Rural sewage treatment process and system

Abstract

The invention relates to the technical field of sewage treatment, in particular to a rural sewage treatment process and system.The rural sewage treatment process comprises the steps that S1, sewage pretreatment is conducted through a raw water pool to remove large floating objects and suspended solids in sewage, and the pretreated sewage is conveyed to a biochemical reaction pool; s2, monitoring water quality parameters in real time through water quality detection equipment in the biochemical reaction tank; s3, inputting the water quality parameters into a pre-trained neural network model to output operation parameters of the biochemical reaction tank; s4, performing biochemical treatment and advanced treatment on the sewage in the biochemical reaction tank according to the operation parameters, the biochemical treatment including removal of organic matters, nitrogen and phosphorus removal and reduction of the number of pathogenic microorganisms, and the advanced treatment including further removal of suspended solids, removal of refractory organic matters and disinfection and sterilization. And the operation parameters are flexibly obtained through the neural network model, so that the sewage treatment flexibility is improved, and the resource waste is reduced.

Description

Technical Field

[0001] The present invention relates to the technical field of sewage treatment, and in particular to a rural sewage treatment process and system. Background Art

[0002] Rural sewage treatment is an important part of rural living environment improvement. Effective treatment of rural sewage can reduce water pollution, protect water resources, and improve the rural ecological environment. After treatment, sewage can be used for agricultural irrigation, landscape water, etc., to improve the utilization rate of water resources, while reducing disease transmission and improving the quality of life of farmers.

[0003] The traditional rural sewage treatment process usually includes a coarse screen, a fine screen, a sand settling tank, a regulating tank, a flocculation reaction tank, anoxic tank, anaerobic tank, aerobic tank, a secondary sedimentation tank, a deep bed filter, etc., and each process is executed in series. For example, a rural sewage treatment system with a publication number of CN114940563A includes: a water distributor, a biofilm reaction unit and an aerator. The water distributor is connected to the water inlet channel and is located at the bottom of the treatment tank. The treatment tank is formed with a sludge zone, an anaerobic zone, anoxic zone and aerobic zone from bottom to top; the biofilm reaction unit is located between the aerobic zone and the anoxic zone; the aerator is arranged between the anoxic zone and the aerobic zone. In the above-mentioned rural sewage treatment system, the water distributor is located at the bottom of the treatment tank so that the sewage flows from bottom to top, and passes through the sludge zone, anaerobic zone, anoxic zone and aerobic zone during the flow process to achieve sewage treatment. It is an environmentally friendly and low-cost rural sewage treatment system.

[0004] Although the prior art has achieved different treatments for sewage by zoning the treatment pool, the operating parameters of the biochemical reaction cannot be flexibly adjusted according to the sewage quality in real time during sewage treatment, resulting in the inability to flexibly treat different types of rural sewage differently, causing a waste of treatment resources. Summary of the invention

[0005] In order to solve the problem that traditional sewage treatment cannot flexibly adjust the operating parameters of the biochemical reaction according to the sewage quality in real time, resulting in the inability to flexibly treat different types of rural sewage in different ways, causing waste of treatment resources, the present invention provides a rural sewage treatment process and system.

[0006] In the first aspect, the present invention provides a rural sewage treatment process, which adopts the following technical scheme: A rural sewage treatment process comprises the following steps: S1, pre-treating the sewage through the raw water tank to remove larger floating objects and suspended solids in the sewage, and transporting the pre-treated sewage to the biochemical reaction tank; S2, real-time monitoring of water quality parameters through water quality testing equipment in the biochemical reaction pool; S3, inputting the water quality parameters into the pre-trained neural network model to output the operating parameters of the biochemical reaction tank; S4, according to the operating parameters, the wastewater is biochemically treated and deeply treated in the biochemical reaction tank. The biochemical treatment includes the removal of organic matter, nitrogen and phosphorus removal and the reduction of the number of pathogenic microorganisms. The deep treatment includes further removal of suspended matter, removal of difficult-to-degrade organic matter and disinfection.

[0007] By adopting the above technical solution, corresponding operation parameters are obtained according to real-time water quality parameters, so that when sewage treatment is carried out, it can adapt to different sewage types, has a strong adaptability to rural sewage, and saves treatment resources.

[0008] Optionally, before training the neural network model, multiple water quality parameters and their corresponding optimal operating parameters are obtained through experiments, specifically including: Obtaining a variety of different types of sewage as test subjects, and obtaining a variety of water quality parameters for each test subject; In each test body, a plurality of operating parameters are set according to a gradient and the test body is treated to obtain a plurality of treated water quality parameters; The treated water quality parameters are subtracted from the standard water quality parameters, and the treated water quality parameters with the smallest absolute value of the difference are taken as the best treatment result, and the corresponding operating parameters are obtained as the best operating parameters.

[0009] By adopting the above technical solution, the data obtained by processing the test body is used as the basis for training the model.

[0010] Optionally, the process of training the neural network model includes: After associating the water quality parameters obtained from the experiment with the optimal operating parameters, they were randomly divided into training and test sets; The neural network model is trained through the training set to obtain the initial prediction model, and the output of the model is the corresponding optimal operating parameters; The initial prediction model is tested through the test set, and the predicted optimal operating parameters are output; Determine a corresponding loss value based on the predicted optimal operating parameters and mean square error output by the model, and judge whether the loss value is less than a threshold; If it is less than the threshold, stop training and use the initial prediction model as the corresponding prediction model; If it is not less than the threshold, the parameters corresponding to the initial prediction model are corrected based on the loss value, and the prediction model is updated.

[0011] Optionally, the water quality parameters include chromaticity data, turbidity data, SS data, PH data, COD data, BOD data, TOC data, ammonia nitrogen data, total nitrogen data, and total phosphorus data; The operating parameters include dosage of reagent, aeration volume, residual sludge discharge volume and processing time.

[0012] Optionally, in step S4, each biochemical treatment and deep treatment of the sewage includes: S4-1, input the current water quality parameters into the neural network model and output the operation parameters; S4-2, biochemical treatment and deep treatment of current sewage according to operating parameters; S4-3, analyzing and comparing the treated water quality parameters with the standard water quality parameters and obtaining the relative error; When the relative error is greater than the preset threshold, return to S4-1; When the relative error is less than the threshold, the process ends.

[0013] By adopting the above technical solution, since there may be deviations between the test and the actual situation, the sewage treatment effect can be improved through actual monitoring.

[0014] Optionally, analyzing and comparing the actual treated water quality parameters with the standard water quality parameters and obtaining the relative errors include: Obtain the water quality parameter removal rate in the test and the actual water quality parameter removal rate; Relative error = (|actual water quality parameter removal rate - water quality parameter removal rate in the test| / water quality parameter removal rate in the test) * 100%.

[0015] In a second aspect, a rural sewage treatment system is provided, comprising: Raw water tank and biochemical reaction tank. The raw water tank pre-treats the sewage to remove larger floating objects and suspended solids in the sewage, and then transports the pre-treated sewage to the biochemical reaction tank. Multiple water quality testing devices are installed in the biochemical reaction pool to monitor water quality parameters in real time; a processor, inputting water quality parameters into a pre-trained neural network model to output biochemical reaction tank operating parameters; The controller controls the biochemical treatment and deep treatment of the sewage in the biochemical reaction tank according to the operating parameters output by the processor. The biochemical treatment includes removing organic matter, removing nitrogen and phosphorus, and reducing the number of pathogenic microorganisms. The deep treatment includes further removing suspended matter, removing difficult-to-degrade organic matter, and disinfecting and sterilizing.

[0016] By adopting the above technical solution, the treatment module is adjusted to match the fixed treatment module for sewage treatment, which greatly saves the floor space. The operation parameters are flexibly obtained through the neural network model, which improves the flexibility of sewage treatment and reduces the waste of treatment resources.

[0017] Optionally, the water quality detection equipment includes: an online colorimetry sensor, an online turbidity sensor, a suspended matter concentration sensor, an online pH sensor, an online COD sensor, an online BOD sensor, an online TOC sensor, an online ammonia nitrogen sensor, an online total nitrogen sensor, and an online total phosphorus sensor.

[0018] Optionally, before training the neural network model, multiple water quality parameters and their corresponding optimal operating parameters are obtained through experiments, specifically including: Obtaining a variety of different types of sewage as test subjects, and obtaining a variety of water quality parameters for each test subject; In each test body, a plurality of operating parameters are set according to a gradient and the test body is treated to obtain a plurality of treated water quality parameters; The treated water quality parameters are subtracted from the standard water quality parameters, and the treated water quality parameters with the smallest absolute value of the difference are taken as the best treatment result, and the corresponding operating parameters are obtained as the best operating parameters.

[0019] Optionally, the process of training the neural network model includes: After associating the water quality parameters obtained from the experiment with the optimal operating parameters, they were randomly divided into training and test sets; The neural network model is trained through the training set to obtain the initial prediction model, and the output of the model is the corresponding optimal operating parameters; The initial prediction model is tested through the test set, and the predicted optimal operating parameters are output; Determine a corresponding loss value based on the predicted optimal operating parameters and mean square error output by the model, and judge whether the loss value is less than a threshold; If it is less than the threshold, stop training and use the initial prediction model as the corresponding prediction model; If it is not less than the threshold, the parameters corresponding to the initial prediction model are corrected based on the loss value, and the prediction model is updated.

[0020] In summary, the present invention has the following beneficial technical effects: When the types of sewage to be treated are different, the operation parameters can be flexibly obtained through the neural network model, which increases the flexibility of sewage treatment. The sewage treatment operation is carried out by combining sewage treatment equipment with raw water tanks and biochemical reaction tanks, which reduces the occupied space and reduces the waste of resources. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 It is a flow chart of a rural sewage treatment process according to an embodiment of the present invention.

[0022] Figure 2 It is a structural schematic diagram of a rural sewage treatment system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0023] The following is combined with Figure 1-2 The present invention is described in further detail.

[0024] Example 1 Reference Figure 1 , a rural sewage treatment process of this embodiment comprises the following steps: S1, pre-treat the sewage through the raw water pool to remove larger floating and suspended matter in the sewage, and transport the pre-treated sewage to the biochemical reaction pool; the raw water pool is equipped with a grid to remove larger suspended matter, floating matter and debris in the sewage, such as branches, leaves, plastic bags, fibers, etc. Sand settling is also carried out in the raw water pool to remove sand, coal slag and other inorganic particles with larger specific gravity in the sewage, so as to reduce the wear and siltation of the subsequent treatment units.

[0025] S2, real-time monitoring of water quality parameters through water quality detection equipment in the biochemical reaction pool; the water quality detection equipment includes a variety of sensors: online colorimetry sensor, online turbidity sensor, suspended solids concentration sensor, online pH sensor, online COD sensor, online BOD sensor, online TOC sensor, online ammonia nitrogen sensor, online total nitrogen sensor, online total phosphorus sensor.

[0026] S3, inputting the water quality parameters into the pre-trained neural network model to output the operating parameters of the biochemical reaction tank; Water quality parameters include chromaticity data, turbidity data, SS data, PH data, COD data, BOD data, TOC data, ammonia nitrogen data, total nitrogen data, and total phosphorus data; Operation parameters include dosage of reagents, aeration volume, residual sludge discharge, and treatment time; The dosage of reagents affects all water quality parameters; the reagents include acid-base regulators, coagulants, oxidants, phosphorus removal agents and disinfectants, etc. The required aeration volume has an impact on BOD data, COD data, and ammonia nitrogen data; the aeration volume refers to the total amount of gas introduced into the sewage treatment system per unit time, which is mainly used to provide oxygen, stir and mix, and remove volatile substances. The amount of aeration directly affects the concentration of dissolved oxygen (DO) in the water. Sufficient DO can ensure the normal metabolism and growth of microorganisms, improve the removal efficiency of organic matter and the biochemical treatment effect. The rising bubbles and airflow movement generated during the aeration process can fully mix the sewage and activated sludge, prevent sludge precipitation, maintain the uniform distribution of materials in the treatment system, and promote the contact and reaction between pollutants and microorganisms. Aeration can blow volatile organic compounds (VOCs) and some harmful gases (such as hydrogen sulfide, methane, etc.) in sewage out of the water, reduce the odor and toxicity of sewage, and improve the working environment of the treatment system.

[0027] The amount of residual sludge discharged affects the SS data, BOD data, COD data, and ammonia nitrogen data; the amount of residual sludge discharged refers to the amount of excess sludge discharged from the biochemical treatment system. By discharging a certain amount of sludge, the sludge residence time (sludge age) in the system is kept within an appropriate range to maintain the activity and metabolic capacity of microorganisms and prevent sludge aging and over-breeding. Removing some invalid or aged sludge provides space for the growth of new microorganisms and maintains the dynamic balance of the microbial population. Excessive filamentous bacteria and other microorganisms that are prone to sludge swelling are discharged in a timely manner to maintain the stable operation of the treatment system.

[0028] Treatment time has an impact on SS data, BOD data, COD data, ammonia nitrogen data, total nitrogen data, and total phosphorus data. For example, the hydraulic retention time of the biochemical treatment pool, the reaction time of the chemical reaction pool, the sedimentation time, the filtration time, the disinfection contact time, etc.

[0029] Before training the neural network model, multiple water quality parameters and their corresponding optimal operating parameters were obtained through experiments, including: Obtaining a variety of different types of sewage as test subjects, and obtaining a variety of water quality parameters for each test subject; In each test body, a plurality of operating parameters are set according to a gradient and the test body is treated to obtain a plurality of treated water quality parameters; The treated water quality parameters are subtracted from the standard water quality parameters, and the treated water quality parameters with the smallest absolute value of the difference are taken as the best treatment result, and the corresponding operating parameters are obtained as the best operating parameters.

[0030] The process of neural network model training includes: After associating the water quality parameters obtained from the experiment with the optimal operating parameters, they were randomly divided into training and test sets; The neural network model is trained through the training set to obtain the initial prediction model, and the output of the model is the corresponding optimal operating parameters; The initial prediction model is tested through the test set, and the predicted optimal operating parameters are output; Determine a corresponding loss value based on the predicted optimal operating parameters and mean square error output by the model, and judge whether the loss value is less than a threshold; If it is less than the threshold, stop training and use the initial prediction model as the corresponding prediction model; If it is not less than the threshold, the parameters corresponding to the initial prediction model are corrected based on the loss value, and the prediction model is updated.

[0031] S4, according to the operating parameters, the wastewater is biochemically treated and deeply treated in the biochemical reaction tank. The biochemical treatment includes the removal of organic matter, nitrogen and phosphorus removal and the reduction of the number of pathogenic microorganisms. The deep treatment includes further removal of suspended matter, removal of difficult-to-degrade organic matter and disinfection.

[0032] The processes corresponding to biochemical treatment and deep treatment include: The activated sludge method, biofilm method and anaerobic biochemical treatment method are used to remove organic matter, and the A² / O process is used for nitrogen and phosphorus removal to reduce the natural inhibition of the number of pathogenic microorganisms in the biochemical treatment process. It also includes: acid-base neutralization treatment with chemicals, chemical filtration, etc. To further remove suspended solids, membrane filtration technology is used, which utilizes the pore size selectivity of the membrane to separate and remove tiny particles, bacteria, viruses, soluble organic matter, etc. in the sewage, such as microfiltration, ultrafiltration, nanofiltration, reverse osmosis, etc.

[0033] Activated carbon adsorption and advanced oxidation technology are used to remove difficult-to-degrade organic matter. The huge specific surface area and rich pore structure of activated carbon are used to adsorb pollutants such as soluble organic matter, color, and odor in sewage. By producing free radicals with strong oxidizing properties (such as ·OH), the difficult-to-degrade organic matter in sewage is oxidized and decomposed into small molecular substances, and even mineralized into carbon dioxide and water, such as ozone oxidation, Fenton oxidation, photocatalytic oxidation, wet oxidation, etc.

[0034] Disinfection and sterilization are carried out using chlorine disinfection, ultraviolet disinfection and ozone disinfection.

[0035] Each biological and advanced treatment of sewage includes: S4-1, input the current water quality parameters into the neural network model and output the operation parameters; S4-2, biochemical treatment and deep treatment of current sewage according to operating parameters; S4-3, analyzing and comparing the treated water quality parameters with the standard water quality parameters and obtaining the relative error; When the relative error is greater than the preset threshold, return to S4-1; When the relative error is less than the threshold, the process ends. In this embodiment, the threshold is set to 5%.

[0036] Analyze and compare the actual treated water quality parameters with the standard water quality parameters and obtain the relative errors including: Obtain the water quality parameter removal rate in the test and the actual water quality parameter removal rate; Relative error = (|actual water quality parameter removal rate - water quality parameter removal rate in the test| / water quality parameter removal rate in the test) * 100%.

[0037] For example, theoretically, the influent COD concentration is 400 mg / L, the ammonia nitrogen concentration is 35 mg / L, and the total phosphorus concentration is 5 mg / L. The designed effluent COD concentration is ≤60 mg / L, the ammonia nitrogen concentration is ≤8 mg / L, and the total phosphorus concentration is ≤1 mg / L. The removal rates of COD, ammonia nitrogen and total phosphorus are 85%, 77% and 80% respectively.

[0038] In actual operation, the average inlet COD concentration was 380 mg / L, the average ammonia nitrogen concentration was 30 mg / L, and the average total phosphorus concentration was 4 mg / L. The average effluent COD concentration was 70 mg / L, the average ammonia nitrogen concentration was 12 mg / L, and the average total phosphorus concentration was 1.5 mg / L. The actual COD, ammonia nitrogen and total phosphorus removal rates were 81.6%, 60% and 62.5% respectively.

[0039] The comparative analysis results at this time are: COD removal rate: The theoretical value is 85%, the actual value is 81.6%, the relative error is 4%, the difference is small, and it is less than the threshold of 5%, so there is no need to adjust the operating parameters.

[0040] Ammonia nitrogen removal rate: The theoretical value is 77%, the actual value is 60%, and the relative error is 22.1%. The difference is large and far greater than the threshold. Analysis shows that nitrification may be insufficient. In this case, the treatment time can be extended or the dosage can be increased to promote a full reaction.

[0041] Total phosphorus removal rate: The theoretical value is 80%, the actual value is 62.5%, and the relative error is 21.9%. The difference is large and far greater than the threshold. Analysis shows that the dosage of chemical phosphorus removal agent is insufficient. At this time, increase the dosage and treat again.

[0042] Example 2 Reference Figure 2 The difference between this embodiment and embodiment 1 is that this embodiment provides a rural sewage treatment system, including: Raw water tank and biochemical reaction tank. The raw water tank pre-treats the sewage to remove larger floating objects and suspended solids in the sewage, and then transports the pre-treated sewage to the biochemical reaction tank. Multiple water quality testing devices are installed in the biochemical reaction pool to monitor water quality parameters in real time; a processor, inputting water quality parameters into a pre-trained neural network model to output biochemical reaction tank operating parameters; The controller controls the biochemical treatment and deep treatment of the sewage in the biochemical reaction tank according to the operating parameters output by the processor. The biochemical treatment includes removing organic matter, removing nitrogen and phosphorus, and reducing the number of pathogenic microorganisms. The deep treatment includes further removing suspended matter, removing difficult-to-degrade organic matter, and disinfecting and sterilizing.

[0043] Before training the neural network model, multiple water quality parameters and their corresponding optimal operating parameters were obtained through experiments, including: Obtaining a variety of different types of sewage as test subjects, and obtaining a variety of water quality parameters for each test subject; In each test body, a plurality of operating parameters are set according to a gradient and the test body is treated to obtain a plurality of treated water quality parameters; The treated water quality parameters are subtracted from the standard water quality parameters, and the treated water quality parameters with the smallest absolute value of the difference are taken as the best treatment result, and the corresponding operating parameters are obtained as the best operating parameters.

[0044] The process of neural network model training includes: After associating the water quality parameters obtained from the experiment with the optimal operating parameters, they were randomly divided into training and test sets; The neural network model is trained through the training set to obtain the initial prediction model, and the output of the model is the corresponding optimal operating parameters; The initial prediction model is tested through the test set, and the predicted optimal operating parameters are output; Determine a corresponding loss value based on the predicted optimal operating parameters and mean square error output by the model, and judge whether the loss value is less than a threshold; If it is less than the threshold, stop training and use the initial prediction model as the corresponding prediction model; If it is not less than the threshold, the parameters corresponding to the initial prediction model are corrected based on the loss value, and the prediction model is updated.

[0045] All pollution control links are integrated into two pools, a raw water pool and a biochemical reaction pool. More than ten traditional sewage pools are simplified into two functional reaction pools. The neural network model is used to obtain different sewage parameters and achieve standard treatment of different sewage, which greatly reduces the sewage treatment cost and improves the efficiency of resource management.

[0046] In this specification, each embodiment is described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the embodiments can be referred to each other. For the process disclosed in the embodiment, since it corresponds to the system disclosed in the embodiment, the description is relatively simple, and the relevant parts can be referred to the process part description.

[0047] Professionals may further appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, computer software, or a combination of the two. In order to clearly illustrate the interchangeability of hardware and software, the composition and steps of each example have been generally described in the above description according to function. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professionals and technicians may use different processes to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of the present invention.

[0048] The above disclosure is only a preferred embodiment of the present invention, but the present invention is not limited thereto. Any non-creative changes that can be thought of by a person skilled in the art, as well as several improvements and modifications made without departing from the principle of the present invention, should fall within the protection scope of the present invention.

Claims

1. A rural sewage treatment process, characterized in that: The following steps are involved: S1, pre-treating the sewage through the raw water tank to remove larger floating objects and suspended solids in the sewage, and transporting the pre-treated sewage to the biochemical reaction tank; S2, real-time monitoring of water quality parameters through water quality testing equipment in the biochemical reaction pool; S3, inputting the water quality parameters into the pre-trained neural network model to output the operating parameters of the biochemical reaction tank; S4, according to the operating parameters, the wastewater is biochemically treated and deeply treated in the biochemical reaction tank. The biochemical treatment includes the removal of organic matter, acid-base neutralization, nitrogen and phosphorus removal and the reduction of the number of pathogenic microorganisms. The deep treatment includes further removal of suspended matter, removal of difficult-to-degrade organic matter and disinfection.

2. The rural sewage treatment process according to claim 1, characterized in that: Before training the neural network model, experiments were conducted to obtain various water quality parameters and their corresponding optimal operating parameters, including: Obtaining a variety of different types of sewage as test subjects, and obtaining a variety of water quality parameters for each test subject; In each test body, a plurality of operating parameters are set according to a gradient and the test body is treated to obtain a plurality of treated water quality parameters; The treated water quality parameters are subtracted from the standard water quality parameters, and the treated water quality parameters with the smallest absolute value of the difference are taken as the best treatment result, and the corresponding operating parameters are obtained as the best operating parameters.

3. The rural sewage treatment process according to claim 2, characterized in that: The process of neural network model training includes: After associating the water quality parameters obtained from the experiment with the optimal operating parameters, they were randomly divided into training and test sets; The neural network model is trained through the training set to obtain the initial prediction model, and the output of the model is the corresponding optimal operating parameters; The initial prediction model is tested through the test set, and the predicted optimal operating parameters are output; Determine a corresponding loss value based on the predicted optimal operating parameters and mean square error of the model output, and judge whether the loss value is less than a threshold; If it is less than the threshold, stop training and use the initial prediction model as the corresponding prediction model; If it is not less than the threshold, the parameters corresponding to the initial prediction model are corrected based on the loss value, and the prediction model is updated.

4. The rural sewage treatment process according to claim 3, characterized in that: The water quality parameters include chromaticity data, turbidity data, SS data, PH data, COD data, BOD data, TOC data, ammonia nitrogen data, total nitrogen data, and total phosphorus data; The operating parameters include dosage of reagent, aeration volume, residual sludge discharge volume and processing time.

5. The rural sewage treatment process according to claim 2, characterized in that: In step S4, each biochemical treatment and deep treatment of the sewage includes: S4-1, input the current water quality parameters into the neural network model and output the operation parameters; S4-2, biochemical treatment and deep treatment of current sewage according to operating parameters; S4-3, analyzing and comparing the treated water quality parameters with the standard water quality parameters and obtaining the relative error; When the relative error is greater than the preset threshold, return to S4-1; When the relative error is less than the threshold, the process ends.

6. The rural sewage treatment process according to claim 5, characterized in that: Analyze and compare the actual treated water quality parameters with the standard water quality parameters and obtain the relative errors including: Obtain the water quality parameter removal rate in the test and the actual water quality parameter removal rate; Relative error = (|actual water quality parameter removal rate - water quality parameter removal rate in the test| / water quality parameter removal rate in the test) * 100%.

7. A rural sewage treatment system, characterized in that: include: Raw water tank and biochemical reaction tank. The raw water tank pre-treats the sewage to remove larger floating objects and suspended solids in the sewage, and then transports the pre-treated sewage to the biochemical reaction tank. Multiple water quality testing devices are installed in the biochemical reaction pool to monitor water quality parameters in real time; a processor, inputting water quality parameters into a pre-trained neural network model to output biochemical reaction tank operating parameters; The controller controls the biochemical treatment and deep treatment of the sewage in the biochemical reaction tank according to the operating parameters output by the processor. The biochemical treatment includes removing organic matter, removing nitrogen and phosphorus, and reducing the number of pathogenic microorganisms. The deep treatment includes further removing suspended matter, removing difficult-to-degrade organic matter, and disinfecting and sterilizing.

8. The rural sewage treatment system according to claim 7, characterized in that: Water quality testing equipment includes: online colorimetry sensor, online turbidity sensor, suspended matter concentration sensor, online pH sensor, online COD sensor, online BOD sensor, online TOC sensor, online ammonia nitrogen sensor, online total nitrogen sensor, and online total phosphorus sensor.

9. The rural sewage treatment system according to claim 7, characterized in that: Before training the neural network model, experiments were conducted to obtain various water quality parameters and their corresponding optimal operating parameters, including: Obtaining a variety of different types of sewage as test subjects, and obtaining a variety of water quality parameters for each test subject; In each test body, a plurality of operating parameters are set according to a gradient and the test body is treated to obtain a plurality of treated water quality parameters; The treated water quality parameters are subtracted from the standard water quality parameters, and the treated water quality parameters with the smallest absolute value of the difference are taken as the best treatment result, and the corresponding operating parameters are obtained as the best operating parameters.

10. The rural sewage treatment system according to claim 9, characterized in that: The process of neural network model training includes: After associating the water quality parameters obtained from the experiment with the optimal operating parameters, they were randomly divided into training and test sets; The neural network model is trained through the training set to obtain the initial prediction model, and the output of the model is the corresponding optimal operating parameters; The initial prediction model is tested through the test set, and the predicted optimal operating parameters are output; Determine a corresponding loss value based on the predicted optimal operating parameters and mean square error of the model output, and judge whether the loss value is less than a threshold; If it is less than the threshold, stop training and use the initial prediction model as the corresponding prediction model; If it is not less than the threshold, the parameters corresponding to the initial prediction model are corrected based on the loss value, and the prediction model is updated.

Images