Landfill leachate low-carbon intelligent treatment and recycling method based on'end-edge-cloud 'cooperative control

Abstract

The invention discloses a landfill leachate low-carbon intelligent treatment and recycling method based on'end-edge-cloud 'cooperative control. The method comprises the following steps: the terminal layer obtains real-time monitoring data of each sensor and sends the data to the edge layer, the edge layer adopts a machine learning algorithm to perform multi-target prediction on an effluent quality index, system comprehensive energy consumption and an anaerobic hydrogen production rate, and then a target function is constructed to obtain an operating parameter ideal value of each module; according to the ideal value and the real-time monitoring data of the terminal layer sensor, the key process operation condition of the system is adjusted, so that the optimal result of the effluent quality index, the anaerobic hydrogen production rate and the comprehensive energy consumption of the system is obtained. According to the invention, the operation parameters of the biochemical system are optimized and controlled through the'end-edge-cloud 'Internet of Things sensing network, so that the intelligent management and control of the landfill leachate treatment and resource / energy conversion overall process flow are realized, and the dynamic adjustment of key influence factors in the landfill leachate treatment and resource conversion process is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of environmental protection and resource utilization, and specifically relates to a low-carbon intelligent treatment and resource utilization method of landfill leachate based on "end-edge-cloud" collaborative control. Background Art

[0002] The content of suspended solids, organic matter, nutrients, salts, heavy metals, etc. in landfill leachate is very high, and it is difficult to treat. The main treatment methods currently used are: physicochemical pretreatment-biochemical treatment (anaerobic-aerobic) and back-end deep treatment processes. Among them, physicochemical pretreatment usually adopts coagulation / flocculation method. Commonly used coagulants / flocculants such as polyaluminium chloride (PAC), polyferric sulfate (PFS) and polyacrylamide (PAM) have limited effects on leachate pretreatment. The dosage of the agent is large, the sludge output is large, the sedimentation process is slow, and the agent addition ratio cannot be adjusted in time according to the changes in the raw water quality, thereby affecting the activity of microorganisms in the biochemical system and the service life of the back-end membrane treatment system.

[0003] In addition, the concentration of organic pollutants, ammonia nitrogen and other pollutants in the leachate is high, the biodegradability is poor, the aeration energy consumption is high, the system operation automation is relatively low, the labor volume is large, and there is a lag in water quality monitoring and feedback control. In actual application, the expected treatment effect cannot be achieved. Therefore, using the sensors integrated in the leachate treatment equipment to collect various data such as production equipment, process, operation and maintenance, and to perform intelligent algorithm optimization and process parameter regulation is of great significance to realize the intelligent operation and maintenance of garbage leachate sewage treatment facilities and improve the operation level and supervision capabilities. Summary of the invention

[0004] The purpose of the present invention is to address the shortcomings of the prior art and provide a low-carbon intelligent treatment and resource utilization method for landfill leachate based on "end-edge-cloud" collaborative control.

[0005] A low-carbon intelligent treatment and resource utilization method for landfill leachate based on "end-edge-cloud" collaborative control, based on a landfill leachate treatment system; the landfill leachate treatment system includes a regulating tank, a flocculation pretreatment module, a biochemical system, and a membrane bioreactor that are cascaded in sequence; The method comprises the following steps: Step S1: Acquisition of real-time monitoring data from sensors at the terminal layer.

[0006] The real-time monitoring data of each sensor include the operating parameters of the regulating tank, the operating parameters of the flocculation pretreatment module, the operating parameters of the anaerobic reactor, the operating parameters of the anaerobic tank and the aerobic tank in the anaerobic-aerobic biochemical system, and the operating parameters of the membrane bioreactor.

[0007] Step S2: The real-time monitoring data of each sensor is sent to the edge layer through the communication network, and a machine learning algorithm is used at the edge layer to perform multi-objective prediction on the effluent water quality indicators, the system comprehensive energy consumption, and the anaerobic hydrogen production rate of the anaerobic reactor.

[0008] Step S3: Construct an objective function based on the effluent water quality index, the system comprehensive energy consumption, and the anaerobic hydrogen production rate of the anaerobic reactor. Under the constraints, use the decision-level optimization model in the cloud server to obtain the ideal values ​​of the operating parameters of the regulating tank, flocculation pretreatment module, anaerobic reactor, anaerobic-aerobic biochemical system, and membrane bioreactor.

[0009] Step S4: Based on the optimization results of the decision-level optimization model and the real-time monitoring data of the terminal layer sensors, the key process operating conditions of the system are adjusted to obtain the optimal results of effluent water quality indicators, anaerobic hydrogen production rate, and system comprehensive energy consumption.

[0010] Preferably, in step S1, the operating parameters of the regulating tank include pH value, water temperature, dissolved oxygen (DO), oxidation-reduction potential (ORP), and influent water quality indicators, wherein the influent water quality indicators include influent flow rate, influent chemical oxygen demand (COD), ammonia nitrogen (NH 4 + -N), total nitrogen (TN).

[0011] The operating parameters of the flocculation pretreatment module include flocculant dosage, sedimentation residence time, pH value, ORP, and water temperature.

[0012] The operating parameters of the anaerobic reactor include the hydraulic retention time of the anaerobic fermentation tank, the dosage of the external redox mediator, the pH value of the anaerobic tank, the water temperature, the ORP, the flow rate, temperature, total pressure and partial pressure of different biogas produced.

[0013] The operating parameters of the anaerobic tank and the aerobic tank in the anaerobic-aerobic biochemical system include: DO, ORP, water temperature, and pH value.

[0014] The operating parameters of the membrane bioreactor include pH, ORP, water temperature, DO, and transmembrane pressure difference.

[0015] The water quality indicators of the effluent after the membrane bioreactor treatment include turbidity, effluent COD, NH 4 + -N, TN.

[0016] Preferably, in step S2, the machine learning algorithm includes at least one of random forest (RF), extreme gradient boosting (XGBoost), long short-term memory network (LSTM), and artificial neural network (ANN).

[0017] Preferably, in step S3, the decision-level optimization model is specifically one of a genetic algorithm (GA) and a particle swarm optimization algorithm (PSO).

[0018] Preferably, in step S3, the objective function is: in, is the weight; Indicates that the overall energy consumption of the system is minimized. Indicates the target value of water quality discharge. Indicates that the anaerobic hydrogen production rate of the anaerobic reactor is the highest; E(x) represents the comprehensive energy consumption of the system under certain operating parameters; α 1 , α 2 , α 3 The water quality indicators of each outlet water are COD, NH 4 + -N and TN are the weights of importance to the optimization results of the emission target value; COD( ) represents the relationship between the effluent COD value and the discharge standard value under certain operating parameters; NH 4 + -N( ) represents the outlet NH of the system under certain operating parameters 4 + -The relationship between N value and emission standard value; TN( ) represents the relationship between the outlet TN value of the system under certain operating parameters and the discharge standard value; It indicates the hydrogen production rate in the anaerobic reactor under certain operating parameters of the system.

[0019] The constraints are specifically: In the formula, E 0 It is the energy consumption of process equipment before the system performs “end-edge-cloud” collaborative control; The actual energy consumption of process equipment after the system performs “end-edge-cloud” collaborative control; E A is the energy consumption of the system during data collection, transmission, calculation, feedback, and control when performing “end-edge-cloud” collaborative control; C is the actual COD value of the system effluent, C （t） is the standard value of effluent COD discharge; N is the actual NH 4 + -N concentration, N （t） For effluent NH 4+ -N emission standard value; T is the actual TN concentration of the effluent, T （t） is the standard value of TN discharge for effluent; is the Sigmoid function; H 2 is the actual hydrogen production rate of the system, H 2（opt） is the maximum hydrogen production rate; Tanh is the hyperbolic tangent function; S 1 , S 2 is the parameter used to scale the data.

[0020] Preferably, in step S4, the key process operating conditions of the system include: key process operating conditions of the flocculation pretreatment module, key process operating conditions of the anaerobic-aerobic biochemical system, and key process operating conditions of the membrane bioreactor system.

[0021] The key process operating conditions of the flocculation pretreatment module include the stirring speed of the stirrer and the rotation speed of the high-speed shearing machine.

[0022] The key process operating conditions of the anaerobic-aerobic biochemical system include aeration intensity, stirring speed of the mixer, and sludge return volume.

[0023] The key process operating conditions of the membrane bioreactor system include the working pressure of the water production pump and the aeration intensity.

[0024] The beneficial effects of the present invention are: The low-carbon intelligent treatment and resource utilization method of landfill leachate with "end-edge-cloud" collaborative control of the present invention realizes the interconnection of leachate treatment and resource / energy process units through the "end-edge-cloud" IoT sensing network; by optimizing and controlling the operating parameters of the biochemical system, the intelligent management and control of the overall process of landfill leachate treatment and resource / energy is realized. The process integrates technologies such as chemical coagulation / magnetic flocculation efficient pretreatment, anaerobic fermentation hydrogen / acid production, and membrane bioreactor based on modular ultrafiltration membranes, which can shorten the overall residence time of the system, reduce the load of the biochemical system, reduce the floor space and system energy consumption, and can generate new energy hydrogen, improve the ability of organic matter in wastewater to ferment and produce volatile fatty acids, and can be used to provide biological efficient nitrogen and phosphorus removal carbon sources, realize the efficient removal of particulate matter, organic matter, and nutrients in the leachate treatment process, and realize the resource, energy and intelligent leachate treatment.

[0025] Under the "cloud-edge-end" intelligent control architecture, the real-time monitoring data of each sensor at the terminal layer of each link of the landfill leachate treatment is first sent to the edge layer through a dedicated network or operator network, and the intelligent algorithm deployed at the edge layer is used to analyze the system's comprehensive energy consumption, effluent water quality (COD, NH 4 +-N, TN), and fermentation hydrogen production rate are predicted in real time. The monitoring data is uploaded to the cloud through the edge, explored in the multi-dimensional parameter space, the objective function value is evaluated, and gradually converged to the optimal solution. Combined with methods such as gradient descent and genetic algorithm, the global optimal point of key influencing factors is accurately located. And the optimal result is returned to the edge. The edge forms corresponding operation instructions and transmits them back to the execution equipment of the key process parameters of the control system to perform specific operations, so as to realize the dynamic adjustment of key influencing factors in the process of landfill leachate treatment and resource utilization, so as to achieve the highest treatment efficiency and hydrogen production rate, and the lowest overall comprehensive energy consumption of the system, and achieve the goal of "energy conservation and emission reduction". BRIEF DESCRIPTION OF THE DRAWINGS

[0026] Figure 1 Schematic diagram of terminal layer sensor deployment in the landfill leachate treatment system.

[0027] Figure 2 Example 1 is a monitoring point map of a landfill leachate treatment system in an incineration power plant.

[0028] Figure 3 Example 2 is a monitoring point map of a leachate sewage treatment station at a domestic waste landfill.

[0029] Figure 4 The model topology of the machine learning algorithm. DETAILED DESCRIPTION

[0030] The present invention is further described below in conjunction with specific embodiments, but the protection scope of the present invention is not limited thereto.

[0031] An embodiment of the present invention provides a low-carbon intelligent treatment and resource utilization method for leachate based on "end-edge-cloud" collaborative control, which is implemented based on a leachate treatment system; the leachate treatment system includes a regulating tank, a flocculation pretreatment module, a biochemical system, and a membrane bioreactor (MBR) cascaded in sequence.

[0032] The regulating tank is responsible for receiving the landfill leachate and regulating the water quality uniformly.

[0033] The flocculation pretreatment module is responsible for the efficient and rapid removal of particulate matter in the influent water; preferably, coagulation / flocculation physicochemical pretreatment or magnetic flocculation pretreatment.

[0034] The biochemical system includes anaerobic reactors connected in cascade, and single-stage or multi-stage anaerobic-aerobic biochemical systems connected in series.

[0035] The anaerobic reactor is responsible for utilizing the metabolism of anaerobic microorganisms to convert organic matter in the wastewater into hydrogen, thereby achieving degradation and stabilization of the organic matter, and is preferably an anaerobic IC reactor or a USAB reactor.

[0036] The anaerobic-aerobic biochemical system comprises an anaerobic tank (A tank), an aerobic aeration tank (O tank), and a clarifier tank which are cascaded in sequence. A portion of the effluent from the aerobic aeration tank (O tank) flows back to the anaerobic tank (A tank); the anaerobic tank and the aerobic aeration tank are responsible for carrying out anaerobic biochemical reactions and aerobic biochemical reactions on the anaerobic fermentation products in sequence to complete preliminary denitrification; the clarifier tank is responsible for carrying out mud-water separation on the effluent from the aerobic aeration tank (O tank).

[0037] The membrane bioreactor (MBR) is responsible for high-efficiency solid-liquid separation of anaerobic-aerobic biochemical products, further enhancing the efficiency of the biochemical reaction.

[0038] The "end-edge-cloud" collaboratively controlled low-carbon intelligent treatment and resource utilization method for landfill leachate includes the following steps: Step S1: Acquisition of real-time monitoring data from sensors at the terminal layer.

[0039] The real-time monitoring data of each sensor include the operating parameters of the regulating tank, the operating parameters of the flocculation pretreatment module, the operating parameters of the anaerobic reactor, the operating parameters of the anaerobic tank (A tank) and the aerobic tank (O tank) in the anaerobic-aerobic biochemical system, and the operating parameters of the membrane bioreactor (MBR).

[0040] The operating parameters of the regulating pool mainly include pH value, water temperature, dissolved oxygen (DO), oxidation-reduction potential (ORP), and influent water quality indicators, including influent flow rate, influent chemical oxygen demand (COD), ammonia nitrogen (NH 4 + -N), total nitrogen (TN).

[0041] The operating parameters of the flocculation pretreatment module include flocculant dosage, sedimentation residence time, pH value, ORP, water temperature, etc.

[0042] The operating parameters of the anaerobic reactor include the hydraulic retention time of the anaerobic fermentation tank, the dosage of the external redox mediator, the pH value of the anaerobic tank, the water temperature, the ORP, the flow rate, temperature, total pressure and partial pressure of different biogas produced; the biogas includes H 2 , CH 4 , CO 2 .

[0043] The operating parameters of the anaerobic pool (A pool) and the aerobic pool (O pool) in the anaerobic-aerobic biochemical system mainly include: DO, ORP, water temperature, and pH value.

[0044] The operating parameters of the membrane bioreactor (MBR) include pH, ORP, water temperature, DO, and transmembrane pressure difference.

[0045] The effluent quality indicators after the membrane bioreactor (MBR) treatment include turbidity, COD, NH4 + -N, TN.

[0046] The deployment of the terminal layer sensor in step S1 is detailed in Figure 1-Figure 3 , as follows: (1) A flow sensor is installed at the water inlet of the regulating tank to mainly monitor the inlet flow of the landfill leachate; a first pH sensor, a first temperature sensor, a first water temperature sensor, a first ORP sensor, a first COD measurement sensor, and a first NH 4 + -N measurement sensor, first TN measurement sensor, first turbidity measurement sensor.

[0047] (2) A second pH sensor and a second water temperature sensor are disposed at the water outlet of the flocculation pretreatment module or inside the module.

[0048] (3) A gas phase component sensor and a state sensor are set at the gas outlet of the anaerobic reactor. The gas phase component sensor is used to detect H 2 , CH 4 , CO 2 The state sensor includes a gas flow sensor, a gas temperature sensor, and a pressure sensor; a third pH sensor, a third water temperature sensor, and a third ORP sensor are provided at the water outlet of the anaerobic reactor.

[0049] (5) A fourth pH sensor, a fourth water temperature sensor, a fourth ORP sensor, and a first DO sensor are provided in both the A pool and the O pool of the anaerobic-aerobic biochemical system.

[0050] (6) The fifth pH sensor, the fifth water temperature sensor, and the fifth ORP sensor are installed in the clarification tank of the anaerobic-aerobic biochemical system to monitor the pH, ORP, and water temperature of the supernatant on-line.

[0051] (7) The sixth pH sensor, the sixth water temperature sensor, the sixth ORP sensor, the second DO sensor, the transmembrane pressure difference and other sensors are installed in the membrane bioreactor (MBR) to monitor the operating status of the MBR pool in real time; the seventh pH sensor, the seventh water temperature sensor, the seventh ORP sensor, the second COD measurement sensor, the second NH 4 + -N measurement sensor, second TN measurement sensor, second turbidity measurement sensor.

[0052] Step S2: The real-time monitoring data of each sensor in S1 is sent to the edge layer through the communication network, and a machine learning algorithm is used at the edge layer to perform multi-objective prediction on the effluent water quality indicators, the system comprehensive energy consumption, and the anaerobic hydrogen production rate of the anaerobic reactor.

[0053] In step S2, the machine learning algorithm includes one or more of random forest (RF), extreme gradient boosting (XGBoost), long short-term memory network (LSTM), artificial neural network (ANN), etc.; the model topology is as follows Figure 4 shown.

[0054] Step S3: Construct an objective function based on effluent water quality indicators, anaerobic hydrogen production rate, and system comprehensive energy consumption. Under constraints, use the decision-level optimization model in the cloud server to obtain the ideal values ​​of operating parameters of the regulating tank, flocculation pretreatment module, anaerobic reactor, anaerobic-aerobic biochemical system, and MBR system.

[0055] In step S3, the decision-level optimization model is specifically one of a genetic algorithm (GA) and a particle swarm optimization algorithm (PSO).

[0056] In step S3, the objective function is that the effluent water quality index meets the discharge standard requirements, the system comprehensive energy consumption reaches the minimum, and the anaerobic hydrogen production rate reaches the maximum. The objective function is expressed as follows:

[0057] in, is the weight of each goal, reflecting the minimization of the system's comprehensive energy consumption ( )、Water quality discharge target value ( ) and the highest hydrogen production rate ( ) between the two.

[0058] Constraints: =1 The system comprehensive energy consumption minimization expression is: In the formula, E(x) represents the comprehensive energy consumption of the system under certain operating parameters; It means minimizing the comprehensive energy consumption during system operation, which can be achieved by adjusting the operating frequency, start and stop time of the equipment.

[0059] Constraint: E(x)≤E 0 ; E(x)=E 0 '+E A In the formula, E 0 E is the energy consumption of process equipment before the system performs “end-edge-cloud” collaborative control; 0 ' is the actual energy consumption of process equipment after the system performs "end-edge-cloud" collaborative control; E AThe energy consumption of each unit module in the "end-edge-cloud" collaborative control system during data collection, transmission, calculation, feedback, and control; The effluent water quality meets the discharge target value Optimize the expression: In the formula, It is used to optimize the target values ​​of various parameters of effluent water quality. The target optimized water quality indicators are COD, NH 4 + -N and TN; among them, α 1 , α 2 , α 3 It is the weight of each water quality index’s importance to the optimization result of the target discharge value. COD( ) represents the relationship between the effluent COD value and the discharge standard value under certain operating parameters; NH 4 + -N( ) represents the outlet NH of the system under certain operating parameters 4 + -The relationship between N value and emission standard value; TN( ) represents the relationship between the outlet TN value and the discharge standard value of the system under certain operating parameters.

[0060] Constraints on effluent quality: Where, C is the actual COD value of the system effluent, C （t） is the standard value of effluent COD discharge (0≤C （t） ≤ 500 mg / L); N is the actual NH 4 + -N concentration, N （t） For effluent NH 4 + -N emission standard value (0≤N （t） ≤ 25 mg / L); T is the actual TN concentration of the effluent, T （t） is the standard value of TN discharge of effluent (0≤T （t） ≤ 70 mg / L); is the Sigmoid function, which is used to optimize the water quality target value so that each water quality index tends to the target value; S 1 is the parameter used to scale the data so that the scales of different variables are comparable.

[0061] The anaerobic hydrogen production rate is maximized expression: In the formula, Indicates that the anaerobic hydrogen production rate is maximized; It indicates the hydrogen production rate in the anaerobic reactor under certain operating parameters of the system.

[0062] Constraints: H 2 is the actual hydrogen production rate of the system, H 2（opt） is the maximum hydrogen production rate; Tanh is the hyperbolic tangent function, which is used to control the hydrogen production rate to make it close to the maximum value; S 2 is the parameter used to scale the data.

[0063] Step S4: Based on the optimization results of the decision-level optimization model (i.e., the ideal values ​​of the operating parameters of the equalization tank, flocculation pretreatment module, anaerobic reactor, anaerobic-aerobic biochemical system, and MBR system) and the real-time monitoring data of the terminal layer sensors, the key process operating conditions of the system are adjusted to obtain the optimal results of the effluent water quality indicators, anaerobic hydrogen production rate, and comprehensive energy consumption of the system.

[0064] The key process operating conditions of the system include: 1. Key process operating conditions of flocculation pretreatment module Working frequency and stirring speed of the mixer: The mixer speed can be adjusted by changing the working frequency of the mixer inverter (20HZ~50HZ). The mixer speed range is 20~600 rpm.

[0065] High-speed shearing machine: When the flocculation unit selects the magnetic flocculation technology, the speed of the high-speed shearing machine is adjusted by changing the operating frequency of the inverter (30HZ~50HZ). The speed range of the high-speed shearing machine is 2000~2000 rpm.

[0066] 2. Key process operating conditions of anaerobic-aerobic biochemical system Aeration intensity: In the aerobic pool (O pool), adjust the inverter operating frequency (25HZ~50HZ) and start and stop time, and adjust the air volume and aeration time of the aeration equipment according to the real-time monitoring, prediction and optimization results of DO.

[0067] Stirring speed: In the anoxic pool (Pool A), the stirrer speed is adjusted by changing the operating frequency of the stirrer inverter (20HZ~45HZ), and the stirrer speed range is 40~1000 rpm.

[0068] Sludge return volume: By adjusting the motor speed of the reflux pump (30HZ~50HZ), the sludge return volume and internal reflux volume can be adjusted to maintain appropriate sludge concentration and denitrification efficiency in the system.

[0069] 3. Key process operating conditions of membrane bioreactor (MBR) system Working pressure of water production pump: Based on the real-time monitoring, prediction and optimization results of trans-membrane pressure difference, the working pressure of the water production pump is adjusted, and regular backwashing or chemical cleaning is performed to ensure the stability of membrane flux and water production of the MBR system.

[0070] Aeration intensity: In the MBR pool, according to the real-time monitoring, prediction and optimization results of DO, adjust the inverter operating frequency (35HZ~50HZ) and start and stop time, and adjust the air volume and aeration time of the aeration equipment.

[0071] Example 1 The sewage treatment station of a waste incineration power plant adopts the treatment process of regulating tank-chemical coagulation-anaerobic fermentation-anaerobic (A) tank-O tank-MBR-tube discharge. The "cloud-edge-end" collaborative control method of the present invention is used to intelligently control all aspects of sewage treatment in the whole plant. The terminal layer sensors of each treatment link are as follows Figure 2 As shown:

[0072] (1) Water inlet 1#: Set up a flow sensor to monitor the water inlet flow in real time.

[0073] (2) Regulating tank 2#: Set up online sensors for pH, ORP, water temperature, turbidity, etc., and automatically collect water samples and pump them to online monitoring equipment for COD and ammonia nitrogen (NH 4 + -N), and total nitrogen (TN).

[0074] (3) After chemical coagulation 3#: Set up online sensors such as pH and water temperature above the clarification tank.

[0075] (4) Anaerobic IC reactors 4# and 5#: Set H 2 / CH 4 / CO 2 Gas phase component sensors, such as temperature / flow / pressure, and other state sensors are used to monitor the partial pressure of the generated biogas components in real time; operating state sensors such as pH, ORP, and water temperature are set in the water phase.

[0076] (5) Anaerobic-aerobic biochemical system (A / O system) #6: pH, DO, ORP, water temperature and other sensors are installed in both the A pool and the O pool to monitor the operating status of the biochemical system in real time.

[0077] (6) A / O system effluent #7: Pump the effluent from the A / O biochemical pool to a small clarification tank outside the system for mud and water separation, and set sensors such as pH, water temperature, and ORP in the supernatant.

[0078] (7) MBR pool 8#: Set up sensors such as pH, ORP, DO, water temperature, water level, and transmembrane pressure difference in the MBR pool to monitor the operating status of the MBR pool in real time.

[0079] (8) MBR effluent 9#: Set up sensors such as pH, ORP, and water temperature in the MBR effluent pool, and pump out water to the online monitoring equipment for COD and NH 4 + -Online monitoring of N, TN and turbidity.

[0080] Subsequently, the historical data of various sensors at the terminal layer were preprocessed and converted into a robust data set through feature engineering. Secondly, the training and optimization of algorithms such as random forest (RF) and extreme gradient boosting (XGBoost) were carried out on the constructed robust data set. At the edge layer, the system comprehensive energy consumption and effluent water quality (COD, NH 4 + -N, TN), fermentation hydrogen production rate, and characteristic importance analysis were performed to determine the key influencing factors affecting the effluent quality of the incineration plant leachate, anaerobic hydrogen production rate, and comprehensive energy consumption of the system.

[0081] Taking the optimized XGBoost algorithm as the core and combining it with the particle swarm optimization algorithm (PSO), a decision-level optimization model (XGBoost-PSO model) of multiple objective functions such as effluent water quality, anaerobic hydrogen production rate, and system comprehensive energy consumption was established at the cloud platform layer. Reverse engineering design was carried out under the overall optimal conditions of the system (lowest comprehensive energy consumption, water quality meeting emission standards, and high hydrogen production rate).

[0082] The weight of the system comprehensive energy consumption on the overall optimization result of the objective function ω 1 =0.32.

[0083] Outlet water quality (COD, NH 4 + -N, TN) weight ω on the overall optimization result of the objective function 2 =0.41; where the weight α of COD on the optimization result of the target emission value is 1 =0.55, NH 4 + -N is the weight of the importance of the optimization result of the emission target value to meet the standard α 2 =0.28, the weight of TN on the optimization result of the target emission value 3 =0.17.

[0084] The weight of anaerobic hydrogen production rate on the overall optimization result of the objective function ω 3 =0.27.

[0085] Under the above multi-objective optimization conditions, based on the real-time monitoring, prediction and optimization data of the terminal layer, the stirring speed of the stirring motor of the chemical flocculation pretreatment module, the air volume and aeration time of the aeration fan of the aerobic tank (O tank) of the anaerobic-aerobic biochemical system, the speed of the mixer of the anoxic tank (O tank), the flow rate of the sludge return pump, the air volume and aeration time of the aeration equipment of the MBR system, and the working pressure of the water production pump are regulated to achieve online adaptive adjustment of key parameters.

[0086] The leachate treatment efficiency and resource utilization level of the waste incineration plant based on the above-mentioned "cloud-edge-end" collaborative control method are shown in Appendix 1.

[0087] Appendix 1. Treatment efficiency and resource utilization level of leachate from waste incineration plants under “cloud-edge-end” collaborative control Example 2 The leachate sewage treatment station of a domestic waste landfill plant adopts a treatment process of regulating tank-magnetic coagulation-UASB-two-stage anaerobic / aerobic (A / O)-MBR-nanofiltration-nanotube discharge. The "cloud-edge-end" collaborative control method of the present invention is used to intelligently control all aspects of the sewage treatment in the whole plant. The terminal layer sensors of each treatment link are as follows Figure 3 As shown:

[0088] (1) Water inlet 1#: Set up a flow sensor to monitor the water inlet flow in real time.

[0089] (2) Regulating tank 2#: Set up online sensors for pH, ORP, water temperature, turbidity, etc., and automatically collect water samples and pump them to online monitoring equipment for COD and ammonia nitrogen (NH 4 + -N), and total nitrogen (TN).

[0090] (3) After magnetic coagulation 3#: Set up online sensors such as pH and water temperature above the clarification tank.

[0091] (4) Anaerobic UASB reactors 4# and 5#: Set H 2 / CH 4 / CO 2 Gas phase component sensors, such as temperature / flow / pressure, and other state sensors are used to monitor the partial pressure of the generated biogas components in real time; operating state sensors such as pH, ORP, and water temperature are set in the water phase.

[0092] (5) Two-stage A / O system 6#: pH, DO, ORP, water temperature and other sensors are installed in the A pool and O pool of the first and second stages to monitor the operating status of the biochemical system in real time.

[0093] (6) A / O system effluent #7: Pump the effluent from the A / O biochemical pool to a small clarification tank outside the system for mud and water separation, and set sensors such as pH, water temperature, and ORP in the supernatant.

[0094] (7) MBR pool 8#: Set up sensors such as pH, DO, ORP, water temperature, water level, and membrane filtration pressure in the MBR pool to monitor the operating status of the MBR pool in real time.

[0095] (8) MBR effluent 9#: Set up sensors for pH, water temperature, turbidity, etc. in the MBR effluent pool, and pump out water to the online monitoring equipment for COD and NH 4 + -Online monitoring of N and TN.

[0096] Subsequently, the historical data of various sensors at the terminal layer were preprocessed and converted into a robust data set through feature engineering. Secondly, the artificial neural network (ANN), long short-term memory network (LSTM) and other algorithms were trained and tuned on the constructed robust data set. The system comprehensive energy consumption and effluent water quality (COD, NH 4 + -N, TN), fermentation hydrogen production rate was predicted, and characteristic importance analysis was performed to determine the key influencing factors affecting the comprehensive energy consumption of the landfill infiltration system, the water quality of the filtrate effluent, and the anaerobic hydrogen production rate.

[0097] With the optimized LSTM algorithm as the core and combined with the genetic algorithm (GA), a decision-level optimization model (XGBoost-GA model) of multiple objective functions such as effluent water quality, anaerobic hydrogen production rate, and system comprehensive energy consumption was established at the cloud platform layer, and reverse engineering design was carried out under the overall optimal conditions of the system (lowest comprehensive energy consumption, water quality meeting emission standards, and high hydrogen production rate).

[0098] The weight of the system comprehensive energy consumption on the overall optimization result of the objective function ω 1 =0.25.

[0099] Outlet water quality (COD, NH 4 + -N, TN) weight ω on the overall optimization result of the objective function 2 =0.45; where α is the weight of COD on the optimization result of the target emission value 1 =0.46, NH 4 + -N is the weight of the importance of the optimization result of the emission target value to meet the standard α 2 =0.30, the weight of TN on the optimization result of the target emission value 3 =0.24.

[0100] The weight of anaerobic hydrogen production rate on the overall optimization result of the objective function ω 3 =0.30.

[0101] Under the above multi-objective optimization conditions, based on the real-time monitoring, prediction and optimization data of the terminal layer, the stirring speed of the stirring motor of the magnetic flocculation pretreatment module, the speed of the high-speed shear machine, the air volume and aeration time of the aeration fan of the aerobic tank (O tank) of the anaerobic-aerobic biochemical system, the speed of the mixer of the anoxic tank (O tank), the flow rate of the sludge return pump, as well as the air volume and aeration time of the aeration equipment of the MBR system, the working pressure of the water production pump, etc. are regulated to achieve online adaptive adjustment of key parameters.

[0102] Appendix 2: Landfill leachate treatment efficiency and resource utilization level under “cloud-edge-end” collaborative control The above embodiments are not limitations of the present invention, and the present invention is not limited to the above embodiments. As long as the requirements of the present invention are met, they belong to the protection scope of the present invention.

Claims

1. A low-carbon intelligent treatment and resource utilization method for landfill leachate based on "end-edge-cloud" collaborative control, based on a landfill leachate treatment system; the landfill leachate treatment system includes a regulating tank, a flocculation pretreatment module, a biochemical system, and a membrane bioreactor that are cascaded in sequence; The regulating tank is responsible for receiving the garbage leachate and regulating the water quality uniformly; The flocculation pretreatment module is responsible for the efficient and rapid removal of particulate matter in the influent water; preferably, coagulation / flocculation physicochemical pretreatment or magnetic flocculation pretreatment; The biochemical system includes anaerobic reactors cascaded in sequence, and single-stage or multi-stage anaerobic-aerobic biochemical systems in series; The anaerobic-aerobic biochemical system comprises an anaerobic tank, an aerobic aeration tank and a clarifier tank which are cascaded in sequence. A portion of the effluent from the aerobic aeration tank flows back to the anaerobic tank. The anaerobic tank and the aerobic aeration tank are responsible for carrying out anaerobic biochemical reactions and aerobic biochemical reactions on the anaerobic fermentation products in sequence to complete the preliminary denitrification and removal of organic matter. The clarifier tank is responsible for carrying out mud-water separation on the effluent from the aerobic aeration tank. The membrane bioreactor is responsible for efficient solid-liquid separation of anaerobic-aerobic biochemical products, further enhancing the efficiency of biochemical reactions; It is characterized in that The method comprises the following steps: Step S1: Acquisition of real-time monitoring data of each sensor at the terminal layer; The real-time monitoring data of each sensor include the operating parameters of the regulating tank, the operating parameters of the flocculation pretreatment module, the operating parameters of the anaerobic reactor, the operating parameters of the anaerobic tank and the aerobic tank in the anaerobic-aerobic biochemical system, and the operating parameters of the membrane bioreactor; Step S2: sending the real-time monitoring data of each sensor to the edge layer through the communication network, and using a machine learning algorithm at the edge layer to perform multi-objective prediction on the effluent water quality index, the system comprehensive energy consumption, and the anaerobic hydrogen production rate of the anaerobic reactor; Step S3: construct an objective function based on the effluent water quality index, the system comprehensive energy consumption, and the anaerobic hydrogen production rate of the anaerobic reactor. Under the constraint conditions, the decision-level optimization model in the cloud server is used to obtain the ideal values ​​of the operating parameters of the regulating tank, the flocculation pretreatment module, the anaerobic reactor, the anaerobic-aerobic biochemical system, and the membrane bioreactor; Step S4: Based on the optimization results of the decision-level optimization model and the real-time monitoring data of the terminal layer sensors, the key process operating conditions of the system are adjusted to obtain the optimal results of effluent water quality indicators, anaerobic hydrogen production rate, and system comprehensive energy consumption.

2. The method according to claim 1, characterized in that In step S1, the operating parameters of the regulating tank include pH value, water temperature, dissolved oxygen, ORP, and influent water quality indicators, and the influent water quality indicators include influent flow rate, influent COD, ammonia nitrogen, and total nitrogen; The operating parameters of the flocculation pretreatment module include flocculant dosage, sedimentation residence time, pH value, ORP, and water temperature; The operating parameters of the anaerobic reactor include the hydraulic retention time of the anaerobic fermentation tank, the dosage of the added redox mediator, the pH value of the anaerobic tank, the water temperature, the ORP, the flow rate, temperature, total pressure of the generated biogas, and the partial pressures of different biogases; The operating parameters of the anaerobic tank and the aerobic tank in the anaerobic-aerobic biochemical system include: dissolved oxygen, ORP, water temperature, and pH value; The operating parameters of the membrane bioreactor include pH, ORP, water temperature, dissolved oxygen, and transmembrane pressure difference; The water quality indicators of the effluent after the membrane bioreactor treatment include turbidity, effluent COD, ammonia nitrogen and total nitrogen.

3. The method according to claim 1, characterized in that In step S2, the machine learning algorithm includes at least one of random forest, extreme gradient boosting, long short-term memory network, and artificial neural network.

4. The method according to claim 1, characterized in that In step S3, the decision-level optimization model is specifically one of a genetic algorithm and a particle swarm optimization algorithm.

5. The method according to claim 1, characterized in that In step S3, the objective function is specifically: in, is the weight; Indicates that the overall energy consumption of the system is minimized. Indicates the target value of water quality discharge. represents the maximization of the anaerobic hydrogen production rate of the anaerobic reactor; E(x) represents the comprehensive energy consumption of the system under certain operating parameters; α1, α2, α3 are the weights of the importance of each effluent water quality index COD, ammonia nitrogen and total nitrogen to the optimization result of the target emission value; COD( ) represents the relationship between the effluent COD value and the discharge standard value under certain operating parameters; NH4 + -N( ) represents the relationship between the effluent ammonia nitrogen value and the discharge standard value under certain operating parameters; TN( ) represents the relationship between the total nitrogen value of the effluent and the discharge standard value under certain operating parameters of the system; It indicates the hydrogen production rate in the anaerobic reactor under certain operating parameters of the system.

6. The method according to claim 5, characterized in that In step S3, the constraint conditions are specifically: In the formula, E0 is the energy consumption of process equipment before the system performs "end-edge-cloud" collaborative control; The actual energy consumption of process equipment after the system performs "end-edge-cloud" collaborative control; E A is the energy consumption of the system during data collection, transmission, calculation, feedback, and control when performing "end-edge-cloud" collaborative control; C is the actual COD value of the system effluent, C （t） is the standard value of effluent COD discharge; N is the actual ammonia nitrogen concentration in effluent, N （t） is the standard value of effluent ammonia nitrogen emission; T is the actual total nitrogen concentration in effluent, T （t） is the standard value of total nitrogen discharge in effluent; is the Sigmoid function; H2 is the actual hydrogen production rate of the system, H 2（opt） is the maximum hydrogen production rate; Tanh is the hyperbolic tangent function; S1 and S2 are parameters used to scale the data.

7. The method according to claim 1, characterized in that In step S4, the key process operating conditions of the system include: key process operating conditions of the flocculation pretreatment module, key process operating conditions of the anaerobic-aerobic biochemical system, and key process operating conditions of the membrane bioreactor system.

8. The method according to claim 7, characterized in that In step S4, the key process operating conditions of the flocculation pretreatment module include the stirring speed of the stirrer and the rotation speed of the high-speed shearing machine.

9. The method according to claim 7, characterized in that In step S4, the key process operating conditions of the anaerobic-aerobic biochemical system include aeration intensity, stirring speed of the mixer, and sludge return volume.

10. The method according to claim 7, characterized in that In step S4, the key process operating conditions of the membrane bioreactor system include the operating pressure of the water production pump and the aeration intensity.

Images