Sludge concentration intelligent all-in-one machine system based on AI technology for water treatment

Abstract

The invention relates to an AI technology-based sludge concentration intelligent all-in-one machine system for water treatment, and relates to the technical field of AI sludge concentration analysis instruments. The system comprises an AI sludge concentration meter which collects related local data and uploads the data to a cloud server based on an AI algorithm technology; the PLC / DCS switch is used for collecting the running state of the sludge discharge pump, making corresponding decision judgment and executing actions; a server cluster is deduced, an automatic control system model based on raw water quality and sludge concentration is constructed, the sewage treatment state is sensed in real time, the growth and aging degree of sludge is directly calculated, and a reasonable decision is directly pushed; and the Internet of Things module is used as a sensor, so that the data is transmitted to the computer through the data acquisition card and the Internet of Things DTU equipment and is input into the artificial intelligence and mechanism hybrid model of the equipment as real-time data, and the optimal decision is obtained through real-time calculation. According to the application, the sludge concentration in the activated sludge process is detected, the sludge discharge amount is intelligently adjusted, and the effect of controlling the MLSS is achieved.

Description

Technical Field

[0001] This application relates to the technical field of AI sludge concentration analysis instruments, and particularly to an intelligent integrated sludge concentration system for water treatment based on AI technology. Background Art

[0002] With the acceleration of industrialization and urbanization processes and the continuous expansion of urban scale, the pressure of urban water resource shortage is increasing. The problems of water resource shortage and water environmental pollution are becoming increasingly prominent. The root cause of the urban water crisis is that the social cycle of water exceeds the scope that the natural cycle of water can bear.

[0003] At present, the activated sludge process is one of the most effective methods for removing organic pollutants. More than 95% of urban sewage treatment and about 50% of industrial wastewater treatment at home and abroad adopt the activated sludge process. This method has a strong purification function, and the efficiency of removing BOD (biochemical oxygen demand) and the concentration of activated sludge in the mixed liquor is high, both reaching more than 95%. Since this method relies on microbial treatment, the operating cost is relatively low, and it is suitable for various organic wastewater and large, medium and small sewage treatment plants. The activated sludge process mainly consists of a primary sedimentation tank, an aeration tank, a secondary sedimentation tank, an aeration system, and a sludge return system. After the wastewater passes through the primary sedimentation tank, it enters the aeration tank simultaneously with the activated sludge returned from the bottom of the secondary sedimentation tank. Through aeration, the activated sludge is in a suspended state and contacts the wastewater fully. The suspended solids and colloidal substances in the wastewater are adsorbed by the activated sludge, and the soluble organic substances in the wastewater are used by the microorganisms in the activated sludge as nutrients for their own reproduction, metabolized and transformed into biological cells, and oxidized into the final products (mainly CO2). The insoluble organic substances need to be first converted into soluble organic substances and then metabolized and utilized. Thus, the wastewater is purified. After purification, the wastewater and the activated sludge are separated in the secondary sedimentation tank. The upper layer of the effluent is discharged, and a part of the separated and concentrated sludge is returned to the aeration tank to ensure a certain concentration of activated sludge in the aeration tank, and the rest is surplus sludge, which is discharged by the sludge discharge pump of the system.

[0004] In the activated sludge process, controlling sludge discharge with sludge concentration (MLSS) means determining the sludge discharge amount under the condition of maintaining a constant sludge concentration in the aeration tank mixed liquor. First, a suitable MLSS concentration value is determined according to the actual process conditions. The MLSS of the conventional activated sludge process is generally between 1500 - 3000 mg / L. When the actual MLSS is higher than the MLSS value to be controlled, the MLSS value should be reduced by discharging surplus sludge. The sludge discharge amount can be calculated using the following formula:

[0005]

[0006] Where, V Wis the sludge discharge amount at this time, MLSS is the measured value in mg / L, MLSS0 is the concentration value determined according to the actual process in mg / L, V is the volume of the aeration tank in m 3 , RSS is the concentration of return sludge in mg / L.

[0007] The key point of biological phosphorus removal is to increase the proportion of polyphosphate-accumulating organisms (PAOs) in the activated sludge system, and at the same time, during the operation of the system, they grow and reproduce in large numbers, and the phosphorus content in the PAOs remains at a relatively high level when discharged from the system; in order to increase the proportion of PAOs in the activated sludge in the system, a more favorable environment and hydraulic conditions for the growth and reproduction of PAOs need to be created, that is, there are good anaerobic and aerobic environments in the process flow. The control of environmental factors in the anaerobic zone is particularly important for the growth and reproduction of PAOs and the realization of the phosphorus removal function. A high sludge concentration in the anaerobic zone is more beneficial to PAOs.

[0008] The efficiency of biological phosphorus removal is closely related to the sludge age. Only under a certain sludge age (about 3 days) can excessive phosphorus be effectively removed to achieve the phosphorus removal function. When the influent is certain, since the sludge concentration is directly proportional to the sludge age, therefore, when the sludge concentration exceeds a certain range, the higher the sludge concentration, the worse the corresponding phosphorus removal effect; in the activated sludge process, the sludge concentration cannot be selected too low for the following reasons: (1) If MLSS is too low, the volume V of the aeration tank will increase correspondingly, which is economically disadvantageous; (2) If MLSS is too low, foam is likely to be generated in the aeration tank. To prevent foam, generally, a sludge concentration of more than 2 kg / m 3 is required; (3) When the sludge concentration is very low, the amount of oxygen required is less. If MLSS is too low and the tank volume increases, the air supply per unit tank volume will be very small, which cannot meet the requirements of mixing in the tank, and additional stirring power needs to be increased; the sludge concentration cannot be selected too high for the following reasons: To increase MLSS, the sludge return ratio must be increased accordingly, the surface load of the secondary sedimentation tank must be reduced, and the residence time of the secondary sedimentation tank must be lengthened, which requires increasing the volume of the secondary sedimentation tank and the energy consumption of the return sludge; by adjusting the sludge discharge amount, the types and growth rates of microorganisms in the activated sludge can be changed, the amount of oxygen required can be changed, and the sedimentation performance of the sludge can be improved; based on the above, the present application provides an AI-based intelligent integrated sludge concentration system for water treatment to intelligently adjust the sludge discharge amount and achieve the control of MLSS. Summary of the Invention

[0009] In order to detect the sludge concentration in the activated sludge process, intelligently adjust the sludge discharge amount, and achieve the control of MLSS, the present application provides an AI-based intelligent integrated sludge concentration system for water treatment.

[0010] The AI-based intelligent integrated sludge concentration system for water treatment provided by the present application adopts the following technical solutions:

[0011] An intelligent integrated system for sludge concentration based on AI technology in water treatment, comprising:

[0012] An AI sludge concentration meter, which is used to collect relevant on-site data based on AI algorithm technology, upload it to the cloud server, and push corresponding decisions;

[0013] A PLC / DCS switch, which is used to realize the communication between the AI sludge concentration meter and the system, collect the operating status of the sludge discharge pump, and make judgments and execution actions for corresponding decisions;

[0014] A deduction server cluster, which is used to build an automatic control system model based on raw water quality and sludge concentration by using algorithms such as neural network, genetic algorithm, and gradient explosion, perceive the sewage treatment status in real time, directly calculate the growth and aging degree of sludge, and directly push reasonable decisions;

[0015] An Internet of Things module, which is used as a sensor to transmit data to a computer through a data acquisition card and an Internet of Things DTU device, and input it as real-time data into the artificial intelligence and mechanism hybrid model of this device to obtain the best decision in real-time calculation.

[0016] Preferably, the AI sludge concentration meter is used to measure the sludge concentration. A machine learning model obtained from a self-deduction server is set in the instrument control unit of the AI sludge concentration meter. The machine learning model is used to process the data of the terminal monitoring instrument fed back from the Internet of Things module in real time. The AI sludge concentration meter executes the inference task of the machine learning model, returns the result to the terminal monitoring instrument or the central control room for terminal execution, and processes relevant data collection and intelligent decision-making.

[0017] Preferably, the deduction server is responsible for executing calculation tasks.

[0018] Preferably, the deduction server cluster includes a training database, and a large amount of data for machine learning model training is stored in the training database.

[0019] Preferably, the deduction server cluster further includes a model observation server, and the model observation server is used to monitor the model server, the training database, and the terminal model calculation integrated machine in real time.

[0020] Preferably, the PLC / DCS switch is also used to segment and isolate the network, divide multiple virtual network segments to improve network security.

[0021] Preferably, the PLC / DCS switch is also used to identify the received data, accurately forward it to the target device, perform traffic control and management on different ports, users, and applications, optimize the network scenario, improve the reliability and stability of the network, optimize the data transmission method, and improve the transmission rate.

[0022] In summary, the present application includes at least one of the following beneficial technical effects:

[0023] 1. By introducing machine learning algorithms and big data analysis techniques, the present application can process and analyze a large amount of data collected by sensors in real time, adjust the sludge discharge strategy according to the real-time indicators of sewage, and predict the future sludge change trend by learning historical data, so as to make adjustments to the sludge discharge sequence in advance;

[0024] 2. The system of the present application adopts algorithm models such as neural networks to perform in-depth learning and pattern recognition on the data collected by sensors. These algorithms can help the system more accurately judge the sludge deposition condition and predict the possible future sludge changes. In this way, the intelligent integrated sludge concentration system based on AI technology can make corresponding AI adjustments before the sludge undergoes sudden changes, ensuring the continuity and stability of sewage treatment. BRIEF DESCRIPTION OF THE DRAWINGS

[0025] Figure 1 is a schematic structural diagram of an intelligent integrated sludge concentration system for water treatment based on AI technology according to an embodiment of the present application.

[0026] Description of the reference numerals: 1, AI sludge concentration meter; 2, PLC / DCS switch; 3, deduction server cluster; 4, Internet of Things module. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0027] The following further describes the present application in detail Figure 1 with reference to the accompanying drawings.

[0028] An embodiment of the present application discloses an intelligent integrated sludge concentration system for water treatment based on AI technology. Referring to Figure 1 , it includes an AI sludge concentration meter 1, a PLC / DCS switch 2, a deduction server cluster 3, and an Internet of Things module 4.

[0029] The AI sludge concentration meter is integrated with AI based on the original sludge concentration meter. A machine learning model obtained from the deduction server is deployed in its analysis instrument and is used to process in real time the data from the terminal monitoring instrument fed back by the Internet of Things module; the AI sludge concentration meter performs model inference tasks and returns the results to the terminal monitoring instrument or the central control room for terminal execution, and processes relevant data collection and intelligent decision-making. Its working principle is: only a small part of the infrared light emitted by the transmitter on the sensor can reach the detector after being absorbed, reflected, and scattered by the measured object during transmission. By measuring the transmittance of the transmitted light, the concentration of the sludge can be calculated.

[0030] PLC / DCS switches are used to realize the communication between AI sludge concentration meter and the system, collect the operating status of sludge pump, make corresponding decisions and execute actions; PLC / DCS switches are also used to segment and isolate the network, divide multiple virtual network segments to improve network security; PLC / DCS switches are also used to identify received data, accurately forward it to the target device, control and manage the flow of different ports, users and applications, optimize network environment, improve network reliability and stability, optimize data transmission methods, and increase transmission rate.

[0031] The inference server cluster is used to build an automatic control system model based on raw water quality and sludge concentration using neural networks, genetic algorithms, gradient explosion and other algorithms, to sense the sewage treatment status in real time, directly calculate the growth and aging degree of the sludge, and directly push reasonable decisions; the inference server includes a training database, which stores a large amount of data for machine learning model training; and also includes a model observation server, which is used to monitor the model server, training database and terminal model-calculating integrated machine in real time.

[0032] Deduction server (cloud backend):

[0033] The inference server is responsible for performing complex computing tasks such as simulation, prediction, and decision support. It usually has high-performance computing capabilities to handle large amounts of data and complex algorithms such as time series prediction models (such as ARIMA, LSTM networks), regression models (such as linear regression, random forests), or deep learning models.

[0034] Connection method: The simulation server is connected to the training database, model server, and intelligent computing machine through a network switch or router. This connection ensures efficient data transmission and coordinated execution of tasks. The simulation process used by the simulation server.

[0035] The Internet of Things module is used as a sensor to transmit data to the computer via the data acquisition card and the Internet of Things DTU device, and input it into the artificial intelligence and mechanism hybrid model of this device as real-time data, and calculate in real time to obtain the best decision; the deduction server cluster also includes a model observation server, which is used to monitor the model server, training database and terminal model-calculating integrated machine in real time.

[0036] Sludge discharge formula based on PID control:

[0037] PID controller is a common control algorithm used to regulate the operation of sludge pumps. Its basic formula is as follows:

[0038]

[0039] in,

[0040] u(t) is the output of the controller and the control signal for the sludge discharge pump;

[0041] K p is the proportional gain, used to adjust the influence of the current error;

[0042] K i is the integral gain, used to adjust the influence of error accumulation;

[0043] K d is the derivative gain, used to predict the change trend of the error;

[0044] e(t) is the error between the set value and the actual sludge concentration He Zijian;

[0045] t is time.

[0046] Sludge discharge formula based on machine learning:

[0047] For more complex AI models, such as neural networks or support vector machines, the formula will be more complex. The following is a simplified formula based on neural networks:

[0048]

[0049] Output is the output of the neural network, that is, the sludge discharge decision;

[0050] Among them,

[0051] σ is the activation function, used to introduce non-linear factors;

[0052] w i is the weight of the i-th neuron;

[0053] f i is the input feature function of the i-th neuron, which can be a linear or non-linear transformation of the input feature x i ;

[0054] b is the bias term;

[0055] n is the number of neurons;

[0056] x i is the input feature, such as sludge concentration, flow rate, SVI, etc.

[0057] Model training objective function:

[0058] During the training process, the AI model usually uses the loss function to measure the difference between the predicted value and the actual value. The following is the loss function formula adopted in this application:

[0059]

[0060] Among them,

[0061] L(θ) is the loss function, and θ is the model parameter (including weight and bias);

[0062] m is the number of training samples;

[0063] h θ (x (i) ) is the predicted value of the model for the i-th sample;

[0064] y (i) is the actual value of the i-th sample.

[0065] The training process of the AI model is to minimize the loss function L(θ) by adjusting the model parameter θ.

[0066] Training database (cloud background):

[0067] Function: The training database stores a large amount of data for machine learning model training. These data may include historical records, sensor data, user behavior, etc.

[0068] Connection method: The training database is directly connected to the deduction server so that the deduction server can directly obtain data from the database for model training.

[0069] Model observation server (cloud background):

[0070] Function: The model observation server can monitor the status and performance of components such as the model server, training database, and terminal intelligent computing all-in-one machine in real time. By collecting and analyzing the real-time data of these components, the model observation server can ensure the stability and reliability of the entire system.

[0071] Connection method: The model observation server is usually connected to the deduction server. It is responsible for coordinating the data flow and task execution between these components.

[0072] An intelligent all-in-one machine system for sludge concentration based on AI technology in water treatment:

[0073] Function: The intelligent all-in-one machine system for sludge concentration based on AI technology deploys a machine learning model obtained from the deduction server and is used to process the data from the terminal monitoring instrument in real time. It performs model inference tasks and returns the results to the terminal monitoring instrument or the deduction server.

[0074] Connection method: The intelligent all-in-one machine system for sludge concentration based on AI technology is connected to the deduction server, PLC / DCS, and local database through a switch. It reads the data from the central control and local database and sends the processing results to the deduction server and provides them for front-end display on the console.

[0075] Terminal monitoring instrument:

[0076] Function: The terminal monitoring instrument is responsible for collecting on-site data such as pH, flow rate, COD, etc. These data are the basis for the training and inference of machine learning models.

[0077] Connection method: The terminal monitoring instrument sends data to the PLC / DCS, and through the PLC / DCS, the data is sent to the local database for historical data storage.

[0078] Functions of the sludge concentration integrated machine equipment with AI technology of the invention:

[0079] The sludge concentration intelligent integrated machine system with AI technology deploys a machine learning model obtained from the deduction server and is used to process the data from the terminal monitoring instrument in real time and give corresponding instructions and controls. The sludge concentration intelligent integrated machine system with AI technology includes the following functions:

[0080] Data access:

[0081] The types of data that the sludge concentration intelligent integrated machine system with AI technology can access usually include the values or indicators obtained by various sensors, as well as the corresponding monitoring time. The data is sent in JSON format, including the data points collected by various monitoring instruments and related metadata, such as timestamps, device identifiers, etc. The data is transmitted to the intelligent computing integrated machine through a communication interface (such as a network interface, serial port, data bus, etc.), and the data is transmitted through a standard data transmission protocol (such as HTTP, etc.). The data access process involves receiving JSON data packets, parsing the fields therein to extract and process the data, and then passing it to the machine learning model for real-time processing and analysis.

[0082] Data analysis:

[0083] The sludge concentration intelligent integrated machine system with AI technology will preprocess, analyze and possibly detect anomalies in the accessed data. First, the accessed data is cleaned to handle the noise, error values and inconsistencies in the data to ensure data quality; for the missing measured values in the data, appropriate strategies are taken for processing, including the extraction of data with missing measured values and linear interpolation to fill in the missing measured data; at the same time, the possible outliers in the data are detected and processed to prevent them from having an adverse impact on the analysis results. Through these steps, it is ensured that the data accessed from the terminal monitoring instrument is a standard time series for real-time decision-making and application by the calculation model.

[0084] Data decision-making:

[0085] After the above data access and analysis, the standard time series data of the terminal monitoring instrument is obtained and input into the calculation model for sludge discharge decision calculation. The model includes an AI prediction module and a sludge discharge calculation module: The AI prediction module is based on a pre-trained large AI model and performs zero-shot prediction on the input time series data, capable of predicting the future data change trend; The sludge discharge calculation module establishes a sludge discharge rule library based on the built-in chemical reaction equation, determines the specific sludge discharge cycle according to the input prediction data, and formulates a sludge discharge decision based on expert experience. The final sludge discharge decision is transmitted to the automatic control system and the terminal sludge discharge equipment through the data interface.

[0086] The implementation principle of the sludge concentration intelligent all-in-one machine system based on AI technology in the embodiment of this application is as follows: This application collects the real-time instruments of the influent water quality of the sewage, and the data such as the influent water quality and sludge concentration are directly uploaded to the AI sludge concentration meter all-in-one machine and the cloud platform through the Internet of Things communication device. The deduction server performs prediction and decision-making, and conducts AI prediction through the on-site set AI sludge concentration meter, decides to control the sludge discharge valve and the sludge discharge pump, predicts the future water quality change trend, and thus makes sludge discharge adjustment in advance to maintain the stability of the sludge concentration and reduce manual operation.

[0087] The above are all the preferred embodiments of this application. Without restricting the protection scope of this application based on this, therefore: Any equivalent changes made according to the structure, shape, and principle of this application shall be covered within the protection scope of this application.

Claims

1. A water treatment sludge concentration intelligent integrated system based on AI technology, characterized in that: include: AI sludge concentration meter is used to collect relevant local data and upload it to the cloud server based on AI algorithm technology, and push corresponding decisions; PLC / DCS switch, used to realize the communication between the AI ​​sludge concentration meter and the system, collect the operating status of the sludge pump and make corresponding decisions and execute actions; The deduction server cluster is used to build an automatic control system model based on raw water quality and sludge concentration using neural networks, genetic algorithms, gradient explosion and other algorithms, to sense the sewage treatment status in real time, directly calculate the growth and aging degree of sludge, and directly push reasonable decisions; The IoT module is used as a sensor to transmit data to the computer via the data acquisition card and the IoT DTU device, and input the data into the artificial intelligence and mechanism hybrid model of this device as real-time data, so as to obtain the best decision through real-time calculation.

2. According to claim 1, a water treatment sludge concentration intelligent integrated machine system based on AI technology is characterized by: The AI ​​sludge concentration meter is used to measure sludge concentration. The instrument control unit of the AI ​​sludge concentration meter is provided with a machine learning model obtained from the deduction server. The machine learning model is used to process in real time the data of the terminal monitoring instrument fed back from the Internet of Things module. The AI ​​sludge concentration meter executes the reasoning task of the machine learning model and returns the result to the terminal monitoring instrument or the central control room executed by the terminal, as well as processes related data collection and intelligent decision-making.

3. According to claim 2, a water treatment sludge concentration intelligent integrated machine system based on AI technology is characterized by: The deduction server is responsible for executing computing tasks.

4. According to claim 3, a water treatment sludge concentration intelligent integrated machine system based on AI technology is characterized by: The deduction server cluster includes a training database, which stores a large amount of data for machine learning model training.

5. According to claim 4, a water treatment sludge concentration intelligent integrated machine system based on AI technology is characterized by: The deduction server cluster also includes a model observation server, which is used to monitor the model server, training database and terminal model-computing integrated machine in real time.

6. According to claim 1, a water treatment sludge concentration intelligent integrated machine system based on AI technology is characterized by: The PLC / DCS switch is also used to segment and isolate the network and divide it into multiple virtual network segments to improve network security.

7. The water treatment sludge concentration intelligent integrated machine system based on AI technology according to claim 6 is characterized by: The PLC / DCS switch is also used to identify received data, accurately forward it to the target device, control and manage traffic for different ports, users and applications, optimize network traffic, improve network reliability and stability, optimize data transmission methods, and increase transmission rates.

Images