0053] Wherein, the convolution operation is represented by , is the output of the th layer convolution, m is the k th output of the i th layer convolution, represents the number of convolution kernels, m represents the bias of the convolution layer

[0054] (3) Classification is carried out through a full connection layer, and the output value calculation of the output layer is:

[0055] ;

[0056] Among them: Represents the output of the first k Layer of the first j Neuron; And The weight and bias in the first k-1 Layer of the first i Neuron and the first k Layer of the first j Neuron in the full connection layer;

[0057] (4) The problem of overfitting and gradient explosion is solved through a residual module, so that the residual term is easier to train;

[0058] As shown in Figure 3 , wherein: the input of the front layer of the network is represented by x ; H(x) represents the output of the next layer of the network; F(x) is used to represent the residual mapping function; Weight Layer is a convolutional layer; ReLU is an activation function; identity represents an identity mapping;

[0059] (5) Signal denoising is carried out through soft thresholding, x Represents the signal feature, T Represents the threshold, and the specific formula is as follows:

[0060] ;

[0061] According to the output result and the expected value, it is judged whether the expected effect can be achieved.

[0062] The aeration prediction model X03 is established based on the DRN model algorithm according to the historical operation data of the integrated sewage treatment equipment, wherein the aeration prediction model takes the historical PH value, the historical exposure rate, the historical quality distribution rate, the historical sludge depth and the historical solid suspended substance concentration parameters in the aerobic tank as inputs, and takes the calculated working power of the air blower as output.

[0063] The integrated sewage treatment equipment disclosed by the application further comprises a data storage module X05, which can store the parameters collected by the sludge detection sensor, the aeration prediction information and the air blower state information, and is provided with a data export interface, so that the original data can be extracted by subsequent maintenance personnel, and certain data support is provided for subsequent maintenance.

[0064] The integrated sewage treatment equipment air blowing stirring sludge intelligent control system designed by the application can continuously adjust the aeration amount in the sewage treatment process through the air blower state and water quality parameters in different stages, improve the sewage treatment effect, and save energy, so as to overcome the existing traditional problems.

[0065] The above examples are only used to illustrate the specific embodiments of the application. It should be noted that for those skilled in the art, without departing from the concept of the application, several modifications and changes can be made, and these modifications and changes should belong to the protection scope of the application.

Claims

1. An intelligent control system for air-blowing and stirring sludge of an integrated sewage treatment device, comprising: a human-computer interaction interface (X01); a sludge detection sensor collection module (X02); an aeration prediction module (X03); an air-blower control module (X04); a data storage module (X05); wherein the human-computer interaction interface (X01) is connected with the sludge detection sensor collection module (X02), the aeration prediction module (X03), and the air-blower control module (X04); the sludge detection sensor collection module (X02) is connected with the aeration prediction module (X03); the aeration prediction module (X03) is connected with the air-blower control module (X04); the data storage module (X05) is connected with the sludge detection sensor collection module (X02), the aeration prediction module (X03), and the air-blower control module (X04); and the connection mode between the above modules is a 485 communication interface mode; the aeration prediction module (X03) uses a DRN neural network model of a residual shrinkage module to predict the working intensity of the air-blower during the aeration process, the input of the model is the pH value, exposure rate, mass distribution rate, sludge depth, and solid suspended substance concentration parameters in the aerobic tank after normalization processing, the parameters are put into the DRN neural network model for training, and after the training is completed, the model is used as an optimization function to output the air-blower power required for aeration, and the processing process includes the following steps: (1) receiving the sludge parameters and water quality parameters collected by each sensor in the sludge detection sensor collection module (X02), performing fast Fourier transform (FFT) on the parameters, converting the signals from the time domain to the frequency domain through FFT, which is more conducive to training the network model, because the characteristics of most signals in the time domain are hidden; the calculation formula of FFT is as follows: ； the amplitude spectrum and energy spectrum of Fourier transform are as follows: ； ； (2) input calculation of the convolution layer, performing convolution operation on the local features X extracted in the input layer, and selecting ReLU activation function to process the features: ； Among them, the convolution operation is used express, It is m The convolution layer k Output, i represents the number of convolution kernels, Representative convolutional layer m Bias; (3) classifying through the full connection layer, and calculating the output value of the output layer: ； where: represents the output of the k layer i-th neuron; j and is the weight and bias in the fully connected layer between the k-1 layer i-th neuron and the k layer j-th neuron.​ (4) solving the overfitting and gradient explosion problems through the residual module, so that the residual term is easier to train: the output of the previous layer of the network is used as the input of the convolution layer, and then enters the next convolution layer through the activation function, and so on, the last input is added to the initial output through the identity mapping, and the output is used as the output of the residual module; (5) Signal denoising is performed by soft thresholding, x denotes a signal feature, T denotes a threshold, and the specific formula is as shown below: ； (6) judging whether the expected effect can be achieved according to the output result and the expected value.

2. The integrated sewage treatment equipment's intelligent control system for air-blast agitator sludge according to claim 1, characterized in that, The aeration prediction model of the aeration prediction module (X03) is established based on the DRN model algorithm according to the historical operation data of the integrated sewage treatment device, wherein the aeration prediction model takes the historical pH value, historical exposure rate, historical mass distribution rate, historical sludge depth, and historical solid suspended substance concentration parameters in the aerobic tank as inputs, and takes the calculated air-blower working power as output.

3. The integrated sewage treatment equipment's intelligent control system for air-blast agitator sludge according to claim 1, characterized in that, The sludge detection sensor collection module (X02) includes a pH sensor, an exposure sensor, a mass distribution sensor, a sludge depth measurement sensor, and a solid suspended substance concentration measurement sensor in the aerobic tank.

4. The integrated sewage treatment plant air-blast sludge mixing intelligent control system as claimed in claim 1, wherein, The human-computer interaction interface (X01) has real-time monitoring, prediction and control functions, can display the real-time state of the sludge in the aerobic tank and the running state of the air blower in the integrated sewage treatment equipment in real time, including various parameter monitoring data, prediction values and alarm states, the interface can display data in the form of charts and curves, provides remote access and control functions, can remotely monitor, predict and manage the equipment, and enables the equipment to automatically and intelligently run.

5. The integrated sewage treatment plant air-blast sludge mixing intelligent control system of claim 1, wherein the features include, The method comprises the following steps: (1) collecting various parameter data of pH value, exposure rate, mass distribution rate, sludge depth and solid suspended substance concentration in the aerobic tank of the sewage treatment; (2) performing FFT preprocessing on the collected data and extracting effective features; (3) determining the structure of the DRN network model, determining the connection and input and output among layers according to the collected data; (4) adding a residual shrinkage module for training and adjustment, selecting part of the data to form a training set to continue the training and optimization of the model, thereby improving the accuracy of the model; (5) using the trained network model to predict aeration in the sewage treatment process, inputting the collected sludge and water quality parameter data in the aerobic tank into the model, performing calculation and analysis through the DRN network model, and finally obtaining the working power prediction value of the air blower; (6) inputting the prediction value into the air blower control module (X04) to control the real working power of the air blower, achieving the purpose of precise aeration, and storing the data into the data storage module (X05) for data analysis.

Images