Sludge discharge control system and method for sewage treatment

Abstract

The invention relates to the technical field of sewage treatment, in particular to a sludge discharge control system for sewage treatment and a method thereof, the sludge discharge control system comprises a data acquisition module, a data processing module, a neural network module and a control module, the data acquisition module is in signal connection with the data processing module, and the neural network module is in signal connection with the control module. The data processing module and the control module are both in signal connection with the neural network module. The method overcomes the defects of high labor cost and inaccurate control of the sludge discharge amount due to manual judgment and regulation of the sludge discharge amount of the biochemical treatment system in the prior art, replaces manual prediction of the sludge discharge amount and controls the biochemical treatment system to discharge sludge according to the predicted sludge discharge amount, can reduce the labor cost, and improves the sludge discharge efficiency. The accuracy of sludge discharge amount control can be improved, so that the chemical consumption is reduced, and the sewage treatment effect is improved. In addition, arrangement and scheduling of the mud truck can be assisted.

Description

Technical Field

[0001] The present invention relates to the technical field of sewage treatment, and more specifically, to a sludge discharge control system and method for sewage treatment. Background Art

[0002] The activated sludge process is one of the most widely used processes in urban sewage treatment plants. Activated sludge is the general term for the microbial community in the biochemical tank and the organic and inorganic substances attached to them, and its function is to promote the growth and reproduction of microorganisms, thereby effectively removing organic substances and nutrients in sewage. The concentration of sludge must be controlled within a certain range, and too large or too small will affect the sewage treatment effect. By reasonably controlling the sludge concentration, the treatment efficiency and stability of the biochemical tank can be effectively improved, ensuring that the sewage treatment meets the expected effect.

[0003] Due to the incomplete separation of urban rainwater and sewage, there will be two different operating conditions, namely the rainy season / dry season, during normal operation. During the rainy season, the influent COD (chemical oxygen demand contained in rainwater when it enters the biochemical treatment system) will drop significantly, and a large amount of SS (suspended solids in water) will be brought in by rainfall scouring the ground. All of these will cause a large change in the sludge concentration of the biochemical treatment system. If the sludge concentration in the biochemical tank is not adjusted in time, the effluent quality will be affected.

[0004] The sludge discharge control of multiple production lines in the existing sewage treatment plant and the scheduling of sludge transportation vehicles rely on the judgment and regulation of production scheduling technicians. However, there are several defects in the way of manual judgment and regulation. First, the labor cost required for manual judgment and regulation is relatively high. Second, it takes a certain amount of time and manpower to train new staff when they start work. Third, due to the strong subjectivity and experience dependence of manual work, technicians with insufficient experience are very likely to cause inaccurate control of the sludge discharge volume, resulting in an increase in drug consumption or a decrease in sewage treatment effect. Summary of the Invention

[0005] Aiming at the problem in the above-mentioned existing technology that the sludge discharge volume of the biochemical treatment system is judged and regulated manually, which not only leads to a relatively high labor cost, but also the control of the sludge discharge volume is not accurate enough, the present invention provides a sludge discharge control system and method for sewage treatment, which can realize the intelligent regulation of the sludge discharge volume, reduce the labor cost, and improve the accuracy of the sludge discharge volume regulation.

[0006] To solve the above technical problems, the technical solution provided by the present invention is:

[0007] A sludge discharge control system for sewage treatment, which includes a data acquisition module, a data processing module, a neural network module, and a control module. The data acquisition module is signal-connected to the data processing module, and both the data processing module and the control module are signal-connected to the neural network module.

[0008] When the above technical solution is implemented, the data acquisition module acquires relevant data and sends the data to the neural network module. After receiving the data, the neural network module constructs and trains a neural network model based on the received data. Then, the data acquisition module acquires data again and sends it to the neural network module. The neural network module then calculates the predicted number of sludge trucks using the neural network model and sends a signal to the control module. After receiving the signal, the control module controls the opening and closing of the sludge pump and sludge valve of the biochemical treatment system according to the average daily sludge discharge volume corresponding to the predicted number of sludge trucks to control the sludge discharge of the biochemical treatment system.

[0009] Preferably, it further includes a data optimization module, and both the data acquisition module and the data processing module are signal-connected to the data optimization module. The data acquisition module first acquires relevant data and sends it to the data optimization module. After the data optimization module removes the outliers in the data, it then sends the data to the neural network module. This can improve the accuracy of the data and is conducive to improving the prediction accuracy of the neural network model.

[0010] The present invention also provides a sludge discharge control method for sewage treatment, which is applied to the aforementioned sludge discharge control system for sewage treatment, and includes the following steps:

[0011] S1: The data acquisition module acquires multiple groups of data. Each group of data includes the drying production capacity, the water volume before sludge discharge treatment, the water volume before sludge discharge, the 5-day biochemical oxygen demand of the influent, the chemical oxygen demand of the influent, the suspended solid concentration of the influent, the average concentration of volatile suspended solids in the influent mixed liquor, the sludge concentration before sludge discharge, the actual number of sludge trucks transported, and the sludge concentration after sludge discharge. Then, these data are sent to the data processing module. Among them, the drying production capacity is the maximum production capacity of the specific drying workshop in the plant area, which represents the maximum sludge discharge volume that the plant area can handle, and the label is the number of sludge trucks transported.

[0012] S2: The data processing module calculates multiple groups of data including the total sludge reduction amount, the sludge self-decay amount, and the influent sludge increment based on the received data; then these data and the drying production capacity are sent to the neural network module.

[0013] S3: The neural network module uses the drying production capacity, the total sludge reduction amount, the sludge self-decay amount, and the influent sludge increment as inputs and the actual number of sludge trucks transported as outputs to construct and train a neural network model; among them, the total sludge reduction amount = the sludge self-decay amount + the sludge discharge amount corresponding to the actual number of sludge trucks transported - the influent sludge increment, and the influent sludge increment is the amount of sludge brought in by the influent. The sludge discharge amount corresponding to the actual number of sludge trucks transported is the amount of sludge discharged by manually controlling the biochemical treatment system.

[0014] S4: The data acquisition module collects data and sends it to the data processing module. After calculating the self-decay amount of the sludge and the influent sludge increment, the data processing module sends them to the neural network module. After receiving the data, the neural network module calculates the predicted number of sludge trucks not exceeding the drying production capacity using the neural network model, and then sends a signal to the control module.

[0015] S5: The control module controls the sludge discharge of the biochemical treatment system according to the predicted number of sludge trucks.

[0016] Preferably, in step S1, the data acquisition module first sends the collected data to the data optimization module. After the data optimization module removes the outliers from the data, it then sends the data to the data processing module. Removing the abnormal data can improve the accuracy of the data and is beneficial to improving the prediction accuracy of the neural network model.

[0017] Preferably, in step S1, the data optimization module uses the interquartile range criterion to detect and remove outliers from the collected data. The specific process is as follows: for the same parameter, first determine the first quartile and the third quartile in the data, and then calculate the difference between the two; then calculate the upper critical value and the lower critical value of the outlier according to the difference between the first quartile and the third quartile; finally, remove the data in the data that is greater than the upper critical value and less than the lower critical value.

[0018] Preferably, in step S3, the specific steps for the neural network module to train the neural network model are as follows:

[0019] S31: The neural network module confirms the data sample set and divides the data sample set into a training set, a validation set, and a test set. Among them, the data sample set uses the drying production capacity, the total sludge reduction amount, the self-decay amount of the sludge, and the influent sludge increment as inputs, and the actual transported sludge trucks as outputs.

[0020] S32: Determine the learning rate, the maximum number of iterations, the specified error, the activation function, the loss function, the optimization algorithm, and the neural network structure. The neural network structure includes an input layer, a hidden layer, and an output layer. The determined content includes the number of hidden layers and the number of neurons in each layer.

[0021] S33: Use the training set to train the model. During the training process, use the activation function in the output layer, truncate the output value exceeding the drying production capacity to the drying production capacity, and continuously update the network weights and biases through the backpropagation algorithm to optimize the network structure.

[0022] S34: Calculate the predicted value of the muck truck demand using the validation set, then calculate the error between the predicted value of the muck truck demand and the actual number of muck trucks transported according to the loss function, then calculate the gradient of the loss function with respect to each parameter using the backpropagation algorithm, and finally update the weights and biases of the model according to the calculated gradient and the optimization algorithm.

[0023] S35: Determine whether the error between the predicted value of the muck truck demand and the actual number of muck trucks transported is less than the specified error or the current iteration number reaches the maximum iteration number. If so, go to step S36; otherwise, return to step S32.

[0024] S36: Select the model with the minimum average error and evaluate the accuracy using the test set. If the accuracy reaches the set standard, determine this model as the desired neural network model; otherwise, re-enter step S32.

[0025] Preferably, in step S2, the calculation formula for the total sludge reduction is:

[0026]

[0027] In the formula, A represents the total sludge reduction; V represents the water volume for treatment before sludge discharge; X 2 represents the sludge concentration before sludge discharge; X 1 represents the sludge concentration after sludge discharge.

[0028] Preferably, the calculation formula for the influent sludge increment is:

[0029] B = Q * d * (S 0 * Y * BOD 5 / COD + f * (SS) 0 );

[0030] In the formula, B represents the influent sludge increment; Q represents the average daily influent water volume; S 0 represents the average daily influent chemical oxygen demand; d represents the number of days of influent; Y represents the sludge production coefficient; BOD5 / COD represents the 5-day biochemical oxygen demand / chemical oxygen demand, taking an empirical value; f represents the sludge conversion rate of suspended solids; (SS) 0 represents the average daily influent suspended solid concentration.

[0031] Preferably, in step S2, the calculation formula for the sludge self-decay amount is:

[0032] C = K d * V * X v * d;

[0033] In the formula, C represents the sludge self-decay amount; V represents the water volume for treatment before sludge discharge; X vIt represents the average concentration of mixed liquor volatile suspended solids in the biochemical tank; d represents the number of days elapsed.

[0034] Preferably, in the step S5, the control module first converts the predicted sludge truck volume into the daily sludge discharge volume, and then controls the biochemical treatment system to discharge sludge according to the daily sludge discharge volume, where the calculation formula for the daily sludge discharge volume is:

[0035] D = G * N * (1 - W);

[0036] In the formula, D represents the daily sludge discharge volume; G represents the predicted sludge truck volume; N represents the sludge transportation volume of a single sludge truck; W represents the moisture content of the sludge.

[0037] Advantages of the present invention:

[0038] 1. The sludge discharge control system of the present invention can replace manual prediction of the sludge discharge volume and control the biochemical treatment system to discharge sludge according to the predicted sludge discharge volume. It can not only reduce labor costs, but also improve the accuracy of sludge discharge volume control, thereby reducing chemical consumption and improving the sewage treatment effect. This system can be used in sewage treatment plants of various types and scales, and is applicable to the scientific scheduling of multiple sewage treatment production lines, which helps to achieve intelligent and refined operation management of sludge discharge.

[0039] 2. The neural network model is trained, verified and tested using historical data, and finally the neural network model is applied to the calculation of the predicted sludge truck volume, so that the control module can accurately control the sludge discharge volume of the biochemical tank, thereby controlling the sludge concentration in the biochemical tank within a reasonable range. While reducing labor costs and chemical consumption and improving the sewage treatment effect of the biochemical tank, it can also assist in the scheduling of sludge trucks. Description of the Drawings

[0040] Figure 1 It is a connection diagram of each module in a sludge discharge control system for sewage treatment;

[0041] Figure 2 It is a flowchart of a sludge discharge control method for sewage treatment;

[0042] Figure 3 It is a flowchart of constructing a neural network model using historical data;

[0043] Figure 4 It is a flowchart of the neural network model in actual application.

[0044] In the drawings: 1 - data acquisition module; 2 - data processing module; 3 - neural network module; 4 - control module; 5 - data optimization module. Detailed Embodiments

[0045] The technical solution of the present invention will be further specifically described below through specific embodiments in conjunction with the accompanying drawings:

[0046] Embodiment 1

[0047] This embodiment is the first embodiment of a sludge discharge control system for sewage treatment. As Figure 1 shown, it includes a data acquisition module 1, a data processing module 2, a neural network module 3, and a control module 4. The data acquisition module 1 is signal-connected to the data processing module 2, and both the data processing module 2 and the control module 4 are signal-connected to the neural network module 3.

[0048] Furthermore, it further includes a data optimization module 5, and both the data acquisition module 1 and the data processing module 2 are signal-connected to the data optimization module 5.

[0049] The working principle or working process of this embodiment: During implementation, the data acquisition module 1 first acquires relevant data and gives it to the data optimization module 5. After the data optimization module 5 eliminates the outliers in the data, it then sends the data to the neural network module 3. After receiving the data, the neural network module 3 constructs and trains a neural network model based on the received data. Then the data acquisition module 1 acquires data again and sends it to the neural network module 3. The neural network module 3 then calculates the predicted number of sludge trucks using the neural network model and then sends a signal to the control module 4. After receiving the signal, the control module 4 controls the opening and closing of the sludge pump and sludge valve of the biochemical treatment system according to the daily average sludge discharge volume corresponding to the predicted number of sludge trucks to control the sludge discharge of the biochemical treatment system.

[0050] 1. The beneficial effects of this embodiment: This system can replace manual prediction of sludge discharge volume and control the biochemical treatment system to discharge sludge according to the predicted sludge discharge volume. It can not only reduce labor costs, but also improve the accuracy of sludge discharge volume control, thereby reducing chemical consumption and improving the sewage treatment effect. This system can be used in sewage treatment plants of various types and scales, and is suitable for the scientific scheduling of multiple sewage treatment production lines, which helps to achieve intelligent and refined operation management of sludge discharge.

[0051] Embodiment 2

[0052] This embodiment is the first embodiment of a sludge discharge control method for sewage treatment, which is applied to the sludge discharge control system of Embodiment 1. In combination with Figures 1 to 4 shown, it includes the following steps:

[0053] S1: The data acquisition module 1 acquires multiple sets of data. Each set of data includes the drying capacity, the water volume before sludge discharge pretreatment, the water volume before sludge discharge, the 5-day biochemical oxygen demand of the influent, the chemical oxygen demand of the influent, the suspended solid concentration of the influent, the average concentration of mixed liquor volatile suspended solids in the influent, the sludge concentration before sludge discharge, the actual number of sludge transport vehicles, and the sludge concentration after sludge discharge. Then, these data are sent to the data processing module 2. Among them, the drying capacity is the maximum drying workshop capacity of the plant area, which represents the maximum sludge discharge amount that the plant area can handle, and the label is the number of sludge transport vehicles.

[0054] S2: The data processing module 2 calculates multiple sets of data including the total sludge reduction amount, the sludge self-decay amount, and the influent sludge increment based on the received data; then, these data and the drying capacity are sent to the neural network module 3.

[0055] S3: The neural network module 3 constructs and trains a neural network model with the drying capacity, the total sludge reduction amount, the sludge self-decay amount, and the influent sludge increment as inputs and the actual number of sludge transport vehicles as the output. Among them, the total sludge reduction amount = the sludge self-decay amount + the sludge discharge amount corresponding to the actual number of sludge transport vehicles - the influent sludge increment, and the influent sludge increment is the amount of sludge brought in by the influent. The sludge discharge amount corresponding to the actual number of sludge transport vehicles is the amount of sludge discharged by manually controlling the biochemical treatment system.

[0056] S4: The data acquisition module 1 acquires data and sends it to the data processing module 2. After the data processing module 2 calculates the sludge self-decay amount and the influent sludge increment, it sends them to the neural network module 3; after receiving the data, the neural network module 3 calculates the predicted number of sludge transport vehicles not exceeding the drying capacity using the neural network model, and then sends a signal to the control module 4.

[0057] S5: The control module 4 controls the sludge discharge pump in the biochemical treatment system to operate according to the predicted number of sludge transport vehicles to discharge sludge.

[0058] Furthermore, in step S1, the data acquisition module 1 first sends all the acquired data to the data optimization module 5. After the data optimization module 5 eliminates the outliers in the data, it then sends the data to the data processing module 2. Eliminating the abnormal data can improve the accuracy of the data and is conducive to improving the prediction accuracy of the neural network model.

[0059] Furthermore, in step S1, the data optimization module 5 uses the interquartile range criterion to detect and eliminate outliers from the acquired data. The specific process is as follows: for the same parameter, first determine the first quartile and the third quartile in the data, and then calculate the difference between the two; then calculate the upper critical value and the lower critical value of the outlier according to the difference between the first quartile and the third quartile; finally, eliminate the data in the data that is greater than the upper critical value and less than the lower critical value.

[0060] Further, in step S3, the specific steps for the neural network module 3 to train the neural network model are as follows:

[0061] S31: The neural network module 3 confirms the data sample set and divides the data sample set into a training set, a validation set, and a test set; wherein the data sample set uses the drying capacity, the total sludge reduction, the sludge self-decay amount, and the influent sludge increment as inputs, and the actual sludge transport trucks as outputs.

[0062] S32: Determine the learning rate, the maximum number of iterations, the specified error, the activation function, the loss function, the optimization algorithm, and the neural network structure. The neural network structure includes an input layer, a hidden layer, and an output layer. The determined content includes the number of hidden layers and the number of neurons in each layer.

[0063] S33: Use the training set to train the model; during the training process, use the activation function in the output layer, truncate the output value exceeding the drying capacity to the drying capacity, and continuously update the network weights and biases through the backpropagation algorithm to optimize the network structure.

[0064] S34: Use the validation set to calculate the predicted value of the sludge truck demand, then calculate the error between the predicted value of the sludge truck demand and the actual number of sludge transport trucks according to the loss function, then use the backpropagation algorithm to calculate the gradient of the loss function with respect to each parameter, and finally update the weights and biases of the model according to the calculated gradient and the optimization algorithm.

[0065] S35: Determine whether the error between the predicted value of the sludge truck demand and the actual number of sludge transport trucks is less than the specified error or the current number of iterations reaches the maximum number of iterations. If so, enter step S36; otherwise, return to step S32.

[0066] S36: Select the model with the smallest average error and evaluate the accuracy using the test set. If the accuracy reaches the set standard, determine this model as the desired neural network model; otherwise, re-enter step S32.

[0067] The beneficial effects of this embodiment: Use historical data to train, validate, and test the neural network model, and finally apply the neural network model to the calculation of predicting the number of sludge trucks, so that the control module can accurately control the sludge discharge amount of the biochemical pool, thereby controlling the sludge concentration of the biochemical pool within a reasonable range. While reducing labor costs and chemical consumption and improving the sewage treatment effect of the biochemical pool, it can also assist in the scheduling of sludge trucks.

[0068] Other features, working principles, and beneficial effects of this embodiment are the same as those of Embodiment 1.

[0069] Embodiment 3

[0070] This embodiment is the second embodiment of a sludge discharge control method for sewage treatment. On the basis of Embodiment 2, this embodiment further elaborates on steps S2 and S5.

[0071] In step S2, the calculation formula for the total sludge reduction is:

[0072]

[0073] In the formula, A represents the total sludge reduction; V represents the water volume for treatment before sludge discharge; X 2 represents the sludge concentration before sludge discharge; X 1 represents the sludge concentration after sludge discharge.

[0074] Furthermore, the calculation formula for the influent sludge increment is:

[0075] B = Q * d * (S 0 * Y * BOD 5 / COD + f * (SS) 0 );

[0076] In the formula, B represents the influent sludge increment; Q represents the average daily influent water volume; S 0 represents the average daily influent chemical oxygen demand; d represents the number of days of influent; Y represents the sludge production coefficient; BOD5 / COD represents the 5-day biochemical oxygen demand / chemical oxygen demand, taking an empirical value; f represents the sludge conversion rate of suspended solids; (SS) 0 represents the average daily influent suspended solid concentration.

[0077] Furthermore, in step S2, the calculation formula for the sludge self-decay amount is:

[0078] C = K d * V * X v * d;

[0079] In the formula, C represents the sludge self-decay amount; V represents the water volume for treatment before sludge discharge; X v represents the average concentration of mixed liquor volatile suspended solids in the biochemical pool; d represents the number of days elapsed.

[0080] Furthermore, in step S5, the control module 4 first converts the predicted sludge truck volume into the daily sludge discharge volume, and then controls the biochemical treatment system to discharge sludge according to the daily sludge discharge volume. The calculation formula for the daily sludge discharge volume is:

[0081] D = G * N * (1 - W);

[0082] In the formula, D represents the daily sludge discharge volume; G represents the predicted sludge truck volume; N represents the sludge transportation volume of a single sludge truck; W represents the sludge moisture content

[0083] The other working principles and beneficial effects of this embodiment are the same as those of Embodiment 2.

[0084] In the specific content of the above specific embodiments, each technical feature can be combined arbitrarily without contradiction. For the sake of concise description, not all possible combinations of the above technical features are described. However, as long as there is no contradiction in the combination of these technical features, it should be considered as the scope recorded in this specification.

[0085] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, rather than limitations on the implementation manners of the present invention. For those of ordinary skill in the art, other different forms of changes or modifications can be made on the basis of the above description. It is not necessary and impossible to enumerate all implementation manners here. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the claims of the present invention.

Claims

1. A sludge discharge control system for sewage treatment, characterized in that: The invention comprises a data acquisition module (1), a data processing module (2), a neural network module (3) and a control module (4), wherein the data acquisition module (1) is connected to the data processing module (2) by signals, and the data processing module (2) and the control module (4) are both connected to the neural network module (3) by signals.

2. The sludge discharge control system for sewage treatment according to claim 1, characterized in that: It also comprises a data optimization module (5), and the data acquisition module (1) and the data processing module (2) are both connected to the data optimization module (5) by signals.

3. A sludge discharge control method for sewage treatment, applied to the sludge discharge control system for sewage treatment according to claim 2, characterized in that: The following steps are involved: S1: The data acquisition module (1) collects multiple sets of data, each set of data includes drying capacity, pre-sludge treatment water volume, pre-sludge water volume, influent 5-day biochemical oxygen demand, influent chemical oxygen demand, influent suspended solids concentration, influent mixed liquor volatile suspended solids average concentration, pre-sludge sludge concentration, actual sludge transport truck volume and post-sludge sludge concentration, and then sends these data to the data processing module (2); S2: The data processing module (2) calculates multiple groups of data including the total sludge reduction, the sludge self-decay and the influent sludge increment according to the received data; and then sends these data and the drying capacity to the neural network module (3); S3: The neural network module (3) takes the drying capacity, the total sludge reduction, the sludge self-decay and the influent sludge increment as inputs, and takes the actual transport sludge truck volume as output, to construct and train a neural network model; S4: the data acquisition module (1) acquires data and sends it to the data processing module (2); the data processing module (2) calculates the sludge self-decay amount and the influent sludge increment and sends them to the neural network module (3); after receiving the data, the neural network module (3) uses the neural network model to calculate the predicted mud truck quantity that does not exceed the drying capacity, and then sends a signal to the control module (4); S5: The control module (4) controls the biochemical treatment system to discharge sludge according to the predicted sludge truck volume.

4. A sludge discharge control method for sewage treatment according to claim 3, characterized in that: In step S1, the data acquisition module (1) first sends the collected data to the data optimization module (5), and the data optimization module (5) removes abnormal values ​​in the data and then sends the data to the data processing module (2).

5. A sludge discharge control method for sewage treatment according to claim 4, characterized in that: In step S1, the data optimization module (5) uses the interquartile range criterion to detect and eliminate outliers in the collected data. The specific process is as follows: for the same parameter, the first quartile and the third quartile in the data are first determined, and then the difference between the two is calculated; then the upper critical value and the lower critical value of the outlier are calculated according to the difference between the first quartile and the third quartile; finally, the data greater than the upper critical value and less than the lower critical value in the data are eliminated.

6. A sludge discharge control method for sewage treatment according to claim 3, characterized in that: In step S3, the specific steps of the neural network module (3) training the neural network model are: S31: the neural network module (3) confirms a data sample set, and divides the data sample set into a training set, a validation set and a test set; wherein the data sample set takes the drying capacity, the total sludge reduction, the sludge self-decay amount and the influent sludge increment as input, and takes the actual sludge transport truck as output; S32: Determine a learning rate, a maximum number of iterations, a specified error, an activation function, a loss function, an optimization algorithm, and a neural network structure, wherein the neural network structure includes an input layer, a hidden layer, and an output layer, and the content determined includes the number of hidden layers and the number of neurons in each layer; S33: training the model using the training set; during the training process, using the activation function in the output layer to truncate the output value exceeding the drying capacity to the drying capacity, and continuously updating the network weights and biases through the back propagation algorithm to optimize the network structure; S34: using the validation set to calculate the predicted value of the mud truck demand, and then calculating the error between the predicted value of the mud truck demand and the actual transport mud truck quantity according to the loss function, and then using the back propagation algorithm to calculate the gradient of the loss function with respect to each parameter, and finally updating the weight and bias of the model according to the calculated gradient and the optimization algorithm; S35: Determine whether the error between the predicted value of the demand for mud trucks and the actual transported mud truck quantity is less than the specified error or the current iteration number reaches the maximum iteration number, if yes, proceed to step S36, otherwise return to step S32; S36: Select the model with the smallest average error and use the test set to evaluate the accuracy. If the accuracy reaches the set standard, the model is determined to be the desired neural network model. Otherwise, re-enter the step S32.

7. A sludge discharge control method for sewage treatment according to claim 3, characterized in that: In step S2, the calculation formula for the total sludge reduction is: Where A represents the total sludge reduction; V represents the amount of treated water before sludge discharge; X2 represents the sludge concentration before sludge discharge; and X1 represents the sludge concentration after sludge discharge.

8. A sludge discharge control method for sewage treatment according to claim 3, characterized in that: In step S2, the calculation formula for the influent sludge increment is: B=Q*d*(S0*Y*BOD5 / COD+f*(SS)0); In the formula, B represents the influent sludge increment; Q represents the average daily influent volume; S0 represents the average daily influent chemical oxygen demand; d represents the number of influent days; Y represents the sludge production coefficient; BOD5 / COD represents the 5-day biochemical oxygen demand / chemical oxygen demand, which is an empirical value; f represents the sludge conversion rate of suspended solids; (SS)0 represents the average daily influent suspended solids concentration.

9. A sludge discharge control method for sewage treatment according to claim 3, characterized in that: In step S2, the calculation formula of the sludge self-decay amount is: C=K d *V*X v *d; Where, C represents the amount of sludge self-destruction; K d represents the attenuation coefficient; V represents the amount of water treated before sludge discharge; X v represents the average concentration of volatile suspended solids in the mixed liquor in the biochemical pool; d represents the number of days experienced.

10. The sludge discharge control method for sewage treatment according to claim 3, characterized in that: In step S5, the control module (4) first converts the predicted mud truck volume into a daily mud discharge volume, and then controls the biochemical treatment system to discharge the sludge according to the daily mud discharge volume, wherein the calculation formula of the daily mud discharge volume is: D = G*N*(1-W); Where D represents the daily sludge discharge; G represents the predicted sludge truck volume; N represents the transported sludge volume of a single sludge truck; and W represents the moisture content of the sludge.

Images