Water quality up-to-standard monitoring method for rainwater treatment and reuse

Abstract

The invention provides a water quality up-to-standard monitoring method for rainwater treatment and reuse, and belongs to the technical field of water quality up-to-standard monitoring. Performing multi-stage processing flow and key parameter control; online water quality monitoring and multi-index comprehensive judgment; and remote monitoring and alarming are carried out. Therefore, after the collected rainwater is treated by a production sewage treatment station, the effluent reaches the water quality standard of non-drinking water and is used for monitoring water for production and environmental protection. According to the method, infrastructures of an existing rainwater treatment station are combined, and an intelligent monitoring and dynamic control mechanism is introduced, so that the operation efficiency and the water quality stability of a sewage treatment system are improved. The invention aims to solve the problems of lagging water quality detection, low automation degree and incapability of judging whether the water quality reaches the standard or not in real time in the existing monitoring method.

Description

TECHNICAL FIELD

[0001] The present application belongs to the technical field of water quality monitoring, and particularly relates to a water quality monitoring method for rainwater treatment and reuse. BACKGROUND

[0002] As mentioned in the prior art with patent publication number CN106836432B, the application of rainwater treatment and reuse is extensive.

[0003] At present, the monitoring method for rainwater collected and treated by a sewage treatment station for production and environmental water reuse in the prior art is not perfect. The existing scheme usually adopts centralized rainwater collection, direct entry into the sewage treatment system, manual sampling detection after treatment, and reuse after confirming that the water quality meets the standard. However, this monitoring method has problems such as response lag, low detection frequency, and inability to real-time feedback of water quality status, and it is difficult to ensure that the treated water continuously meets the “Water Quality Standard for Miscellaneous Use”. In addition, the online monitoring means is not effectively integrated, and there is a lack of automatic judgment and control mechanism, which may lead to misuse of non-standard water, equipment corrosion, environmental pollution, or unstable process, etc. SUMMARY

[0004] To solve the defects in the prior art, the present application provides a water quality monitoring method for rainwater treatment and reuse, aiming to solve the problems of water quality detection lag, low automation, and inability to real-time judge whether the water quality meets the standard in the existing monitoring method, and to provide a monitoring method that can realize real-time monitoring and automatic control of the water quality of treated rainwater, ensure that the reused water continuously and stably meets the standard, and improve the safety and efficiency of water resource reuse.

[0005] The present application uses the following technical solutions.

[0006] A water quality monitoring method for rainwater treatment and reuse, comprising:

[0007] Step 1: rainwater initial flow separation and pretreatment control;

[0008] Step 2: multi-stage treatment process and key parameter control;

[0009]

[0009] Step 3: online water quality monitoring and multi-index comprehensive judgment;

[0010] Step 4: remote monitoring and alarm.

[0011] Further, in step 1, during the rainwater collection stage, an intelligent primary flow separation device is arranged, which is equipped with a liquid level sensor and an electric control valve. The PLC controller automatically intercepts the rainwater collected in the previous 5-10 minutes by periodically closing the electric control valve every 5-10 minutes. Each time the PLC controller closes the electric control valve for 5-10 minutes, it subsequently opens the electric valve for 5-10 minutes to allow the settled rainwater to be sent to the sewage treatment system through the pipeline. During the period when the PLC controller closes the electric control valve, the separated rainwater enters the sedimentation tank through the rainwater collection pipeline for preliminary solid-liquid separation, removing suspended particles and macromolecular pollutants.

[0012] Further, in step 1, the liquid level sensor and the electric control valve of the intelligent primary flow separation device are connected to the PLC controller. The PLC controller is also connected to the central monitoring system through a 4G module connected thereto. The electric control valve is arranged on the pipeline connecting the sedimentation tank and the sewage treatment system, which communicates with the rainwater collection pipeline. The liquid level sensor is arranged in the sedimentation tank to detect the liquid level value of the sedimentation tank and transmit it to the PLC controller.

[0013] Further, in step 1, the intelligent primary flow separation device is also equipped with a water quality monitoring module connected to the PLC controller. The water quality monitoring module includes a COD sensor one, a turbidity sensor one, and a suspended solids concentration sensor one arranged in the sedimentation tank and connected to the PLC controller. The COD sensor one, the turbidity sensor one, and the suspended solids concentration sensor one detect the COD value, the turbidity value, and the suspended solids concentration value of the initial rainwater in the sedimentation tank and transmit them to the PLC controller. If the PLC controller detects that the COD value, the turbidity value, and the suspended solids concentration value of the initial rainwater in the sedimentation tank exceed the preset COD threshold value, the turbidity threshold value, and the suspended solids concentration threshold value, respectively, the PLC controller will double the periodic interval of 5-10 minutes, and transmit a warning message to the central monitoring system.

[0014] Further, in step 2, after the rainwater is settled in the sedimentation tank, it enters the multi-stage treatment unit of the sewage treatment station, which includes a coagulation sedimentation unit, a sand filtration or membrane filtration unit, and a disinfection unit in sequence. The coagulation sedimentation unit uses an intelligent dosing system connected to the PLC controller to automatically adjust the dosage of chemicals based on the turbidity and COD of the rainwater entering the coagulation sedimentation unit. The turbidity and COD of the rainwater entering the coagulation sedimentation unit are collected by turbidity sensor two and COD sensor two arranged in the coagulation sedimentation unit and connected to the PLC controller, and transmitted to the PLC controller. The sand filtration or membrane filtration unit is provided with a differential pressure sensor connected to the PLC controller. When the filter layer resistance value collected by the differential pressure sensor and transmitted to the PLC controller exceeds the set value, the PLC controller automatically starts the backwashing program. The disinfection unit is provided with a microorganism index detection device and a dosing device connected to the PLC controller to ensure that the microorganism index meets the standards.

[0015] Further, in step 3, a multi-parameter online water quality monitor is deployed at the effluent end of the sewage treatment system, which includes a turbidity sensor three, a pH meter, a COD sensor three, an ammonia nitrogen sensor, and a total coliform rapid detector connected with the PLC controller. The turbidity sensor three, the pH meter, the COD sensor three, the ammonia nitrogen sensor, and the total coliform rapid detector respectively detect the turbidity, pH, COD, ammonia nitrogen, and total coliform of the effluent end of the sewage treatment system in real time and transmit them to the PLC controller. The PLC controller calculates the turbidity, pH, COD, ammonia nitrogen, and total coliform of the effluent end of the sewage treatment system by using the built-in water quality standard judgment algorithm, sets a threshold value according to the calculation results, automatically compares and comprehensively evaluates each index, and judges whether the water quality meets the reuse standard.

[0016] Further, in step 3, the calculation formula of the water quality standard judgment algorithm built in the PLC controller is as follows:

[0017]

[0018] wherein is the comprehensive score of water quality, is the weight of the i-th index, is the measured value of the i-th key index, is the target value set for the i-th key index, is the maximum allowable deviation value set for the i-th key index.

[0019] Further, in step 3, when the comprehensive score is less than the threshold value, the PLC controller determines that the water quality does not meet the standard, automatically closes the effluent valve at the effluent end of the sewage treatment system, and returns the water at the effluent end to the sewage treatment system for secondary treatment. The PLC controller also transmits a pre-warning notice to the central monitoring system. When the comprehensive score is greater than or equal to the threshold value, the PLC controller determines that the water quality meets the standard, automatically opens the effluent valve at the effluent end of the sewage treatment system, and sends the treated water to the water storage tank connected with the effluent end.

[0020]

[0021] Further, in step 4, historical data is analyzed by using a machine learning algorithm to identify abnormal operation patterns of the sludge treatment system and provide early warning.

[0022] Further, in step 4, the method for analyzing historical data by using a machine learning algorithm to identify abnormal operation patterns of the sludge treatment system and provide early warning specifically includes:​​​​​​

[0023] Step 4-1: Data collection and preprocessing: Deploy the following monitoring devices in the wastewater treatment system:

[0024] Install a filter pool differential pressure sensor on the wastewater treatment system connected to the PLC controller: collect the filter pool differential pressure value every 5 minutes and transmit it to the PLC controller; install a water quality online analyzer for inlet and outlet water on the wastewater treatment system connected to the PLC controller: collect the COD, ammonia nitrogen and turbidity of inlet and outlet water every hour and transmit them to the PLC controller; the filter pool differential pressure value collected by the filter pool differential pressure sensor and the COD, ammonia nitrogen and turbidity of inlet and outlet water collected by the water quality online analyzer are the original data of the COD, ammonia nitrogen and turbidity of inlet and outlet water;

[0025] Step 4-2: Feature engineering and operation mode modeling, which specifically includes:

[0026] Step 4-2-1: Extract the following three types of features:

[0027] Time domain features: calculate the mean, variance, skewness, kurtosis and change rate of each item of multi-source data in the sliding window;

[0028] Frequency domain features: the first three main frequency amplitudes obtained after fast Fourier transform (FFT) of each item of multi-source data in the sliding window;

[0029] Correlation features: calculate the multivariate Pearson correlation coefficient matrix , is the first data of one item of multi-source data in the original data, is the first data of another item of multi-source data in the original data, is the mean of one item of multi-source data in the original data, is the mean of another item of multi-source data in the original data;

[0030] Step 4-2-2: Use principal component analysis method to reduce the dimension of the three types of features to obtain 12-dimensional features;

[0031] Step 4-2-3: Combine historical operation records to label data and build a supervised learning training set, which contains the following working condition labels:

[0032] Normal operation;

[0033] Filter clogging;

[0034] Medicine failure;

[0035] Equipment failure;

[0036] Step 4-3: Abnormal detection model construction and training to obtain abnormal score;​​

[0037] Step 4-4: If the abnormality score is higher than the pre-set value, the PLC controller transmits the abnormality score and alarm message to the central monitoring system for remote monitoring and alarm.

[0038] Furthermore, step 4-3 specifically includes:

[0039] Build the hybrid model architecture:

[0040] Supervised learning module: Uses LSTM network to identify time series anomalies; LSTM network includes:

[0041] Input layer: 72-hour time window × 12-dimensional features

[0042] Hidden layer: 2 layers of LSTM units, 64 nodes per layer

[0043] Output layer: Softmax classifier, outputting 4 types of working condition probabilities

[0044] Loss function: cross entropy loss ,in It is one of the original data. data, is the mean of one of the multiple source data in the original data;

[0045] Unsupervised learning module: Isolation forest is used to detect unknown anomalies, which includes:

[0046] 256 isolated trees were constructed, and the sampling depth was limited to 100;

[0047] Anomaly scoring formula: ,in For abnormality score, is the expected path length of 256 isolated trees, is the normalization factor.

[0048] The beneficial effects of the present invention are as follows:

[0049] This invention utilizes rainwater primary flow separation and pretreatment control; multi-stage treatment processes and key parameter control; online water quality monitoring and comprehensive multi-indicator assessment; and remote monitoring and alarming. This method monitors the quality of the collected rainwater after treatment at a production wastewater treatment station, ensuring that the effluent meets miscellaneous water quality standards and is suitable for production and environmental water use. This method integrates existing rainwater treatment station infrastructure and introduces intelligent monitoring and dynamic control mechanisms to improve the operational efficiency and water quality stability of the wastewater treatment system. BRIEF DESCRIPTION OF THE DRAWINGS

[0050] Figure 1is a flow chart of a water quality standard monitoring method for rainwater treatment and reuse in the present application. DETAILED DESCRIPTION

[0051] To make the purpose, technical solutions and advantages of the present application clearer, the technical solutions of the present application will be expressed clearly and completely below by combining the drawings in the embodiments of the present application. The embodiments expressed in the present application are only a part of the embodiments of the present application, not all the embodiments. According to the spirit of the present application, other embodiments obtained by those skilled in the art without making creative efforts are within the protection scope of the present application.

[0052] As shown in Figure 1 A water quality standard monitoring method for rainwater treatment and reuse, comprising:

[0053] Step 1: rainwater initial flow separation and pretreatment control;

[0054] The purpose of step 1 is to set an initial flow separation device to remove high-concentration pollutants in initial rainwater before the rainwater enters a sewage treatment station.

[0055] In the preferred but non-limiting embodiment of the present application, in step 1, during the rainwater collection stage, an intelligent initial flow separation device is set, which can automatically separate initial high-pollution rainwater in combination with a rainfall amount and pollutant concentration model to prevent it from entering the sewage treatment system. The intelligent initial flow separation device is equipped with a liquid level sensor and an electric control valve. According to a preset initial flow time or pollutant concentration threshold, a PLC controller automatically intercepts rainwater in the previous 5-10 minutes by periodically closing the electric control valve every 5-10 minutes. The PLC controller closes the electric control valve for 5-10 minutes each time and then opens the electric valve for 5-10 minutes to let the precipitated rainwater pass through the pipeline into the sewage treatment system. During the period when the PLC controller closes the electric control valve, the separated rainwater enters the sedimentation tank through the rainwater collection pipeline for preliminary solid-liquid separation to remove suspended particles and macromolecular pollutants and reduce the subsequent treatment load. The process is automatically operated by the PLC controller, and the liquid level value of the sedimentation tank is uploaded to the central monitoring system for monitoring.

[0056] In the preferred but non-limiting embodiment of the present application, in step 1, the liquid level sensor and the electric control valve of the intelligent initial flow separation device are connected with the PLC controller. The PLC controller is also in communication connection with the central monitoring system (which can be a computer) through a 4G module connected therewith. The electric control valve is arranged on a pipeline that is in communication between the sedimentation tank in communication with the rainwater collection pipeline and the sewage treatment system (i.e., the sewage treatment station). The liquid level sensor is arranged in the sedimentation tank to detect the liquid level value of the sedimentation tank and transmit it to the PLC controller.

[0057] In the preferred but non-limiting embodiment of the present application, in step 1, the intelligent initial flow separation device is also equipped with a water quality monitoring module connected with the PLC controller, the water quality monitoring module includes a COD sensor one, a turbidity sensor one and a suspended solids concentration sensor one arranged in the sedimentation tank and connected with the PLC controller, the COD sensor one, the turbidity sensor one and the suspended solids concentration sensor one detect the COD value, the turbidity value and the suspended solids concentration value of the initial rainwater in the sedimentation tank respectively and transmit them to the PLC controller, and when the PLC controller detects that the COD value, the turbidity value and the suspended solids concentration value of the initial rainwater in the sedimentation tank exceed the preset COD threshold value, turbidity threshold value and suspended solids concentration threshold value respectively, the PLC controller will double the periodic time interval of 5-10 minutes, and transmit the early warning information to the central monitoring system.

[0058] In the embodiment of the present application, in the rainwater collection system of a certain industrial plant, an intelligent initial flow separation device is arranged, which includes a liquid level sensor, an electrically controlled valve and a PLC controller. According to historical data and local rainfall characteristics, the initial flow time is set to the first 5 minutes. When the rainfall starts, the PLC controller starts timing, and the rainwater in the first 5 minutes is guided to the sedimentation tank for separate treatment through the electrically controlled valve, and the subsequent rainwater enters the sewage treatment system in the main treatment process for treatment.

[0059] The initial flow separation device also combines a water quality monitoring module to detect the COD, turbidity and suspended solids concentration of the initial rainwater in real time. For example, when the COD of the initial rainwater is detected to exceed the set threshold value of 150 mg / L (i.e. COD > 150 mg / L), the PLC controller automatically extends the initial flow time to 10 minutes and sends early warning information to the central monitoring system.

[0060] Step 2: Perform multi-stage treatment process and key parameter control;

[0061] The purpose of step 2 is to treat the rainwater in the sewage treatment station, and the treatment process includes sedimentation, filtration and disinfection.

[0062] In the preferred but non-limiting embodiment of the present application, in step 2, the rainwater is precipitated through the sedimentation tank and then enters the multi-stage treatment unit of the sewage treatment station, which sequentially includes a coagulation sedimentation unit, a sand filtration or membrane filtration unit, and an advanced oxidation or ultraviolet-chlorine combined disinfection unit; the coagulation sedimentation unit adopts an intelligent dosing system connected with a PLC controller, which automatically adjusts the dosage of reagents according to the turbidity and COD of the rainwater entering the coagulation sedimentation unit, thereby improving the removal efficiency; the turbidity and COD of the rainwater entering the coagulation sedimentation unit are collected by turbidity sensor two and COD sensor two respectively, which are arranged in the coagulation sedimentation unit and connected with the PLC controller; the sand filtration or membrane filtration unit is provided with a differential pressure sensor connected with the PLC controller; when the filter layer resistance value collected by the differential pressure sensor and transmitted to the PLC controller exceeds the set value, the PLC controller automatically starts the backwashing program; the structure of the PLC controller automatically starting the backwashing program is that a backwashing pump is arranged on the water supply pipeline facing the sand filtration or membrane filtration unit, and the backwashing pump is connected with the PLC controller; when the filter layer resistance value collected by the differential pressure sensor and transmitted to the PLC controller exceeds the set value, the PLC controller automatically starts the backwashing pump to backwash the sand filtration or membrane filtration unit; the disinfection unit is provided with a microorganism index detection device (such as a total coliform bacteria rapid detector) and a dosing device connected with the PLC controller, to ensure that the microorganism index meets the standard; when the microorganism index value of the disinfection unit is detected by the microorganism index detection device and transmitted to the PLC controller, the PLC controller automatically increases the dosage to the set amount and prolongs the contact time to the set value when the microorganism index value is close to the critical value. The entire treatment process is controlled by the PLC controller for data acquisition and process control, to ensure stable operation of each treatment unit.

[0063] In the embodiment of the present application, after the rainwater enters the sewage treatment station, it sequentially passes through the following treatment units:

[0064] Coagulation sedimentation unit: an intelligent dosing system is adopted to automatically adjust the dosages of polyaluminum chloride (PAC) and polyacrylamide (PAM) according to the turbidity and COD of the influent. For example, when the turbidity is > 50 NTU, the PAC dosage is automatically increased to 20 mg / L, and the PAM dosage is set to 0.5 mg / L.

[0065] Sand filtration unit: double-layer filter material (quartz sand + anthracite) is adopted, and a differential pressure sensor is arranged to monitor the filter layer resistance. When the differential pressure exceeds the set value (such as 0.1 MPa), the system automatically starts the backwashing program to ensure the continuous and efficient operation of the filter tank.

[0066] Disinfection unit: Utilizing a combined UV and sodium hypochlorite disinfection method, the PLC controller dynamically adjusts the sodium hypochlorite dosage based on microbial indicators (such as total coliform bacteria). When microbial indicators approach critical values, the system automatically increases the dosage to 0.5 mg / L and extends the contact time to 30 minutes.

[0067] Step 3: Online water quality monitoring and comprehensive judgment of multiple indicators;

[0068] The purpose of step 3 is to set up an online water quality monitoring device at the outlet of the sludge treatment system to detect key water quality indicators such as turbidity, pH value, COD, and total coliform group in real time; and compare the test results.

[0069] In a preferred but non-restrictive embodiment of the present invention, in step 3, a multi-parameter online water quality monitor is deployed at the outlet of the sewage treatment system. The multi-parameter online water quality monitor includes a turbidity sensor 3, a pH meter, a COD sensor 3, an ammonia nitrogen sensor, and a total coliform group rapid detector connected to a PLC controller. The turbidity sensor 3, the pH meter, the COD sensor 3, the ammonia nitrogen sensor, and the total coliform group rapid detector respectively detect key indicators such as turbidity, pH, COD, ammonia nitrogen, and total coliform group at the outlet of the sewage treatment system in real time and transmit them to the PLC controller; the PLC controller can use advanced detection technologies such as spectral analysis, electrochemical sensing, and flow cytometry to ensure data accuracy and response speed. The PLC controller has a built-in water quality compliance judgment algorithm to calculate the turbidity, pH, COD, ammonia nitrogen, and total coliform group at the outlet of the sewage treatment system, set thresholds based on this, and automatically compare and comprehensively evaluate each indicator to determine whether it meets the reuse standard. The turbidity, pH, COD, ammonia nitrogen and total coliform bacteria at the outlet of the sewage treatment system constitute the key indicators.

[0070] In a preferred but non-limiting embodiment of the present invention, in step 3, the calculation formula of the water quality compliance determination algorithm built into the PLC controller is:

[0071]

[0072] in is a comprehensive water quality score (between 0 and 1, 1 is the best), For the The weight of each indicator (which can be dynamically adjusted according to the reuse purpose), For the The measured values ​​of key indicators, The water quality of industrial water for urban wastewater recycling" (GB / T19923-2005) or the water quality of urban miscellaneous water for urban wastewater recycling" (GB / T18920-2020) The target value of the key index (such as turbidity 0.5 NTU), The first The maximum allowable deviation value of the key index.

[0073] In the preferred but non-limiting embodiment of the present application, in step 3, when the comprehensive score is not up to standard, the PLC controller automatically closes the outlet valve on the outlet end of the sewage treatment system, and the water on the outlet end is returned to the sewage treatment system for secondary treatment, and the PLC controller also transmits a warning notice to the central monitoring system;

[0074] When the comprehensive score is up to standard, the PLC controller automatically opens the outlet valve on the outlet end of the sewage treatment system, and the treated water is transported to the water storage tank connected to the outlet end for purposes such as factory greening, road spraying, and equipment cooling. The PLC controller is connected to the outlet valve. The structure for returning the water on the outlet end to the sewage treatment system for secondary treatment is that a pipeline is connected between the outlet end and the sewage treatment system, and a return pump connected to the PLC controller is provided on the pipeline, and the PLC controller can control the return pump to return the water on the outlet end to the sewage treatment system.

[0075] The PLC controller also sets an emergency treatment mode, which automatically starts a high-level oxidation unit (such as ozone oxidation) connected to the PLC controller to improve the treatment intensity when the water is still not up to standard after being treated by the sewage treatment system for three consecutive times.

[0076] Step 4: Remote monitoring and alarm.

[0077] The purpose of step 4 is to set up remote monitoring and alarm, and automatically notify the operator to intervene in abnormal situations.

[0078] In the preferred but non-limiting embodiment of the present application, in step 4, historical data is analyzed by a machine learning algorithm to identify abnormal operation patterns of the sludge treatment system (such as filter clogging and ineffective chemicals), and early warning is provided.

[0079] In the preferred but non-limiting embodiment of the present application, in step 4, the method for analyzing historical data by a machine learning algorithm to identify abnormal operation patterns of the sludge treatment system (such as filter clogging and ineffective chemicals) and providing early warning specifically includes:

[0080] Step 4-1: Data collection and preprocessing: the following monitoring equipment is deployed in the sewage treatment system:

[0081] A filter pool pressure difference sensor connected with the PLC controller is arranged on the sewage treatment system: the filter pool pressure difference value is collected every 5 minutes and transmitted to the PLC controller, the range is 0-100 kPa, and the accuracy is ±0.5%; an online water quality analyzer connected with the PLC controller is arranged on the sewage treatment system: the COD, ammonia nitrogen and turbidity of the influent and effluent are collected every hour and transmitted to the PLC controller, and the detection accuracy is ±2%; the PLC controller records the process parameters such as reagent dosage, equipment start-stop state and backwashing frequency; the filter pool pressure difference value collected by the filter pool pressure difference sensor and the COD, ammonia nitrogen and turbidity of the influent and effluent collected by the online water quality analyzer are the original data of the COD, ammonia nitrogen and turbidity of the influent and effluent;

[0082] The original data is processed by the edge computing node in the PLC controller, and the following steps are performed:

[0083] Outlier rejection: 3σ principle is used to filter outliers (such as COD >1000 mg / L) exceeding the process limit in the original data;

[0084] Missing value interpolation: linear interpolation method is used to complete the short breakpoint data of the original data;

[0085] Data normalization: Z-score standardization is performed on the multi-source data of the original data:

[0086]

[0087] wherein is the standardized value of a certain item of multi-source data of the original data, is the mean value of a certain item of multi-source data of the original data, is the standard deviation of a certain item of multi-source data of the original data;

[0088] Sliding window processing: a 72-hour time window is constructed as a sliding window, which is updated every hour with newly collected original data to form a structured time series data set;

[0089] Step 4-2: feature engineering and operation mode modeling, which specifically includes:

[0090] Step 4-2-1: the following three types of features are extracted:

[0091] Time domain features: mean, variance, skewness, kurtosis and change rate of each item of multi-source data in the sliding window are calculated;

[0092] Frequency domain features: the first three main frequency amplitudes obtained after fast Fourier transform (FFT) of each item of multi-source data in the sliding window;

[0093] Correlation features: multivariate Pearson correlation coefficient matrix , is the first data of one of the multi-source data of the original data is the first data of another of the multi-source data of the original data is the mean of one of the multi-source data of the original data is the mean of another of the multi-source data of the original data

[0094] Step 4-2-2: Dimensionality reduction is performed on the three types of features by using principal component analysis method to obtain 12-dimensional features; that is:

[0095] Calculate the feature covariance matrix;

[0096] Solve the eigenvalues and eigenvectors;

[0097] Retain the first k (the value of k can be set by oneself) principal components with cumulative contribution rate of 90%, construct a dimensionality reduction matrix, and obtain the feature vectors after dimensionality reduction.

[0098] Step 4-2-3: Combine the historical operation and maintenance record labeled data to construct a supervised learning training set, which contains the following working condition labels:

[0099] Normal operation (Normal);

[0100] Filter clogging (Filter Clogging);

[0101] Coagulant failure (Coagulant Failure);

[0102] Pump failure (Pump Failure);

[0103] Step 4-3: Abnormality detection model construction and training to obtain an abnormality score;

[0104] In the preferred but non-limiting embodiment of the application, step 4-3 specifically includes:

[0105] Construct a hybrid model architecture:

[0106] Supervised learning module: use LSTM network to identify time series anomaly; the LSTM network includes:

[0107] Input layer: 72-hour time window x 12-dimensional features

[0108] Hidden layer: 2 layers of LSTM units, each layer with 64 nodes

[0109] Output layer: Softmax classifier, outputting 4-class working condition probabilities

[0110] Loss function: cross-entropy loss ,​​ is the mean of the one of the plurality of sources of raw data is the mean of the one of the plurality of sources of raw data

[0111] Unsupervised learning module: Isolation Forest is used to detect unknown anomalies, which includes:

[0112] 256 Isolation Trees are constructed, and the sampling depth limit is 100;

[0113] Anomaly score formula: wherein is the anomaly score, is the path length expectation of the 256 Isolation Trees, is a normalization factor.

[0114] Step 4-4: If the anomaly score is higher than the pre-set value, the PLC controller transmits the anomaly score and the alarm message to the central monitoring system for remote monitoring and alarm, so as to inform the maintenance personnel to maintain the sewage treatment system.

[0115] The present application provides a monitoring method for the collected rainwater treated by the production sewage treatment station, and the effluent reaches the miscellaneous water quality standard, and is used for production and environmental protection water. The method combines the existing rainwater treatment station infrastructure, and improves the operation efficiency and water quality stability of the sewage treatment system by introducing intelligent monitoring and dynamic control mechanism.

[0116] After implementing the monitoring method of the present application in an industrial park, the operation data of the present application is shown in Table 1 as follows:

[0117] Table 1

[0118]

[0119] By introducing the monitoring method, the present application realizes comprehensive control of the rainwater treatment process, improves the water quality stability and operation efficiency, and meets the production and environmental protection water demand of the plant.

[0120] The beneficial effects of the present application are as follows compared with the prior art:

[0121] The present application controls the rainwater initial flow separation and pretreatment; performs multi-stage treatment process and key parameter control; performs online water quality monitoring and multi-index comprehensive judgment; and performs remote monitoring and alarm. Thus, the collected rainwater is treated by the production sewage treatment station, and the effluent reaches the miscellaneous water quality standard, and is used for production and environmental protection water. The method combines the existing rainwater treatment station infrastructure, and improves the operation efficiency and water quality stability of the sewage treatment system by introducing intelligent monitoring and dynamic control mechanism.​

[0122] It should be pointed out finally that the above embodiments are only used to illustrate the technical solutions of the present application but not to limit it, and although the present application has been described in detail with reference to the above embodiments, those ordinarily skilled in the art should understand that modifications or equivalent replacements to the specific embodiments of the present application can still be made without departing from the spirit and scope of the present application, and any modifications or equivalent replacements should be covered within the protection scope of the claims of the present application.

Claims

1. A method for monitoring the water quality of rainwater treatment and reuse, characterized in that: include: Step 1: Rainwater primary flow separation and pretreatment control; Step 2: Conduct multi-stage processing and key parameter control; Step 3: Online water quality monitoring and comprehensive judgment of multiple indicators; Step 4: Conduct remote monitoring and alarm.

2. The method for monitoring the water quality of rainwater treatment and reuse according to claim 1, characterized in that: In step 1, during the rainwater collection stage, an intelligent primary flow separation device is set up. The intelligent primary flow separation device is equipped with a liquid level sensor and an electric control valve. The PLC controller automatically intercepts the rainwater in the first 5 to 10 minutes by periodically closing the electric control valve every 5 to 10 minutes. Each time the PLC controller closes the electric control valve for 5 to 10 minutes, it then opens the electric valve for 5 to 10 minutes to allow the settled rainwater to be sent to the sewage treatment system through the pipe. During the period when the PLC controller closes the electric control valve, the separated rainwater is allowed to enter the sedimentation tank through the rainwater collection pipe for preliminary solid-liquid separation to remove suspended particles and large molecular pollutants.

3. The method for monitoring the water quality of rainwater treatment and reuse according to claim 2, characterized in that: In step 1, the liquid level sensor and electric control valve equipped with the intelligent primary flow separation device are connected to the PLC controller, and the PLC controller is also connected to the central monitoring system through the 4G module connected to it; the electric control valve is set on the pipe connecting the sedimentation tank connected to the rainwater collection pipe and the sewage treatment system; the liquid level sensor is set in the sedimentation tank to detect the liquid level value of the sedimentation tank and transmit it to the PLC controller.

4. The method for monitoring the water quality of rainwater treatment and reuse according to claim 3 is characterized in that: In step 1, the intelligent primary flow separation device is also equipped with a water quality monitoring module connected to the PLC controller. The water quality monitoring module includes a COD sensor 1, a turbidity sensor 1 and a suspended matter concentration sensor 1 connected to the PLC controller and arranged in the sedimentation tank. The COD sensor 1, the turbidity sensor 1 and the suspended matter concentration sensor 1 respectively detect the COD value, turbidity value and suspended matter concentration value of the initial rainwater in the sedimentation tank and transmit them to the PLC controller. When the PLC controller detects that the COD value, turbidity value and suspended matter concentration value of the initial rainwater in the sedimentation tank exceed the preset COD threshold, turbidity threshold and suspended matter concentration threshold respectively, the PLC controller will periodically double the duration every 5 to 10 minutes and transmit early warning information to the central monitoring system.

5. The method for monitoring the water quality of rainwater treatment and reuse according to claim 4, characterized in that: In step 2, the rainwater enters the multi-stage treatment unit of the sewage treatment station after sedimentation in the sedimentation tank, which includes a coagulation sedimentation unit, a sand filtration or membrane filtration unit and a disinfection unit in sequence; the coagulation sedimentation unit adopts an intelligent dosing system connected to the PLC controller, and automatically adjusts the dosage of the agent according to the turbidity and COD of the rainwater entering the coagulation sedimentation unit. The turbidity and COD of the rainwater entering the coagulation sedimentation unit are respectively collected by turbidity sensor 2 and COD sensor 2 installed in the coagulation sedimentation unit and connected to the PLC controller and transmitted to the PLC controller; the sand filtration or membrane filtration unit is provided with a pressure difference sensor connected to the PLC controller. When the filter layer resistance value collected by the pressure difference sensor and transmitted to the PLC controller exceeds the set value, the PLC controller automatically starts the backwash program; the disinfection unit is provided with a microbial index detection device and a dosing device connected to the PLC controller to ensure that the microbial index meets the standard.

6. The method for monitoring the water quality of rainwater treatment and reuse according to claim 5, characterized in that: In step 3, a multi-parameter online water quality monitor is deployed at the outlet of the sewage treatment system. The multi-parameter online water quality monitor includes a turbidity sensor 3, a pH meter, a COD sensor 3, an ammonia nitrogen sensor, and a total coliform rapid detector connected to a PLC controller. The turbidity sensor 3, the pH meter, the COD sensor 3, the ammonia nitrogen sensor, and the total coliform rapid detector respectively detect key indicators such as turbidity, pH, COD, ammonia nitrogen, and total coliform at the outlet of the sewage treatment system in real time and transmit the results to the PLC controller. The PLC controller has a built-in water quality compliance algorithm that calculates the turbidity, pH, COD, ammonia nitrogen, and total coliform bacteria at the outlet of the sewage treatment system. It then sets thresholds based on these values ​​and automatically compares and comprehensively evaluates each indicator to determine whether it meets the reuse standards.

7. The method for monitoring the water quality of rainwater treatment and reuse according to claim 6, characterized in that: In step 3, the calculation formula of the water quality compliance determination algorithm built into the PLC controller is: ； in For the comprehensive water quality score, For the The weight of the indicator, For the The measured values ​​of key indicators, For the The target value set for each key indicator, For the The maximum allowable deviation value set for each key indicator; In step 3, when the comprehensive score When the PLC controller determines that the water quality does not meet the standards, the PLC controller automatically closes the outlet valve on the outlet end of the sewage treatment system and returns the water at the outlet end to the sewage treatment system for secondary treatment. The PLC controller also transmits an early warning notification to the central monitoring system; When the comprehensive score When the PLC controller determines that the water quality meets the standard, the PLC controller automatically opens the outlet valve on the outlet end of the sewage treatment system and transports the treated water to the water storage tank connected to the outlet end.

8. The method for monitoring the water quality of rainwater treatment and reuse according to claim 7, characterized in that: In step 4, historical data is analyzed through machine learning algorithms to identify abnormal operating modes of the sludge treatment system and provide early warning.

9. The method for monitoring the water quality of rainwater treatment and reuse according to claim 8, characterized in that: In step 4, the machine learning algorithm is used to analyze historical data, identify abnormal operation patterns of the sludge treatment system, and provide early warning methods, including: Step 4-1: Data collection and pre-processing: Deploy the following monitoring equipment in the sewage treatment system: The sewage treatment system is equipped with a filter tank differential pressure sensor connected to a PLC controller: the filter tank differential pressure value is collected every 5 minutes and transmitted to the PLC controller; the sewage treatment system is equipped with an online water quality analyzer connected to the PLC controller for inlet and outlet water: the COD, ammonia nitrogen, and turbidity of the inlet and outlet water are collected every hour and transmitted to the PLC controller; the filter tank differential pressure value collected by the filter tank differential pressure sensor and the COD, ammonia nitrogen, and turbidity of the inlet and outlet water collected by the online water quality analyzer are the original data of the COD, ammonia nitrogen, and turbidity of the inlet and outlet water; Step 4-2: Feature engineering and operational modeling, which specifically includes: Step 4-2-1: Extract the following three types of features: Time domain features: Calculate the mean, variance, skewness, kurtosis, and rate of change of each multi-source data within the sliding window; Frequency domain features: The first three main frequency amplitudes are obtained by fast Fourier transforming the multi-source data in the sliding window; Correlation features: Obtaining the multivariate Pearson correlation coefficient matrix , It is one of the original data. data, It is the first item of another multi-source data in the original data. data, is the mean of one of the multi-source data in the original data, is the mean of another multi-source data in the original data; Step 4-2-2: Use principal component analysis to reduce the dimension of the three types of features to obtain 12-dimensional features; Step 4-2-3: Combine the annotated data of historical operation and maintenance records to construct a supervised learning training set, which includes the following working condition labels: Normal operation; Filter blockage; The medicine is ineffective; Equipment failure; Step 4-3: Build and train an anomaly detection model to obtain anomaly scores; Step 4-4: If the abnormality score is higher than the pre-set value, the PLC controller transmits the abnormality score and alarm message to the central monitoring system for remote monitoring and alarm.

10. The method for monitoring the water quality of rainwater treatment and reuse according to claim 9, characterized in that: Step 4-3 specifically includes: Build the hybrid model architecture: Supervised learning module: Uses LSTM network to identify time series anomalies; LSTM network includes: Input layer: 72-hour time window × 12-dimensional features Hidden layer: 2 layers of LSTM units, 64 nodes per layer Output layer: Softmax classifier, outputting 4 types of working condition probabilities Loss function: cross entropy loss ,in It is one of the original data. data, is the mean of one of the multiple source data in the original data; Unsupervised learning module: Isolation forest is used to detect unknown anomalies, which includes: 256 isolated trees were constructed, and the sampling depth was limited to 100; Anomaly scoring formula: ,in For abnormality score, is the expected path length of 256 isolated trees, is the normalization factor.

Images