Sludge backflow intelligent control system of integrated sewage treatment equipment

Abstract

The invention discloses a sludge backflow intelligent control system of integrated sewage treatment equipment. The system comprises a human-computer interaction interface, a sludge detection module, a filtration control module, a backflow control module, a sludge parameter prediction module and a data storage module. Wherein the sludge detection module comprises pre-filtration and post-filtration detection. The filter control device adopted by the invention can control the on-off of the pump, so that the quality of filtered water can meet the required requirements. The sludge parameter prediction module used in the invention adopts a long-short-term memory network algorithm, sludge parameter information collected by a sensor is input into a neural network, and sludge parameters obtained after sewage is repeatedly filtered by an anaerobic tank are predicted, so that accurate regulation and control of the filtration control device are achieved. The problems that in the using process of existing integrated sewage treatment equipment, sludge content parameters cannot be accurately controlled, process parameters are uncontrollable, the parameters in the sludge are inaccurate, and the efficiency of the treatment process of the integrated sewage treatment equipment is low are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of sewage treatment, and in particular to a sludge return intelligent control system for an integrated sewage treatment device. Background Art

[0002] With the development of the times, the advancement of science and technology, the continuous improvement of living standards, and the continuous increase of domestic sewage. The use of integrated sewage treatment equipment can effectively treat domestic sewage, realize water circulation, and reduce water resource waste and environmental pollution.

[0003] The integrated sewage treatment equipment currently used is mainly controlled by PLC, and there is no equipment to control sewage filtration. This results in the process from the anaerobic tank to the sedimentation tank being uncontrollable, and the reflow efficiency is consistent. It cannot be adjusted accordingly based on changes in the surrounding environment and the sludge content. This will directly lead to poor sewage treatment effects and slower rates, resulting in waste of resources and increased costs. Summary of the invention

[0004] In view of the above problems, the present invention proposes a sludge return intelligent control system for an integrated sewage treatment equipment.

[0005] To achieve this goal, the plan is as follows: The present invention proposes a sludge return intelligent control system for an integrated sewage treatment equipment, comprising: Human-computer interaction interface; Sludge pre-filtration detection module; Sludge post-filtration detection module; Filter control module; Sludge parameter prediction module; Reflux control module; Data storage module; Among them, the human-computer interaction interface is connected with the sludge pre-filtration detection module, the sludge post-filtration detection module, the filtration control module, and the reflux control module; the sludge pre-filtration detection module and the sludge post-filtration detection module are respectively connected with the filtration control module and the sludge parameter prediction module; the filtration control module and the sludge parameter prediction module are connected with the reflux control module; the filtration control module and the sludge parameter prediction module are connected with each other; the data storage module is connected with the sludge pre-filtration detection module, the sludge post-filtration detection module, the filtration control module, the sludge parameter prediction module, and the reflux control module; the connection method between the above modules is all 485 communication interface method.

[0006] Furthermore, the sludge parameter prediction module adopts the model predictive control method to comprehensively consider the pH value, sludge depth, and suspended solids concentration of the sludge in the anaerobic tank and sedimentation tank; calculates the impact of the filtration system on the sludge parameters in the anaerobic tank and sedimentation tank in the integrated sewage treatment equipment, and controls the filtration control module to precisely control the filtration switch control by calculating the parameter influence value obtained.

[0007] Furthermore, the sludge parameter prediction module uses the long short-term memory network algorithm, i.e., the LSTM algorithm, to predict the parameter values ​​of the sludge after passing through the filter during the filtration system control process. The input of the model is the pH value, sludge depth, and solid suspended matter concentration parameters of the sludge in the anaerobic tank and sedimentation tank after normalization. These parameters are put into the long short-term memory neural network for training. After the training is completed, the model is used as an optimization function to output the optimal time and optimal duration of switching the filter in the filtration control module. This process includes the following steps: (1) Accept the sludge parameters collected by various sensors in the sludge pre-filtration detection module and the sludge post-filtration detection module, and perform fast Fourier transform (FFT) on these parameters. Since the characteristics of most signals in the time domain will be hidden, the signal is transformed from the time domain to the frequency domain through FFT, which is more conducive to training the network model. The calculation formula of FFT is as follows: ; The formulas for the amplitude spectrum and energy spectrum of Fourier transform are as follows: ; ; (2) Initialize the LSTM neural network first. Initialize the hidden state at the beginning of the sequence. and cell status ; Then loop each time step to perform calculations; The input is , respectively represented as cell state: carries information about the observation sequence and is transmitted between time steps; hidden state: contains information accumulated by the network before; input data: frequency domain data after FFT transformation of text data collected by the sludge detection sensor; (3) Perform forget gate calculation and discard some information. The formula is as follows: ; in, represents the output of the forget gate at time step t, which determines how much information is retained from the cell state. The sigmoid activation function maps the input to a value between 0 and 1. represents the weight matrix of the forget gate, represents the combination of the hidden state of the previous time step and the input of the current time step, Represents the bias term of the forget gate; This calculation process forgets the information through the sigmoid layer. And this step As input, then enter a value between 0 and 1 to assign to , where 0 means complete forgetting and 1 means complete retention; (4) Perform input layer gate calculation and store information. The formula is as follows: ; ; in, represents the output of the input gate at time step t. The input gate determines how much new information is stored in the cell state. The sigmoid activation function maps the input to a value between 0 and 1. represents the weight matrix of the input gate, represents the combination of the hidden state of the previous time step and the input of the current time step, represents the bias term of the input gate, represents the candidate cell state at time step t. The candidate cell state represents the information that may be added to the cell state in the current time step. Tanh represents the hyperbolic tangent activation function, which maps the input to a value between -1 and 1. and Represent the weight matrix and bias term of the candidate cell state respectively; This step determines the information content stored in the cell state and selectively adds new information to the cell state. It includes two steps: 1. The sigmoid layer determines the information that needs to be updated; 2. The tanh layer creates new candidate cell states, which are added to the cell state; these two steps are combined to update the cell state; Update the cell state, the formula is as follows: ; in, Indicates the state of the cell. Indicates the degree of forgetting the previous information state, represents the degree to which the candidate information value is added to the cell state; (5) Perform output gate calculations and output information, including hidden states. The formula is as follows: ; ; in, represents the output of the output gate at time step t, represents the sigmoid activation function, and Represent the weight matrix and bias term of the output gate respectively, represents the combination of the hidden state of the previous time step and the input of the current time step, represents the hidden state at time step t, represents the cell state at time step t; The output gate determines how much information in the cell state is related to the previous time step. The hidden state is the output of the entire neural network and is used to represent the information so far in the sequence. The output gate can regulate the flow of information from the cell state to the hidden layer. (6) Determine whether the expected effect can be achieved based on the output results and expected values.

[0008] Furthermore, the sludge parameter prediction model of the sludge parameter prediction module is established based on the LSTM long short-term memory network algorithm according to the historical operation data of the integrated sewage treatment equipment, wherein the sludge parameter prediction model takes the pH value, sludge depth, and solid suspended matter concentration parameters of the sludge in the anaerobic tank and the sedimentation tank as input, and takes the calculated filter switch and its duration as output.

[0009] Furthermore, the sludge pre-filtration detection module and the sludge post-filtration detection module include pH sensors, exposure sensors, sludge depth measurement sensors, and solid suspended matter concentration measurement sensors in the anaerobic tank and the sedimentation tank.

[0010] Furthermore, a filtering device is provided in the filtering control module. The filtering device is a device for treating sludge. Its main function is to remove solid particles and moisture from the sludge to reduce the volume of sludge, reduce processing costs and realize sludge resource utilization. The sludge filtering device usually includes a filtering medium. The sludge is pushed through the filtering medium by gravity, pressure or vacuum, the solid particles are retained, and the liquid is discharged through the medium.

[0011] Furthermore, the reflux control module is jointly controlled by the filtration control module and the sludge parameter prediction module. The opening and closing time of the reflux pump is regulated by the time series, and the obtained reflux data is provided to the sludge parameter prediction module for more in-depth network weight update.

[0012] Furthermore, the human-computer interaction interface has real-time monitoring, prediction and control functions, which can reflect the real-time status of sludge in the anoxic tank and sedimentation tank in the integrated sewage treatment equipment and the switch status of the filter and reflux pump, including various parameter monitoring data, predicted values, and alarm status. The interface can display data in the form of icons and curves, provide remote access and control functions, and can remotely monitor, predict and manage the equipment, so that the equipment can operate automatically and intelligently.

[0013] The present invention also provides a sludge return intelligent control system for an integrated sewage treatment device, comprising the following working steps: (1) Collect data on the pH value, sludge depth, and suspended solids concentration of sludge in the anaerobic tank and sedimentation tank of the integrated sewage treatment equipment; (2) Perform FFT preprocessing on the collected data to extract effective features; (3) Determine the LSTM neural network model structure and confirm the number of cells and the parameters of each gate based on the collected data; (4) Add actual data for training and adjustment, select part of the data to form a training set for model training and optimization, thereby improving the accuracy of the model; (5) Use the trained network model to predict sludge parameters in the sewage treatment process. Input the collected sludge parameter data from the anaerobic tank and sedimentation tank into the model, calculate and analyze it through the LSTM network model, and finally obtain the filter switch and its duration prediction value; (6) The predicted value is input into the sludge parameter prediction module to control the filter switch and its actual value of duration to achieve the purpose of precise temperature control. At the same time, the data is stored in the data storage module for analysis.

[0014] Compared with the prior art, the present invention designs a set of intelligent control systems for sludge filtration and return flow in the sewage treatment process, which can effectively predict the working time and working effect of the filter, and control the filter to accurately switch on and off to adjust the sludge content and its internal components, so as to achieve the purpose of improving the sewage treatment rate, effect and saving resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0015] Figure 1 It is an overall schematic diagram of the system of the present invention; Figure 2 This is a diagram of an integrated sewage treatment plant; Figure 3 This is a schematic diagram of the LSTM neural network structure; Figure 4 It is the flow chart of prediction algorithm; Figure 5 is the internal feedback diagram of a single cell; Figure 6 Detailed schematic diagram of the system of the present invention. Specific implementation methods

[0016] The following is a detailed description with reference to the accompanying drawings and the invention.

[0017] like Figure 1 As shown, a sludge return intelligent control system for an integrated sewage treatment equipment includes seven modules: a human-computer interaction interface X01, a sludge pre-filtration detection module X02, a sludge post-filtration detection module X03, a filtration control module X04, a sludge parameter prediction module X05, a return control module X06, and a data storage module X07.

[0018] Among them, the human-machine interaction interface X01 and the other five modules transmit information through 485 communication technology, with real-time monitoring, prediction and control functions, which can reflect the real-time status of sludge in the anoxic tank and sedimentation tank in the integrated sewage treatment equipment and the switch status of the filter and reflux pump in real time, including various parameter monitoring data, predicted values, and alarm status. The interface can display data in the form of icons and curves, provide remote access and control functions, and can remotely monitor, predict and manage the equipment, so that the equipment can operate automatically and intelligently.

[0019] The sludge parameter prediction module X05 in this system uses the long short-term memory network algorithm, namely the LSTM algorithm, to predict the parameter values ​​of the sludge after passing through the filter during the control process of the filtration system. The input of the model is the pH value, sludge depth, and solid suspended matter concentration parameters of the sludge in the anaerobic tank and sedimentation tank after normalization. These parameters are put into the long short-term memory neural network for training. After the training is completed, the model is used as an optimization function to output the optimal time and optimal duration of switching the filter in the filtration control module. This process includes the following steps: (1) Accept the sludge parameters collected by various sensors in the sludge pre-filtration detection module X02 and the sludge post-filtration detection module X03, and perform fast Fourier transform (FFT) on these parameters. Since the characteristics of most signals in the time domain will be hidden, the signal is converted from the time domain to the frequency domain through FFT, which is more conducive to training the network model.

[0020] The calculation formula of FFT is as follows: ; The formulas for the amplitude spectrum and energy spectrum of Fourier transform are as follows: ; ; (2) Initialize the LSTM neural network first. Initialize the hidden state at the beginning of the sequence. and cell status ; Then loop and calculate each time step.

[0021] The input is , respectively represented as cell state: carries information about the observation sequence and is transmitted between time steps; hidden state: contains the information accumulated by the network before; input data: frequency domain data after FFT transformation of text data collected by the sludge detection sensor.

[0022] (3) Perform forget gate calculation and discard some information. The formula is as follows: ; in, represents the output of the forget gate at time step t, which determines how much information is retained from the cell state. The sigmoid activation function maps the input to a value between 0 and 1. represents the weight matrix of the forget gate, represents the combination of the hidden state of the previous time step and the input of the current time step, Represents the bias term of the forget gate.

[0023] This calculation process forgets the information through the sigmoid layer. And this step As input, then enter a value between 0 and 1 to assign to , where 0 means complete forgetting and 1 means complete retention.

[0024] (4) Perform input layer gate calculation and store information. The formula is as follows: ; ; in, represents the output of the input gate at time step t. The input gate determines how much new information is stored in the cell state. The sigmoid activation function maps the input to a value between 0 and 1. represents the weight matrix of the input gate, represents the combination of the hidden state of the previous time step and the input of the current time step, represents the bias term of the input gate, represents the candidate cell state at time step t. The candidate cell state represents the information that may be added to the cell state in the current time step. Tanh represents the hyperbolic tangent activation function, which maps the input to a value between -1 and 1. and Represent the weight matrix and bias term of the candidate cell state respectively.

[0025] This step determines the information content stored in the cell state and selectively adds new information to the cell state. It includes two steps: 1. The sigmoid layer determines the information that needs to be updated; 2. The tanh layer creates new candidate cell states and adds them to the cell state; these two steps are combined to update the cell state.

[0026] Update the cell state, the formula is as follows: ; in, Indicates the state of the cell. Indicates the degree of forgetting the previous information state, Indicates the degree to which the candidate information value is added to the cell state.

[0027] (5 Perform output gate calculation and output information, including hidden state. The formula is as follows: ; ; in, represents the output of the output gate at time step t, represents the sigmoid activation function, and Represent the weight matrix and bias term of the output gate respectively, represents the combination of the hidden state of the previous time step and the input of the current time step, represents the hidden state at time step t, represents the cell state at time step t.

[0028] The output gate determines how much of the information in the cell state is related to the previous time step. The hidden state is the output of the entire neural network and is used to represent the information so far in the sequence; the output gate can regulate the flow of information from the cell state to the hidden layer.

[0029] (6) Determine whether the expected effect can be achieved based on the output results and expected values.

[0030] The integrated sewage treatment equipment involved in the present invention is also provided with a data storage module X07, which can store the sludge parameter prediction information and filter working status collected by the sludge parameter sensor, and is provided with a data export interface to facilitate subsequent maintenance personnel to extract the original data and provide certain data support for subsequent maintenance.

[0031] The sludge return intelligent control system of the integrated sewage treatment equipment designed by the present invention can continuously adjust the filter startup state during the sewage treatment process through the sludge parameters and filter states at different stages, improve the sewage treatment efficiency and effect, save energy, and overcome the existing traditional problems.

[0032] The above embodiments are only used to illustrate the specific implementation of the present invention. It should be noted that, for those skilled in the art, several modifications and changes can be made without departing from the concept of the present invention, and these modifications and changes should all fall within the protection scope of the present invention.

Claims

1. An intelligent control system for sludge return of an integrated sewage treatment plant, characterized in that: include: Human-computer interaction interface (X01); Sludge pre-filtration detection module (X02); Sludge post-filtration detection module (X03); Filter control module (X04); Sludge parameter prediction module (X05); Reflux control module (X06); Data storage module (X07); Among them, the human-computer interaction interface (X01) is connected to the sludge pre-filtration detection module (X02), the sludge post-filtration detection module (X03), the filtration control module (X04), and the reflux control module (X06); the sludge pre-filtration detection module (X02) and the sludge post-filtration detection module (X03) are respectively connected to the filtration control module (X04) and the sludge parameter prediction module (X05); the filtration control module (X04) and the sludge parameter prediction module (X05) are connected to the reflux control module (X06); the filtration control module (X04) and the sludge parameter prediction module (X05) are connected to each other; the data storage module (X07) is connected to the sludge pre-filtration detection module (X02), the sludge post-filtration detection module (X03), the filtration control module (X04), the sludge parameter prediction module (X05), and the reflux control module (X06); the connection method between the above modules is the 485 communication interface method.

2. The sludge return intelligent control system of the integrated sewage treatment equipment according to claim 1 is characterized in that: The sludge parameter prediction module (X05) adopts the model predictive control method to comprehensively analyze the pH value, sludge depth, and suspended solids concentration of the sludge in the anaerobic tank and sedimentation tank; calculates the impact of the filtration system on the sludge parameters in the anaerobic tank and sedimentation tank in the integrated sewage treatment equipment, and controls the filtration control module (X04) to precisely control the filtration switch control by calculating the parameter influence value obtained.

3. The sludge return intelligent control system of the integrated sewage treatment equipment according to claim 1 is characterized in that: The sludge parameter prediction module (X05) uses the long short-term memory network algorithm, i.e., the LSTM algorithm, to predict the parameter values ​​of the sludge after passing through the filter during the control process of the filtration system. The input of the model is the pH value, sludge depth, and solid suspended matter concentration parameters of the sludge in the anaerobic tank and sedimentation tank after normalization. These parameters are put into the long short-term memory neural network for training. After the training is completed, the model is used as an optimization function to output the optimal time and optimal duration of switching the filter in the filtration control module. This processing process includes the following steps: (1) Accept the sludge parameters collected by various sensors in the sludge pre-filtration detection module (X02) and the sludge post-filtration detection module (X03), and perform fast Fourier transform (FFT) on these parameters. Since most of the signal features in the time domain will be hidden, the signal is transformed from the time domain to the frequency domain through FFT, which is more conducive to training the network model. The calculation formula of FFT is as follows: ； The formulas for the amplitude spectrum and energy spectrum of Fourier transform are as follows: ； ； (2) Initialize the LSTM neural network first. Initialize the hidden state at the beginning of the sequence. and cell status ; Then loop each time step to perform calculations; The input is , respectively represented as cell state: carries information about the observation sequence and is transmitted between time steps; hidden state: contains information accumulated by the network before; input data: frequency domain data after FFT transformation of text data collected by the sludge detection sensor; (3) Perform forget gate calculation and discard some information. The formula is as follows: ； in, represents the output of the forget gate at time step t, which determines how much information is retained from the cell state. The sigmoid activation function maps the input to a value between 0 and 1. represents the weight matrix of the forget gate, represents the combination of the hidden state of the previous time step and the input of the current time step, Represents the bias term of the forget gate; This calculation process forgets the information through the sigmoid layer. And this step As input, then enter a value between 0 and 1 to assign to , where 0 means complete forgetting and 1 means complete retention; (4) Perform input layer gate calculation and store information. The formula is as follows: ； ； in, represents the output of the input gate at time step t. The input gate determines how much new information is stored in the cell state. The sigmoid activation function maps the input to a value between 0 and 1. represents the weight matrix of the input gate, represents the combination of the hidden state of the previous time step and the input of the current time step, represents the bias term of the input gate, represents the candidate cell state at time step t. The candidate cell state represents the information that may be added to the cell state in the current time step. Tanh represents the hyperbolic tangent activation function, which maps the input to a value between -1 and 1. and Represent the weight matrix and bias term of the candidate cell state respectively; This step determines the information content stored in the cell state and selectively adds new information to the cell state. It includes two steps:

1. The sigmoid layer determines the information that needs to be updated; 2. The tanh layer creates new candidate cell states, which are added to the cell state; these two steps are combined to update the cell state; Update the cell state, the formula is as follows: ； in, Indicates the state of the cell. Indicates the degree of forgetting the previous information state, represents the degree to which the candidate information value is added to the cell state; (5) Perform output gate calculations and output information, including hidden states. The formula is as follows: ； ； in, represents the output of the output gate at time step t, represents the sigmoid activation function, and Represent the weight matrix and bias term of the output gate respectively, represents the combination of the hidden state of the previous time step and the input of the current time step, represents the hidden state at time step t, represents the cell state at time step t; The output gate determines how much information in the cell state is related to the previous time step. The hidden state is the output of the entire neural network and is used to represent the information so far in the sequence. The output gate can regulate the flow of information from the cell state to the hidden layer. (6) Determine whether the expected effect can be achieved based on the output results and expected values.

4. The sludge return intelligent control system of the integrated sewage treatment equipment according to claim 3, characterized in that: The sludge parameter prediction model of the sludge parameter prediction module (X05) is established based on the LSTM long short-term memory network algorithm according to the historical operation data of the integrated sewage treatment equipment, wherein the sludge parameter prediction model takes the pH value, sludge depth, and solid suspended matter concentration parameters of the sludge in the anaerobic tank and the sedimentation tank as input, and takes the calculated filter switch and its duration as output.

5. The sludge return intelligent control system of the integrated sewage treatment equipment according to claim 1, characterized in that: The sludge pre-filtration detection module (X02) and the sludge post-filtration detection module (X03) include pH sensors, exposure sensors, sludge depth measurement sensors, and solid suspended matter concentration measurement sensors in the anaerobic tank and the sedimentation tank.

6. The sludge return intelligent control system of the integrated sewage treatment equipment according to claim 1, characterized in that: A filtering device is provided in the filtering control module (X04). The filtering device is a device used to treat sludge. Its main function is to remove solid particles and moisture from the sludge to reduce the volume of sludge, reduce processing costs and realize sludge resource utilization. The sludge filtering device usually includes a filtering medium. The sludge is pushed through the filtering medium by gravity, pressure or vacuum, and the solid particles are retained, while the liquid is discharged through the medium.

7. The sludge return intelligent control system of the integrated sewage treatment equipment according to claim 1, characterized in that: The reflux control module (X06) is jointly controlled by the filtration control module (X04) and the sludge parameter prediction module (X05). The opening and closing time of the reflux pump is regulated by time series. The obtained reflux data is provided to the sludge parameter prediction module (X05) for more in-depth network weight update.

8. The sludge return intelligent control system of the integrated sewage treatment equipment according to claim 1, characterized in that: The human-computer interaction interface (X01) has real-time monitoring, prediction and control functions. It can reflect the real-time status of sludge in the anoxic tank and sedimentation tank in the integrated sewage treatment equipment and the switch status of the filter and reflux pump, including various parameter monitoring data, predicted values, and alarm status. The interface can display data in the form of icons and curves, provide remote access and control functions, and can remotely monitor, predict and manage the equipment, so that the equipment can operate automatically and intelligently.

9. The sludge return intelligent control system of the integrated sewage treatment equipment according to claim 1, characterized in that: The following steps are included: (1) Collect data on the pH value, sludge depth, and suspended solids concentration of sludge in the anaerobic tank and sedimentation tank of the integrated sewage treatment equipment; (2) Perform FFT preprocessing on the collected data to extract effective features; (3) Determine the LSTM neural network model structure and confirm the number of cells and the parameters of each gate based on the collected data; (4) Add actual data for training and adjustment, select part of the data to form a training set for model training and optimization, thereby improving the accuracy of the model; (5) The trained network model is used to predict the sludge parameters in the sewage treatment process. The collected sludge parameter data in the anaerobic tank and sedimentation tank are input into the model, and the LSTM network model is used for calculation and analysis to finally obtain the filter switch and its duration prediction value; (6) The predicted value is input into the sludge parameter prediction module (X05) to control the filter switch and its actual value of duration, so as to achieve the purpose of precise temperature control. At the same time, the data is stored in the data storage module (X07) for analysis.

Images