Early warning method and system for proliferation of nitrite oxidizing bacteria in synchronous short-cut nitrification and denitrification

Abstract

The invention discloses an early warning method for proliferation of nitrite oxidizing bacteria in synchronous short-cut nitrification and denitrification, which comprises the following steps of: detecting ammonia nitrogen and COD (Chemical Oxygen Demand) of inlet and outlet water in a continuous flow synchronous short-cut nitrification and denitrification process by using an online water quality detector; the water quality data is processed by means of a programming platform, and whether the NOB in the reaction system is rapidly proliferated or not is early warned by utilizing the trained XGBoost model, so that the breakdown of the short-cut nitrification system caused by the mass growth of the NOB is avoided. According to the method, accurate early warning can be carried out in advance at the initial stage of rapid proliferation of the NOB in the synchronous short-cut nitrification and denitrification system, the detection cost of the NOB can be reduced while stable operation of short-cut nitrification is maintained, and online monitoring of the NOB in the reaction system and long-term stable and efficient operation of the synchronous short-cut nitrification and denitrification process are facilitated.

Description

Technical Field

[0001] The present invention relates to sewage treatment technology, and in particular to a method and system for early warning of rapid proliferation of nitrite oxidizing bacteria (NOB) in synchronous short-range nitrification and denitrification by using XGBoost (extreme gradient boosting model). Background Art

[0002] Simultaneous Partial Nitrification and Denitrification (SPND) refers to the simultaneous occurrence of short-term nitrification and denitrification reactions in the same reactor, controlling the nitrification reaction in the nitrification stage and using nitrite nitrogen for denitrification. Compared with the traditional biological denitrification nitrification-denitrification process, the synchronous short-term nitrification and denitrification has the advantages of reducing the production of residual sludge, saving 25% of aeration energy consumption, and 40% of carbon source addition; compared with the short-term nitrification and denitrification process, it has the advantages of saving land and reducing greenhouse gas emissions. However, in the process of converting ammonia nitrogen to nitrite nitrogen, it is easy to peroxidize nitrite nitrogen to nitrate nitrogen, turning the short-term nitrification and denitrification reaction into a nitrification and denitrification reaction. Therefore, the accumulation of nitrite nitrogen is the key to the normal progress of the synchronous short-term nitrification and denitrification.

[0003] At present, most measures to prevent the short-term nitrification process from turning into a nitrification process are to control the operating conditions to inhibit and elute nitrite oxidizing bacteria (NOB), such as strictly controlling dissolved oxygen (DO) and adjusting sludge retention time (SRT) to inhibit the growth of NOB. However, under long-term inhibition conditions, NOB will develop a certain tolerance, that is, the short-term nitrification process cannot maintain long-term stability. Therefore, even under strict operating conditions, it is still necessary to predict whether NOB will proliferate rapidly. Since ammonia nitrogen oxidation occurs in an aerobic environment and nitrate nitrogen and nitrite nitrogen reduction occur in an anoxic environment, the DO concentration in the reaction system will directly affect the nitrification and denitrification processes. Therefore, the concentration of DO in the synchronous short-term nitrification and denitrification system can be reflected by indicators such as ammonia oxidation rate and organic matter removal rate. Since NOB is more sensitive to DO concentration than AOB, these indicators can be used to further reflect the activity of NOB. However, there is a complex nonlinear relationship between these indicators, and it is difficult to make an intuitive judgment on the growth of NOB based on these indicators based on manual experience and traditional control methods alone.

[0004] In recent years, the application of machine learning in the field of sewage treatment process control has been increasing. Its powerful ability to process complex data has provided new possibilities for solving the problem of predicting the rapid proliferation of NOB in synchronous short-range nitrification and denitrification. Among them, the XGBoost (extreme gradient boosting) model, as a machine learning algorithm based on gradient boosting tree, has performed well in prediction and classification problems and has been widely used in various fields. The XGBoost model is an ensemble learning algorithm that improves the accuracy of prediction by training multiple weak classifiers simultaneously and combining them into a strong classifier. The model shows excellent performance in processing large-scale data and complex features, and has strong generalization ability. In the prediction of NOB proliferation in synchronous short-range nitrification and denitrification, the XGBoost model can accurately predict the trend of NOB proliferation by learning patterns and association laws in a large amount of data. Summary of the invention

[0005] In view of the problem that it is difficult to accurately predict the large-scale proliferation of NOB in the process of synchronous short-range nitrification and denitrification in advance, the present invention provides a method for early warning of the rapid proliferation of NOB in synchronous short-range nitrification and denitrification using XGBoost, and introduces the intelligent algorithm model XGBoost to solve the problem of insufficient accuracy and adaptability of NOB proliferation prediction in the prior art. By establishing an early warning system based on the XGBoost model, it is possible to more accurately reflect the rapid proliferation of NOB in the process of synchronous short-range nitrification and denitrification, and provide technical guarantee for the long-term and efficient operation of the synchronous short-range nitrification and denitrification process.

[0006] To achieve the above object, the present invention provides an early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification, which is characterized in that it includes:

[0007] S1. Use an online ammonia nitrogen detector and an online COD detector to collect water quality data at the inlet and outlet of the synchronous short-range nitrification and denitrification reactor;

[0008] S2. Preprocess and extract features of the collected water quality data to determine the most valuable characteristic variables for predicting rapid proliferation of NOB;

[0009] S3. Input the water quality data features obtained in step S2 into the pre-trained XGBoost model to obtain the rapid proliferation warning results of NOB in the process of synchronous short-range nitrification and denitrification.

[0010] Furthermore, the aeration flow rate was controlled by a rotor flowmeter; an online pH electrode was used to monitor the pH in real time and an alkali dosing pump was used to ensure that the pH of the reaction system was not less than 6.7; the reaction temperature was maintained at 35°C by heating. Water pumps were placed at the inlet and outlet of the reactor to take samples and send them to the online ammonia nitrogen detector and online COD detector for water quality testing.

[0011] Furthermore, online pH meters, online water quality testers and other equipment transmit water quality parameters to the client's programming software through PLC for processing;

[0012] Furthermore, the nitrate nitrogen in the effluent when collecting sample data in the previous step was measured using a spectrophotometer, and the collected data were artificially grouped. It was stipulated that if the currently measured nitrate nitrogen concentration was 10 mg / L higher than the nitrate nitrogen concentration at the previous moment, it was considered that NOB rapid proliferation occurred at this time, and the data group with NOB rapid proliferation was marked as 1, and the data group without NOB rapid proliferation was marked as 0;

[0013] Furthermore, the collected data are preprocessed, including outlier removal and missing value interpolation;

[0014] Furthermore, the random forest model is used to score the contribution of all collected variables, and the variables with higher contribution to the prediction results are selected as the input features of the model. The collected variables are processed by dimensionality reduction to reduce the calculation amount of the model, improve the model operation efficiency, and avoid model overfitting;

[0015] Furthermore, the XGBoost model was established by training the preprocessed data, and the model parameters were optimized to improve the prediction performance of the model, so that the trained model could learn the rapid proliferation law of NOB in the process of synchronous short-range nitrification and denitrification.

[0016] Furthermore, the trained XGBoost model is used to warn the real-time data of the system operation. By inputting the real-time characteristic variable values ​​into the trained model, the rapid proliferation warning results of NOB in the synchronous short-range nitrification and denitrification process can be obtained.

[0017] The present invention also provides a system for implementing the above-mentioned early warning method, which is characterized in that it includes:

[0018] Synchronous short-range nitrification and denitrification reaction device;

[0019] An online COD detector and an online ammonia nitrogen detector are respectively arranged at the water inlet and the water outlet of the synchronous short-range nitrification and denitrification reaction device;

[0020] A data processing unit, used for receiving and processing data from the detector and generating characteristic data;

[0021] XGBoost model training and prediction unit, used to train the XGBoost model and receive real-time data features to predict the rapid proliferation of NOB;

[0022] The early warning and operation adjustment unit is used to issue an early warning according to the prediction results of the XGBoost model and adjust the operating conditions of the reaction device according to the early warning results.

[0023] Compared with the prior art, the present invention has the following beneficial effects:

[0024] 1) By inputting the inlet and outlet water quality data into the trained XGBoost model, the growth of NOB can be accurately predicted without the need for additional detection equipment, reminding operators whether to implement measures to suppress NOB, thereby achieving long-term, efficient and stable operation of the short-range nitrification process.

[0025] 2) The XGBoost machine learning model introduced can learn from a large amount of data the laws of NOB proliferation and other water quality parameters during the synchronous short-range nitrification and denitrification process, and make accurate predictions based on the more easily available water quality parameters. When evaluating the classification effect of the model, the accuracy of the model reached 100%. Compared with traditional statistical models and empirical formulas, the XGBoost model can better capture complex system behaviors and nonlinear relationships, and improve prediction accuracy.

[0026] 3) The XGBoost model used is a general machine learning algorithm with the advantage of almost zero cost to achieve NOB prediction. The XGBoost model and the prediction system based on the model have a fast training and prediction speed. This enables the method to provide real-time early warning of the proliferation of NOB during the synchronous short-range nitrification and denitrification process. Operators can take corresponding control measures based on the early warning results, such as adjusting the aeration intensity, controlling the water inlet load, etc., to avoid the accumulation of nitrate nitrogen caused by excessive proliferation of NOB, providing important guidance for process control.

[0027] Combined with the above advantages, the present invention has significant technical effects such as high accuracy, strong adaptability, high efficiency and optimized process operation, which can effectively solve the problem of difficult to accurately warn of NOB proliferation in the prior art, and can provide early warning in the early stage of rapid NOB proliferation, providing important support and guidance for the regulation of short-term nitrification and denitrification processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0028] Figure 1 It is a schematic diagram of the early warning system for the proliferation of nitrite oxidizing bacteria in the synchronous short-range nitrification and denitrification of the present invention.

[0029] Figure 2 It is a data processing flow chart of the early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification of the present invention.

[0030] Figure 3It is a model variable importance measurement diagram of the early warning method for the proliferation of nitrite oxidizing bacteria in the synchronous short-range nitrification and denitrification of the present invention.

[0031] Figure 4 It is a result schematic diagram of a preferred embodiment of the early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification of the present invention.

[0032] Figure 5 It is a result schematic diagram of another preferred embodiment of the early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification of the present invention. DETAILED DESCRIPTION

[0033] In order to make the technical problems, technical solutions and beneficial effects to be solved by the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0034] In the following description, specific details such as specific internal procedures and techniques are provided for the purpose of illustration rather than limitation, so as to provide a thorough understanding of the embodiments of the present invention. However, it should be clear to those skilled in the art that the present invention may be implemented in other embodiments without these specific details. In other cases, detailed descriptions of well-known systems, devices, circuits, and methods are omitted to prevent unnecessary details from obstructing the description of the present invention.

[0035] An early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification, the steps are as follows:

[0036] Build a synchronous short-range nitrification and denitrification reactor, such as Figure 1 An aeration sand head is placed at the bottom of the reactor and the rotor flowmeter is used to control the aeration amount of the aeration pump into the reactor; an online pH detection electrode is placed in the reactor to monitor the pH change in real time and to link with the alkali pump to ensure that the pH value of the reaction system is not less than 6.7; there is a water bath heating layer on the outer layer of the reactor to keep the reaction system at 35°C, so that the ammonia oxidizing bacteria and denitrifying bacteria are in suitable environmental conditions.

[0037] The reactor water quality data processing process is as follows Figure 2 As shown in the figure, the online ammonia nitrogen detector and online COD detector are used to automatically and regularly detect the ammonia nitrogen and COD concentrations at the outlet and inlet of the reactor, and the test results are exchanged with the PLC through the 485 protocol. The programming software in the host computer completes the reading and writing of data in the PLC through the Modbus TCP protocol.

[0038] The data collected in the early stage was used to train the XGBoost model and the trained model was used to predict whether NOB would proliferate rapidly.

[0039] The details are as follows:

[0040] First, the reactor operation data was collected. The collected data included influent ammonia nitrogen concentration, HRT, effluent ammonia nitrogen concentration, effluent nitrite nitrogen concentration, effluent nitrate nitrogen concentration, aeration rate, influent nitrogen load (NLR), influent organic load (OLR), ammonia oxidation rate (ARR), organic oxidation rate (ORR) and pH. The influent and effluent ammonia nitrogen concentrations were measured by an online ammonia nitrogen detector, and the effluent nitrite nitrogen concentrations and effluent nitrate nitrogen concentrations were measured by a spectrophotometer. Whether NOB proliferated rapidly was marked based on the effluent nitrate nitrogen concentration.

[0041] The collected data is then preprocessed. First, the obviously outlier abnormal data and missing data are corrected and filled, and the abnormal data or missing data are replaced or filled by averaging the previous and next values ​​of the outlier. Then feature extraction is performed on the various types of data collected. Since there are many features collected in the early stage, more features will cause the model to overfit, the generalization ability will decrease, and the prediction performance will decrease. At the same time, it will also lead to an increase in the cost of training the model and predicting data. Therefore, feature selection is needed to reduce the risk of overfitting, improve the generalization ability of the model, reduce the computational cost, improve the interpretability of the model, and improve the accuracy of the prediction to a certain extent. The present invention uses a random forest algorithm to calculate the importance of each input variable in the model, such as Figure 3 As shown. Figure 3 It can be seen that the importance scores of the three variables ARR, ORR, and effluent ammonia nitrogen concentration are relatively high. In order to further explore the relationship between the three variables and further reduce the dimension of the features, the three selected variables are transformed by linear combination, differentiation, etc., and then the importance of the variables is predicted for the transformed data. Finally, the features of the training model are determined to be effluent ammonia nitrogen concentration, ARR, d'ARR, and d'ORR, among which d'ARR and d'ORR are the first-order differentials of ARR and ORR.

[0042] After determining the feature variables, the prediction model can be trained. The preprocessed data is randomly divided into training sets and test sets, and the training parameters of the XGBoost binary classifier are set, such as learning rate, iteration rounds, loss function, etc. Then the model is trained and the performance of the model is evaluated by accuracy. The results show that the accuracy of the XGBoost model predicted by the 60-day feature data collected in the early stage has reached 100%.

[0043] After the model is trained, it is only necessary to collect the ammonia nitrogen and COD concentrations measured by the online water quality detector. There is no need to measure the nitrate nitrogen by spectrophotometer. At this time, the real-time data can be processed and an early warning can be issued on whether a rapid growth of NOB has occurred in the synchronous short-range nitrification and denitrification system.

[0044] Example 1

[0045] This embodiment provides a method for predicting the rapid proliferation of NOB in synchronous short-range nitrification and denitrification using XGBoost. The reaction device is as follows: Figure 1 As shown, the programming platform used is Python platform, and online COD detector and online ammonia nitrogen detector are set at the water inlet and outlet of the UASB reactor to collect water quality data.

[0046] First, the collected data set is annotated. If the nitrate nitrogen concentration at the current moment is 10 mg / L higher than the nitrate nitrogen concentration at the previous moment, it is annotated as 1, otherwise it is annotated as 0. The annotation 1 indicates that rapid proliferation of NOB has occurred at this time, and the annotation 0 indicates that rapid proliferation of NOB has not occurred. Subsequently, the XGBoost model was trained with the annotated effluent ammonia nitrogen concentration, ARR, d'ARR, and d'ORR and used for real-time early warning of rapid proliferation of NOB in synchronous short-range nitrification and denitrification. At the same time, in order to verify the effectiveness of the model, the effluent nitrate nitrogen concentration was still monitored daily by a spectrophotometer during the real-time early warning stage of the model.

[0047] On the 17th day of the real-time operation phase of the model, the online ammonia nitrogen detector detected that the concentration of inlet ammonia nitrogen was 395.8 mg / L, and the concentration of effluent ammonia nitrogen was 75.62 mg / L. The online COD detector detected that the concentration of inlet COD was 742 mg / L, and the concentration of effluent COD was 548.98 mg / L. The online water quality detector transmitted the detected data to the python platform through the communication protocol for processing, such as Figure 2 As shown in Figure 2, after data processing, the ARR of the reactor is 0.5763 Kg-N / m 3 d, the first-order differential of ARR is 0.1074, and the first-order differential of ORR is -0.3114. The four features obtained by processing are then input into the trained XGBoost model for prediction. The prediction result of the model is 1, that is, the model issues a warning of rapid proliferation of NOB at this time. The prediction results of the model and the effluent nitrate nitrogen concentration measured by the spectrophotometer are shown in Figure 4 As shown in the figure, it can be seen that the nitrate nitrogen concentration in the effluent increased from 9.81 mg / L on the 16th day to 20.75 mg / L on the 17th day, and the model also accurately issued an early warning on the 17th day. After seeing the early warning information, we reduced the aeration volume and timely controlled the nitrate nitrogen in the effluent to a lower level. This embodiment shows that the early warning model of the present invention can timely inhibit the activity of NOB.

[0048] Example 2

[0049] The device and implementation method of this embodiment are the same as those of Embodiment 1 and will not be described here. On the 30th day of the real-time operation stage of the model, the online ammonia nitrogen detector detected that the concentration of ammonia nitrogen in the inlet water was 241.2 mg / L, and the concentration of ammonia nitrogen in the outlet water was 29.78 mg / L. The online COD detector detected that the concentration of COD in the inlet water was 1330.19 mg / L, and the concentration of COD in the outlet water was 356.98 mg / L. The system input the processed characteristic values ​​into the model, and the model's prediction result was 1. The prediction results of the model and the nitrate nitrogen concentration of the outlet water measured by the spectrophotometer are shown in Figure 1. Figure 5 As shown in the figure, it can be seen that the nitrate nitrogen in the effluent at this time is 8.11 mg / L, which is significantly increased compared with the nitrate nitrogen of 1.32 mg / L measured at the previous sampling time, indicating that NOB has rapidly proliferated. Subsequently, we took measures to reduce HRT and increase the influent load, so that the concentration of nitrate nitrogen in the effluent gradually decreased. This embodiment shows that even if the nitrate nitrogen in the current effluent is not 10 mg / L higher than the nitrate nitrogen at the previous sampling time, the XGBoost model proposed in the present invention can still warn of the rapid proliferation of NOB in the synchronous short-range nitrification and denitrification system, indicating the good adaptability of the model.

[0050] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. An early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification, It is characterized in that include: S1. Use an online ammonia nitrogen detector and an online COD detector to collect water quality data at the inlet and outlet of the synchronous short-range nitrification and denitrification reactor; S2. Preprocess and extract features of the collected water quality data to determine the most valuable characteristic variables for predicting rapid proliferation of NOB; S3. Input the water quality data features obtained in step S2 into the pre-trained XGBoost model to obtain the rapid proliferation warning results of NOB in the process of synchronous short-range nitrification and denitrification.

2. The early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification according to claim 1, It is characterized in that In the step S1, sampling points are set at the water inlet and outlet of the synchronous short-range nitrification and denitrification reactor, samples are taken by a water pump and sent to an online ammonia nitrogen detector and an online COD detector for water quality detection, and the ammonia nitrogen and COD concentrations at the water inlet and outlet are automatically detected at regular intervals.

3. The early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification according to claim 2, It is characterized in that An aeration sand head is provided at the bottom of the synchronous short-range nitrification and denitrification reactor to control the aeration volume, ensure that the dissolved oxygen concentration in the reactor is moderate, and maintain the pH value of the reaction system not less than 6.7 through the linkage of the online pH detection electrode and the alkali pump, thereby ensuring the appropriate growth range of ammonia oxidizing bacteria and denitrifying bacteria.

4. The early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification according to claim 2 or 3, It is characterized in that The synchronous short-range nitrification and denitrification reactor uses a water bath heating layer to maintain the temperature of the reaction system at 35°C.

5. The early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification according to claim 1, It is characterized in that The preprocessing in step S2 includes correcting and filling abnormal and missing data; the feature extraction uses a random forest model to rank the importance of the input variables of the XGBoost model, and selects the effluent ammonia nitrogen concentration, ammonia nitrogen oxidation rate, the first-order differential of the ammonia nitrogen oxidation rate, and the first-order differential of the organic matter oxidation rate as input features of the XGBoost model.

6. The early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification according to claim 1, It is characterized in that The pre-trained XGBoost model in step S3 refers to optimizing the performance of the XGBoost model by using the characteristics of water quality data and setting appropriate training parameters (such as learning rate, iteration rounds, etc.).

7. The early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification according to claim 1, It is characterized in that The warning result in step S3 includes that if the nitrate nitrogen concentration in the effluent at the current moment is higher than that in the effluent at the previous moment by more than 10 mg / L, it is considered that rapid proliferation of NOB has occurred and the model issues a warning; otherwise, the model does not issue a warning.

8. The early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification according to claim 1, It is characterized in that The step S3 also includes monitoring the nitrate nitrogen concentration in the effluent by a spectrophotometer to verify the accuracy of the model warning, and adjusting the operating conditions according to the warning results to control the activity of NOB.

9. A system for implementing the early warning method for the proliferation of nitrite oxidizing bacteria in synchronous short-range nitrification and denitrification according to any one of claims 1 to 8, It is characterized in that include: Synchronous short-range nitrification and denitrification reaction device; An online COD detector and an online ammonia nitrogen detector are respectively arranged at the water inlet and the water outlet of the synchronous short-range nitrification and denitrification reaction device; A data processing unit, used for receiving and processing data from the detector and generating characteristic data; XGBoost model training and prediction unit, used to train the XGBoost model and receive real-time data features to predict the rapid proliferation of NOB; The early warning and operation adjustment unit is used to issue an early warning according to the prediction results of the XGBoost model and adjust the operating conditions of the reaction device according to the early warning results.

Images