A soy product wastewater treatment process optimization method and system based on digital twinning

Abstract

The application discloses a kind of based on digital twinning soy products wastewater treatment process optimization method and system, specifically related to soy products processing wastewater biological denitrification processing technical field, for solving the complex influence of existing carbon source dosing control method only optimization denitrification efficiency and neglecting its to microbial community, leading to secondary process problems such as foam and floating mud;By obtaining real-time operation data and inputting to the digital twin model containing water quality conversion and microbial community succession submodel, based on model simulation for different candidate carbon source dosing strategy simultaneously predicts denitrification efficiency value, denitrification foam formation risk value and interspecific competition imbalance risk value, and then from the strategy that meets the requirement of denitrification selects the optimization strategy of biological risk comprehensive optimization to execute carbon source dosing;Realize the collaborative optimization of denitrification performance and system biological stability, effectively inhibit the generation of foam and flora imbalance, guarantee that effluent water quality is stable and up to standard.

Description

Technical Field

[0001] This invention relates to the field of biological denitrification treatment technology for soybean product processing wastewater, and more specifically, to a method and system for optimizing soybean product wastewater treatment process based on digital twins. Background Technology

[0002] Wastewater from soybean product processing contains high concentrations of organic matter and ammonia nitrogen. It is often treated using biological denitrification processes such as anaerobic-anoxic-aerobic processes. To meet stringent total nitrogen discharge requirements, it is common practice to add external carbon sources to maintain the carbon-to-nitrogen ratio needed for denitrification. Currently, the control of carbon source addition is mainly based on real-time monitoring or empirical setting of limited parameters such as influent ammonia nitrogen and nitrate nitrogen, aiming to achieve efficient nitrogen removal with the lowest possible carbon source consumption.

[0003] However, the above control methods only focus on the stoichiometric efficiency of denitrification as the core optimization target, ignoring the complex impact of carbon source addition strategies on the metabolic behavior of microbial communities. In the scenario of soybean product wastewater treatment, dynamic addition of carbon sources can easily trigger excessive proliferation and metabolic transformation of denitrifying bacteria, leading to the secretion of a large amount of hydrophobic extracellular polymers and the formation of stubborn biochemical foam. This can cause sludge to float in the secondary sedimentation tank, severely damaging the solid-liquid separation effect, and in turn causing the concentration of suspended solids and total nitrogen in the effluent to exceed the standard. Existing technologies rely on the local control of a single water quality target and cannot synergistically optimize denitrification efficiency and system operation stability. In essence, this is due to the lack of understanding and coordination of the full-process, multi-dimensional biological response triggered by carbon source addition, which, while improving denitrification performance, causes secondary process problems that affect the overall treatment effect. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a method and system for optimizing the treatment process of soybean product wastewater based on digital twins to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for optimizing the treatment process of soybean product wastewater based on digital twins includes the following steps:

[0007] S1: Obtain real-time operating data of the soybean product wastewater treatment system;

[0008] S2: Input real-time operational data into the digital twin model;

[0009] S3: Based on real-time operating data, set several candidate carbon source addition strategies. Each candidate carbon source addition strategy includes the amount of candidate carbon source added and the acceleration rate mode of candidate carbon source addition. Use a digital twin model to simulate the denitrification efficiency value and the denitrification foam formation risk value under each strategy, and establish the mapping relationship between the candidate carbon source addition strategy and the denitrification efficiency value and the denitrification foam formation risk value.

[0010] S4: For each candidate carbon source addition strategy, based on the simulation process of the digital twin model, evaluate the risk value of interspecific competition imbalance corresponding to the changes in the internal distribution structure of the denitrifying bacteria caused by the candidate carbon source addition acceleration mode.

[0011] S5: Based on the mapping relationship and the risk value of interspecific competition imbalance, select the optimal carbon source addition strategy from the candidate carbon source addition strategies that meets the preset denitrification efficiency requirements and has the best comprehensive risk value of denitrification foam formation and interspecific competition imbalance.

[0012] S6: Perform carbon source addition operation according to the optimized carbon source addition strategy.

[0013] Furthermore, S1 includes:

[0014] Obtain real-time influent parameters at the inlet of the soybean product wastewater treatment system;

[0015] Obtain real-time status parameters of the biochemical reaction tank within the soybean product wastewater treatment system;

[0016] Obtain historical data on carbon source addition for the soybean product wastewater treatment system.

[0017] Furthermore, S2 includes:

[0018] Input the real-time influent parameters into the influent water quality parameter interface of the digital twin model;

[0019] Input real-time status parameters into the biochemical reaction process parameter interface of the digital twin model;

[0020] Input historical carbon source addition data into the carbon source addition record interface of the digital twin model.

[0021] Furthermore, digital twin models include:

[0022] A water quality transformation kinetics sub-model is used to simulate the ammonia nitrogen transformation, nitrification and denitrification reaction processes based on real-time influent parameters and historical carbon source addition data.

[0023] Microbial community succession model is used to simulate the growth, metabolism, and population structure dynamics of denitrifying functional microbial communities based on real-time state parameters;

[0024] The water quality transformation kinetics sub-model and the microbial community succession sub-model are coupled together. The microbial metabolic activity parameters output by the microbial community succession sub-model serve as the regulatory factors for the denitrification reaction rate in the water quality transformation kinetics sub-model.

[0025] Furthermore, S3 includes:

[0026] Based on real-time influent parameters and historical carbon source addition data, several candidate carbon source addition strategies are generated, including different candidate carbon source addition amounts and different candidate carbon source addition acceleration rate modes.

[0027] Each candidate carbon source addition strategy is input into the digital twin model, which then drives the digital twin model to simulate and execute the biochemical reaction process under the candidate carbon source addition strategy.

[0028] The simulated total nitrogen concentration of the effluent output by the digital twin model is obtained to calculate the denitrification efficiency value, and the simulated microbial metabolic indicators are obtained to assess the risk value of denitrification foam formation.

[0029] Each candidate carbon source addition strategy is associated with its corresponding denitrification efficiency value and denitrification foam formation risk value, forming a mapping relationship.

[0030] Furthermore, the microbial metabolic indicators on which the risk value of denitrification foam formation is assessed include the proportion of hydrophobic components and the secretion rate in the simulated extracellular polymeric material.

[0031] Furthermore, S4 includes:

[0032] For each candidate carbon source addition strategy, obtain the denitrifying functional microbial community simulation data output by the digital twin model during the simulation of the biochemical reaction under that candidate carbon source addition strategy;

[0033] Based on simulation data of denitrifying functional bacteria, we analyzed the relative abundance and metabolic activity trends of different denitrifying bacteria genera during the simulation period.

[0034] Based on the trends in relative abundance and metabolic activity, an index characterizing the stability of the internal distribution structure of denitrifying bacteria was calculated as a risk value for interspecific competition imbalance.

[0035] The stability indicators of the internal distribution structure of denitrifying bacteria, which are used to calculate the risk value of interspecific competition imbalance, include the decline of the denitrifying bacteria genus diversity index and the frequency of dominant genus replacement during the simulation period.

[0036] Furthermore, S5 includes:

[0037] Based on the mapping relationship, candidate carbon source addition strategies that meet the preset denitrification efficiency requirements are selected from all candidate carbon source addition strategies;

[0038] For each candidate carbon source addition strategy selected, the corresponding risk value of denitrification foam formation and the risk value of interspecific competition imbalance are weighted and calculated to obtain a comprehensive risk assessment value.

[0039] From the selected candidate carbon source addition strategies, the candidate carbon source addition strategy with the lowest comprehensive risk assessment value is selected as the optimized carbon source addition strategy.

[0040] Furthermore, S6 includes:

[0041] Extract the optimal carbon source addition amount and optimal carbon source addition rate mode from the optimized carbon source addition strategy;

[0042] Set the total carbon source addition control parameters for the carbon source addition equipment based on the optimized carbon source addition amount;

[0043] Based on the optimized carbon source dosing rate mode, a rate control command is generated for the carbon source dosing equipment during the dosing cycle.

[0044] According to the total amount control parameters and rate control instructions, the carbon source dosing equipment is driven to add carbon source to the anoxic tank of the soybean product wastewater treatment system.

[0045] On the other hand, the present invention provides a digital twin-based wastewater treatment process optimization system for soybean products, comprising the following modules:

[0046] The data acquisition module is used to acquire real-time operating data of the soybean product wastewater treatment system;

[0047] The model input module is used to input real-time running data into the digital twin model;

[0048] The strategy simulation module is used to set several candidate carbon source addition strategies based on real-time operation data. Each candidate carbon source addition strategy includes the candidate carbon source addition amount and the candidate carbon source addition acceleration rate mode. The digital twin model is used to simulate the denitrification efficiency value and the denitrification foam formation risk value under each strategy, and to establish the mapping relationship between the candidate carbon source addition strategy and the denitrification efficiency value and the denitrification foam formation risk value.

[0049] The risk assessment module is used to assess the risk value of interspecies competition imbalance caused by changes in the internal distribution structure of denitrifying bacteria due to the candidate carbon source addition rate mode, based on the simulation process of the digital twin model for each candidate carbon source addition strategy.

[0050] The strategy optimization module is used to select the optimal carbon source addition strategy from the candidate carbon source addition strategies based on the mapping relationship and the risk value of interspecific competition imbalance. The strategy meets the preset denitrification efficiency requirements and has the best comprehensive risk value of denitrification foam formation and interspecific competition imbalance.

[0051] The carbon source addition module is used to execute carbon source addition operations according to the optimized carbon source addition strategy.

[0052] Compared with the prior art, the present invention has the following beneficial effects:

[0053] 1. By constructing a digital twin model that integrates water quality transformation and microbial community dynamics, a quantitative simulation and forward-looking assessment of the entire process biological response triggered by carbon source addition strategies were achieved. This not only predicted the denitrification efficiency of the strategy, but also, more importantly, simultaneously assessed the risks of denitrification foam formation and interspecies competition imbalance in the microbial community. Thus, denitrification performance and system biological stability were inherently coordinated at the strategy generation stage. This effectively overcomes the limitations of existing technologies that solely pursue denitrification stoichiometric efficiency while neglecting secondary process issues. It also avoids microbial metabolic imbalance caused by improper carbon source addition, providing a synergistic control method for resolving the contradiction between the high carbon-to-nitrogen ratio requirement and operational obstacles such as foam and floating sludge in soybean product wastewater treatment.

[0054] 2. By comprehensively optimizing the selection of candidate strategies that meet the denitrification requirements based on biological risks, the optimized carbon source addition strategy ensures that while guaranteeing the denitrification effect, it also minimizes the excessive secretion of hydrophobic extracellular polymers and the drastic fluctuations in the denitrifying bacterial community structure. This elevates carbon source addition from a replenishment operation based on limited parameters to a regulatory means of maintaining the system's microbial ecological balance, thereby significantly enhancing the operational stability and robustness of the entire biological denitrification process. Ultimately, this achieves long-term synergistic compliance of total nitrogen and suspended solids concentrations in the effluent, avoiding the cycle of compliance, disturbance, and re-excess. Attached Figure Description

[0055] Figure 1 This is a flowchart of a method for optimizing the treatment process of soybean product wastewater based on digital twins according to the present invention;

[0056] Figure 2 This is a schematic diagram of the structure of a digital twin-based wastewater treatment process optimization system for soybean products according to the present invention.

[0057] Figure 3 This is a graph showing the changes in the denitrification efficiency, foam risk, and competition imbalance indicators of this invention over operating time. Detailed Implementation

[0058] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0059] Example 1: Figure 1This invention presents a method for optimizing the treatment process of soybean product wastewater based on digital twins, which includes the following steps:

[0060] S1: Obtain real-time operating data of the soybean product wastewater treatment system;

[0061] S2: Input real-time operational data into the digital twin model;

[0062] S3: Based on real-time operating data, set several candidate carbon source addition strategies. Each candidate carbon source addition strategy includes the amount of candidate carbon source added and the acceleration rate mode of candidate carbon source addition. Use a digital twin model to simulate the denitrification efficiency value and the denitrification foam formation risk value under each strategy, and establish the mapping relationship between the candidate carbon source addition strategy and the denitrification efficiency value and the denitrification foam formation risk value.

[0063] S4: For each candidate carbon source addition strategy, based on the simulation process of the digital twin model, evaluate the risk value of interspecific competition imbalance corresponding to the changes in the internal distribution structure of the denitrifying bacteria caused by the candidate carbon source addition acceleration mode.

[0064] S5: Based on the mapping relationship and the risk value of interspecific competition imbalance, select the optimal carbon source addition strategy from the candidate carbon source addition strategies that meets the preset denitrification efficiency requirements and has the best comprehensive risk value of denitrification foam formation and interspecific competition imbalance.

[0065] S6: Perform carbon source addition operation according to the optimized carbon source addition strategy.

[0066] S1: Obtain real-time operational data from the soybean product wastewater treatment system, implemented as follows:

[0067] When acquiring real-time operational data of the soybean product wastewater treatment system, the operation of acquiring real-time influent parameters at the system's inlet is performed. These real-time influent parameters include influent flow rate, influent ammonia nitrogen concentration, influent total nitrogen concentration, and influent chemical oxygen demand (COD) concentration. The influent flow rate is measured in real-time using an electromagnetic flowmeter installed on the main influent pipe. The electromagnetic flowmeter collects instantaneous flow rate values ​​at fixed time intervals and outputs them as current signals, for example, at 1 minute intervals. The influent ammonia nitrogen concentration is measured using an online analyzer based on the ammonia-sensitive electrode method. The online analyzer automatically collects water samples from the inlet periodically and completes the analysis, with an analysis cycle of, for example, 20 minutes. The influent total nitrogen concentration is measured using an online analyzer based on the alkaline potassium persulfate digestion ultraviolet spectrophotometry method. The online analyzer performs a complete cycle of sampling, digestion, and measurement, with a complete cycle of, for example, 30 minutes. The influent COD concentration is measured using an online water quality analyzer based on the potassium dichromate method, with a measurement cycle of, for example, 1 hour. All online analyzers transmit the measurement results to the data processing unit in real time via digital communication interface. The data processing unit stores the received real-time influent parameters according to timestamps.

[0068] This process acquires real-time status parameters of the biochemical reactors within the soybean product wastewater treatment system. The biochemical reactors include anaerobic, anoxic, and aerobic tanks. Real-time status parameters include the dissolved oxygen concentration in the anaerobic, anoxic, and aerobic tanks; the mixed liquor sludge concentration in the anaerobic, anoxic, and aerobic tanks; the mixed liquor temperature in the anaerobic, anoxic, and aerobic tanks; and the mixed liquor temperature in the aerobic tanks. Dissolved oxygen concentration is monitored using a dissolved oxygen sensor submerged at a depth in the middle of the tank, with a data acquisition frequency of, for example, once every 10 seconds. Mixed liquor sludge concentration is measured using an optical sludge concentration meter installed on the outlet pipe of the mixed liquor circulation pump within the tank, with a measurement interval of, for example, 15 minutes. Mixed liquor temperature is measured using a platinum resistance temperature sensor installed within the tank, with a sampling frequency of, for example, 1 minute. Measurement data from dissolved oxygen sensors, optical sludge concentration meters, and platinum resistance temperature sensors are uploaded in real time to the data storage area of ​​the central monitoring system via a fieldbus network, forming a real-time status parameter set. The data storage area records the name of the reaction tank from which each data item in the real-time status parameter set originated and the timestamp of the acquisition.

[0069] The process involves retrieving historical carbon source dosing data from the soybean product wastewater treatment system. This historical data is extracted from the historical record database of the carbon source dosing control system. The data includes multiple chronologically ordered historical dosing event records. Each record contains the time of the event, the amount of carbon source added at that time, and the carbon source dosing acceleration mode used to control the carbon source dosing equipment for a period after that time. The carbon source dosing acceleration mode is a sequence of control commands describing the change in carbon source dosing per unit time over time. This sequence may include constant rate commands, incremental rate commands, or proportional rate commands linked to the influent flow rate. The storage time span of the historical record database is set according to process adjustment requirements, for example, 30 days. The process of retrieving historical carbon source dosing data involves sending a data query request to the carbon source dosing control system. The system then retrieves historical dosing event records from the historical record database that meet the time range criteria and returns the retrieved records. The returned historical carbon source dosing event records are organized into a structured list of historical carbon source dosing data. Each record in the list contains a complete time point field, a carbon source dosing amount field, and a carbon source dosing acceleration rate mode code field. The carbon source dosing acceleration rate mode code field corresponds to a specific control command sequence. Real-time water inflow parameters, real-time status parameters, and historical carbon source dosing data together constitute the real-time operating data used for the digital twin model. If communication with any data source is interrupted during the acquisition process, the most recently successfully acquired valid data value from that data source is used, and a status log marking the data delay is generated in the central monitoring system. The status log records the name of the interrupted data source and the time of occurrence.

[0070] S2: Input real-time operational data into the digital twin model, implemented as follows:

[0071] Real-time influent parameters are input to the influent water quality parameter interface of the digital twin model. The influent water quality parameter interface is a programmed entry point within the digital twin model with a specific data format convention. The input operation involves passing a data set containing influent flow rate, influent ammonia nitrogen concentration, influent total nitrogen concentration, and influent chemical oxygen demand (COD) concentration to the influent water quality parameter interface. Upon receiving the data set, the influent water quality parameter interface performs data validity verification. The verification rule determines whether each value falls within a preset reasonable value range; for example, the preset reasonable value range for influent ammonia nitrogen concentration is 0 mg / L to 500 mg / L. Data that passes verification is stored within the digital twin model in a dedicated storage location to maintain the latest influent water quality data.

[0072] Real-time status parameters are input to the biochemical reaction process parameter interface of the digital twin model. The biochemical reaction process parameter interface is a programmed entry point in the digital twin model used to receive operating data from each biochemical reaction tank. The input operation involves passing a data set containing values ​​for dissolved oxygen concentration, sludge concentration, and temperature of the anaerobic, anoxic, and aerobic mixed liquors. The biochemical reaction process parameter interface then maps each value in the data set to a state variable representing the corresponding reaction tank and parameter within the digital twin model, according to the reaction tank name and parameter type associated with that value, overwriting the previous value of that state variable.

[0073] Historical carbon source addition data is input into the carbon source addition record interface of the digital twin model. The carbon source addition record interface is a programmed entry point in the digital twin model for receiving chronologically ordered records of carbon source addition events. The input operation involves passing a list to the carbon source addition record interface. Each element in the list is a record, containing a time point representing a specific moment in the past, a carbon source addition quantity representing the mass of carbon source added at that moment, and a carbon source addition acceleration rate pattern encoding data defining the acceleration rate variation pattern. After receiving this list, the carbon source addition record interface stores it in a data storage structure within the digital twin model that supports querying by time point.

[0074] The digital twin model includes a water quality transformation kinetics sub-model. This sub-model is a set of computational rules used to calculate changes in nitrogen speciation concentrations in water. It uses real-time influent parameters such as ammonia nitrogen concentration, total nitrogen concentration, and chemical oxygen demand (COD) concentration as initial or boundary conditions. It also uses historical carbon source addition data to obtain information on carbon source addition amounts at any point within the simulation timeframe. The sub-model simulates the ammonia nitrogen conversion process based on the chemical reaction rate equation for the conversion of ammonia nitrogen to nitrite nitrogen, with the reaction rate influenced by the dissolved oxygen concentration in the aerobic tank. Finally, it simulates the nitrification process based on the chemical reaction rate equation for the conversion of nitrite nitrogen to nitrate nitrogen. The water quality transformation kinetics sub-model simulates the denitrification process through calculation. This calculation is based on the chemical reaction rate equation for the reduction of nitrate nitrogen to nitrogen gas. This equation includes a term affected by carbon source availability, which is determined by the carbon source dosage data and the influent chemical oxygen demand concentration at the simulation time. The water quality transformation kinetics sub-model advances the calculation in discrete time steps, updating the simulated concentration values ​​of various nitrogen forms in all reaction tanks within each time step. The length of the time step can be set, for example, to 1 minute.

[0075] The digital twin model includes a microbial community succession sub-model. This sub-model is a set of computational rules used to calculate the changes in the quantity and composition of denitrifying functional bacteria. It uses real-time state parameters such as mixed liquor temperature, dissolved oxygen concentration, and mixed liquor sludge concentration as environmental inputs. The sub-model divides the denitrifying functional bacteria into multiple virtual genera, each assigned different growth characteristic parameters, such as maximum specific growth rate and substrate half-saturation constant. The sub-model simulates the growth process of the denitrifying functional bacteria through computation, based on the biomass growth equation for each genera. The rate term of the growth equation is influenced by the environmental inputs and simulated nitrate nitrogen concentration and carbon source availability information from the water quality transformation kinetics sub-model. Finally, the sub-model simulates the metabolic processes of the denitrifying functional bacteria, outputting the extracellular polymer secretion rate for each genera. The microbial community succession sub-model simulates the population structure dynamics of denitrifying functional bacteria by calculating and solving the biomass change equations for all genera to obtain the change in the proportion of each genera in the total biomass, i.e., the relative abundance, over time. The microbial community succession sub-model uses the same time step as the water quality transformation kinetics sub-model.

[0076] The water quality transformation kinetics sub-model and the microbial community succession sub-model are coupled. This coupling is achieved by having the two sub-models perform calculations alternately and exchange data within the same simulation time step. In the calculation sequence of each simulation time step, the microbial community succession sub-model performs calculations first. The calculation of the microbial community succession sub-model consumes the environmental conditions input, simulated nitrate nitrogen concentration, and carbon source availability information for the current step, and produces a calculation result called the microbial metabolic activity parameter. This parameter is a numerical value that comprehensively reflects the total metabolic capacity of all denitrifying functional bacterial genera within the current step. In subsequent calculation sequences within the same simulation time step, the water quality transformation kinetics sub-model performs calculations. When calculating the rate equation for the denitrification process, the water quality transformation kinetics sub-model requires the basic reaction rate value calculated from the rate equation. The water quality transformation kinetics sub-model acquires the microbial metabolic activity parameter generated by the microbial community succession sub-model in this step. The water quality transformation kinetics sub-model multiplies the baseline reaction rate value with the microbial metabolic activity parameter. The result of this multiplication is used as the actual denitrification reaction rate value for that time step, updating the simulated nitrate nitrogen concentration. By executing this sequential operation at each time step—first generating microbial metabolic activity parameters from the microbial community succession sub-model, and then using these parameters to correct the reaction rate from the water quality transformation kinetics sub-model—coupling between the two sub-models is achieved. The digital twin model completes the entire set simulation duration by repeatedly executing this cycle of step progression and data exchange.

[0077] S3: Based on real-time operational data, several candidate carbon source addition strategies are set. Each candidate carbon source addition strategy includes the candidate carbon source addition amount and the candidate carbon source addition acceleration rate mode. A digital twin model is used to simulate the denitrification efficiency value and the denitrification foam formation risk value under each strategy, and a mapping relationship is established between the candidate carbon source addition strategy and the denitrification efficiency value and the denitrification foam formation risk value. The implementation is as follows:

[0078] The process involves generating several candidate carbon source addition strategies based on real-time influent parameters and historical carbon source addition data, including different candidate carbon source addition amounts and acceleration rates. When generating candidate carbon source addition strategies, the influent total nitrogen concentration and influent chemical oxygen demand (COD) concentration values ​​from the real-time influent parameters are used. The carbon source addition amount data from the most recent period in the historical carbon source addition data is used; the length of the most recent period is determined based on the process adjustment frequency, for example, set to 24 hours. The specific process for generating candidate carbon source addition amounts involves calculating a baseline carbon requirement, which is equal to the influent total nitrogen concentration multiplied by a theoretical carbon-to-nitrogen ratio (CNR). The theoretical CNR is determined based on batch denitrification experimental data of soybean product wastewater. The average ratio between the amount of carbon source consumed and the amount of total nitrogen removed during denitrification is obtained by fitting the experimental data; this average ratio is used as the theoretical CNR. A typical range for the theoretical CNR is, for example, 2.8 to 3.5. Based on the baseline carbon demand, a set of discrete candidate carbon source addition values ​​are generated by multiplying them by a series of preset proportional adjustment coefficients. These proportional adjustment coefficients are set to explore a reasonable addition range around the baseline demand. For example, these coefficients include 0.9, 1.0, 1.1, and 1.2, corresponding to 90%, 100%, 110%, and 120% of the baseline carbon demand, respectively. The specific process for generating candidate carbon source addition acceleration rate patterns involves defining three carbon source addition acceleration rate pattern templates with different carbon source addition time distribution characteristics. The first template is the constant rate pattern, which stipulates that the carbon source addition equipment operates at a constant mass flow rate throughout the set future addition period. The instruction sequence corresponding to the constant rate pattern is a constant value. The second template is the early-stage enhancement mode. This mode specifies that the carbon source dosing equipment operates at a higher first mass flow rate for a certain proportion of the future dosing period, and at a lower second mass flow rate for the remaining proportion of the time. The early-stage enhancement mode requires defining the time segmentation ratio and the flow rate ratio; for example, the time segmentation ratio could be set to the first one-third of the time and the last two-thirds, and the first mass flow rate to second mass flow rate ratio could be set to 2:1. The third template is the intermittent pulse mode. This mode specifies that the carbon source dosing equipment operates in a fixed cycle, with each cycle containing a period when the carbon source dosing equipment is on and a period when it is off. The intermittent pulse mode requires defining the percentage of the cycle that the on period is in; for example, the percentage could be set to 30%. Each candidate carbon source dosing acceleration mode template corresponds to a digital instruction set containing a sequence of time and flow rate instructions that can be parsed and executed by the carbon source dosing control system. Combining each candidate carbon source dosing quantity value generated by the above methods with each candidate carbon source dosing acceleration mode template defined by the above methods constitutes a complete candidate carbon source dosing strategy.

[0079] The process involves inputting each candidate carbon source addition strategy into the digital twin model, driving the model to simulate the biochemical reaction process under that strategy. For each candidate carbon source addition strategy to be simulated, a new set of carbon source addition data representing the future plan is input into the carbon source addition record interface of the digital twin model. This new set of carbon source addition data is generated based on the candidate carbon source addition amount and acceleration rate pattern included in the candidate carbon source addition strategy, specifying in detail the amount of carbon source input that the model should use at each time point in the future period starting from the current simulation time. Driving the digital twin model to simulate means that within the digital twin model, starting from the current time and ending at a time point with a future time span, the coupled calculation process of the water quality transformation kinetics sub-model and the microbial community succession sub-model is executed according to the inherent calculation logic of the model. The length of the simulation time span needs to be selected to ensure that the complete impact of the carbon source addition strategy on the biochemical reaction process can be observed. The length of the simulation time span is determined based on the average hydraulic retention time of the soybean wastewater in the target biochemical reaction tank, i.e., the anoxic tank. The simulation time span should not be less than the average hydraulic retention time. For example, the simulation time span is set to 1.5 times the average hydraulic retention time. If the average hydraulic retention time is 8 hours, then the simulation time span is set to 12 hours. The digital twin model advances the calculation step by step according to its internally set simulation time step. The simulation time step is set according to the model's calculation stability and efficiency requirements. For example, the simulation time step is 1 minute. Within each simulation time step, the model iteratively updates the values ​​of all simulation variables, including the concentrations of various nitrogen forms, microbial biomass, and metabolite concentrations, based on the planned carbon source input corresponding to the current simulation moment and combined with other real-time operating data that have been input.

[0080] The process involves obtaining the simulated effluent total nitrogen concentration from the digital twin model output to calculate the denitrification efficiency value, and acquiring simulated microbial metabolic indicators to assess the risk of denitrification foam formation. After the digital twin model completes the simulation of a candidate carbon source addition strategy, the simulated effluent total nitrogen concentration at the end of the simulation time span is extracted from the model output data. The unit of the simulated effluent total nitrogen concentration value is milligrams per liter. The denitrification efficiency value is calculated by subtracting the simulated effluent total nitrogen concentration value from the real-time influent parameters, obtaining a concentration difference, dividing the concentration difference by the influent total nitrogen concentration value, and then multiplying the result by 100 to obtain the denitrification efficiency value in percentage form. Simultaneously, time-series data of microbial metabolic indicators generated during the simulation time span are obtained from the digital twin model output data. The microbial metabolic indicators used to assess the risk of denitrification foam formation include the proportion of hydrophobic components and the secretion rate of simulated extracellular polymers. The process for assessing the risk of denitrification foam formation involves selecting a period within the complete simulation time span where the process conditions tend to stabilize. This assessment period begins at the midpoint of the simulation span and ends at the end of the simulation; for example, the latter half of the simulation might be chosen. If the simulation span is 12 hours, the assessment period would be the last 6 hours. Within this assessment period, the arithmetic mean of the hydrophobic component proportions across all simulation time steps is calculated to obtain the average hydrophobicity proportion used for risk assessment. Similarly, the arithmetic mean of the extracellular polymeric substance secretion rates across all simulation time steps is calculated to obtain the average secretion rate used for risk assessment, expressed in milligrams per liter per hour. A high-risk threshold for hydrophobicity ratio is set. This threshold is determined by tracing historical events of severe foaming problems, extracting actual operational data for the periods of the events, and inputting this data into a digital twin model for inversion simulation. Typical high-values ​​of the hydrophobic component ratio are then statistically analyzed from the microbial metabolic indicators output by the inversion simulation. These typical high-values ​​are used as the basis for setting the high-risk threshold for hydrophobicity ratio. For example, based on statistics from multiple event inversion simulations, the high-risk threshold for hydrophobicity ratio is set to 40%. Similarly, a high-risk threshold for secretion rate is set. This threshold is also determined by tracing historical events of severe foaming problems, extracting actual operational data for the periods of the events, and inputting this data into a digital twin model for inversion simulation. Typical high-values ​​of the extracellular polymeric secretion rate are then statistically analyzed from the microbial metabolic indicators output by the inversion simulation. These typical high-values ​​are used as the basis for setting the high-risk threshold for secretion rate. For example, based on statistics from multiple event inversion simulations, the high-risk threshold for secretion rate is set to 20 mg / L / hour. The calculation rule for the risk value of denitrification foam formation is to first assign risk scores to the proportion of hydrophobic components and the secretion rate.For the average hydrophobicity ratio, if it is greater than or equal to the high-risk threshold for hydrophobicity ratio, the risk score is 2; if it is greater than or equal to but less than the high-risk threshold, the risk score is 1; and if it is less than the high-risk threshold, the risk score is 0. For the average secretion rate, if it is greater than or equal to the high-risk threshold for secretion rate, the risk score is 2; if it is greater than or equal to but less than the high-risk threshold, the risk score is 1; and if it is less than the high-risk threshold, the risk score is 0. The weighting of the hydrophobicity ratio risk score and the secretion rate risk score is based on the analysis of the dominant factors in foam formation in engineering experience, assuming that their contributions are equivalent, so a direct addition method is used. Adding the hydrophobicity ratio risk score and the secretion rate risk score yields the denitrification foam formation risk value corresponding to the candidate carbon source addition strategy. The denitrification foam formation risk value is an integer ranging from 0 to 4.

[0081] The process involves associating each candidate carbon source addition strategy with its corresponding denitrification efficiency value and denitrification foam formation risk value to form a mapping relationship. A unique numerical identifier is assigned to each candidate carbon source addition strategy that completes the simulation. A structured mapping table is created, containing the following fields: strategy identifier, candidate carbon source addition amount, candidate carbon source addition acceleration mode code, denitrification efficiency value, and denitrification foam formation risk value. The strategy identifier, the specific candidate carbon source addition amount, the candidate carbon source addition acceleration mode code, the calculated denitrification efficiency value, and the assessed denitrification foam formation risk value for each candidate carbon source addition strategy are written as a complete record into the corresponding field of the mapping table. This mapping table, populated with all candidate strategy records, constitutes the mapping relationship from the candidate carbon source addition strategy to its simulation results, namely the denitrification efficiency value and the denitrification foam formation risk value. This mapping relationship allows retrieval of the corresponding denitrification efficiency value and denitrification foam formation risk value by querying the strategy identifier or a specific combination of strategy parameters.

[0082] The microbial metabolic indicators relied upon for assessing the risk of denitrification foam formation include the proportion of hydrophobic components and the secretion rate in the simulated extracellular polymers (APIs). In the microbial community succession sub-model of the digital twin model, the total amount of simulated APIs is classified into hydrophobic and non-hydrophobic components based on their chemical properties. The proportion of hydrophobic components is calculated at each simulation time step by dividing the simulated concentration of hydrophobic components at the end of that step by the total simulated concentration of APIs at the end of that step. The secretion rate is calculated at each simulation time step by subtracting the total simulated concentration of APIs at the end of the previous step from the total simulated concentration of APIs at the end of that step, and then dividing the resulting concentration change by the length of the simulation time step. The microbial community succession sub-model outputs the proportion of hydrophobic components and the secretion rate at each step throughout the entire simulation time span. These values ​​are arranged chronologically to form time-series data, which serves as the direct input data source for assessing the risk of denitrification foam formation.

[0083] S4: For each candidate carbon source addition strategy, based on the simulation process of the digital twin model, assess the risk value of interspecies competition imbalance corresponding to the changes in the internal distribution structure of the denitrifying bacteria community caused by the candidate carbon source addition acceleration mode. This is implemented as follows:

[0084] The steps involve executing a simulation of each candidate carbon source addition strategy and obtaining the denitrifying functional microbial community simulation data output by the digital twin model during the biochemical reaction process under that strategy. After the digital twin model completes the simulation of a candidate carbon source addition strategy, the denitrifying functional microbial community simulation data is extracted from the microbial community succession sub-model output of the digital twin model. The denitrifying functional microbial community simulation data is a dataset arranged chronologically according to the simulation time. Each record in the dataset corresponds to a simulation time step. Each record contains the total biomass concentration data of the denitrifying functional microbial community calculated at the end of that simulation time step, expressed in milligrams per liter. Each record contains biomass concentration data for each denitrifying bacteria genus calculated at the end of the simulation time step. The unit of biomass concentration data for each denitrifying bacteria genus is milligrams per liter. Each denitrifying bacteria genus is assigned a unique identifier code in the microbial community succession sub-model for identification. For example, the microbial community succession sub-model divides the denitrifying functional bacteria community into three genera, identified by the codes DN1, DN2, and DN3. Each record also contains metabolic activity intensity data for each denitrifying bacteria genus calculated at the end of the simulation time step. The metabolic activity intensity value is a dimensionless value calculated by the microbial community succession sub-model based on the growth status and environmental conditions of the genus. The specific operation of obtaining the simulation data of the denitrifying functional bacteria community is to call the data output interface function of the digital twin model, passing in the unique identifier of the target candidate carbon source addition strategy. The data output interface function returns a complete time series data set corresponding to the candidate carbon source addition strategy, containing all the above data fields.

[0085] The process involves analyzing the relative abundance and metabolic activity trends of different denitrifying bacteria genera within a simulation period, based on simulated data of denitrifying functional microbial communities. The analysis is performed for each simulation time step within the simulation period. For a specific denitrifying bacteria genus and time step within the simulation period, the relative abundance of that genus and time step is calculated. Relative abundance is equal to the biomass concentration of that genus and time step in the simulation time step record divided by the total biomass concentration of the denitrifying functional microbial community in the simulation time step record. By calculating the relative abundance of each denitrifying bacteria genus and time step across all simulation time steps within the simulation period, a data sequence of the relative abundance value of each genus and time step is obtained; this data sequence is called the relative abundance trend of that genus and time step. Simultaneously, the analysis process also directly reads data for each simulation time step within the simulation period. For a specific denitrifying bacteria genus and time step within the simulation period, the metabolic activity intensity value of that genus and time step is directly read from the simulation time step record. By reading the metabolic activity intensity data of each denitrifying bacteria genus at all simulation time steps within the simulation period, a data sequence of the metabolic activity intensity data of each denitrifying bacteria genus over time is obtained. This data sequence is called the metabolic activity change trend of that denitrifying bacteria genus. For the relative abundance change trend of each denitrifying bacteria genus, the starting and ending values ​​of the relative abundance change trend are calculated. The starting value is defined as the relative abundance value of the genus calculated at the first simulation time step within the simulation period. The ending value is defined as the relative abundance value of the genus calculated at the last simulation time step within the simulation period. For the metabolic activity change trend of each denitrifying bacteria genus, the starting and ending values ​​of the metabolic activity change trend are calculated. The starting value is defined as the metabolic activity intensity data of the genus recorded at the first simulation time step within the simulation period. The ending value is defined as the metabolic activity intensity data of the genus recorded at the last simulation time step within the simulation period.

[0086] The procedure involves calculating an index characterizing the stability of the internal distribution structure of denitrifying bacteria, based on trends in relative abundance and metabolic activity, as a risk value for interspecific competition imbalance. The calculation process utilizes trends in relative abundance and metabolic activity obtained from simulation data of denitrifying functional bacteria. Indicators characterizing the stability of the internal distribution structure of denitrifying bacteria include the decrease in the denitrifying genus diversity index and the frequency of dominant genus turnover within the simulation period. The Shannon diversity index is used to calculate the denitrifying genus diversity index. For each simulation time step within the simulation period, the Shannon diversity index for that time step is calculated. The calculation method involves first obtaining the relative abundance values ​​of all denitrifying genus groups for that simulation time step. Then, for each denitrifying genus group, the relative abundance value is multiplied by its natural logarithm. Finally, the products of all denitrifying genus groups are summed, and the negative of the sum is taken to obtain the Shannon diversity index for that simulation time step. The Shannon diversity index is a dimensionless value. The Shannon diversity index was calculated for each simulation time step within the simulation period, forming a Shannon diversity index time series. The decrease in the denitrifying bacteria genus diversity index within the simulation period was assessed by comparing the initial and final segments of the Shannon diversity index time series. Specifically, the average Shannon diversity index for all simulation time steps in the first quarter of the simulation period was taken as the initial diversity index. The average Shannon diversity index for all simulation time steps in the second quarter of the simulation period was taken as the final diversity index. The decrease was equal to the difference between the initial and final diversity indices, divided by the initial diversity index, and the result was expressed as a percentage. The dominance genus turnover frequency was calculated by determining the dominant genus for each simulation time step within the simulation period. The dominant genus was defined as the denitrifying bacteria genus group with the highest relative abundance value at that simulation time step. By analyzing the identity of the dominant genus for all simulation time steps within the simulation period, a sequence of dominant genus changes over time was formed. The dominance genus turnover frequency was defined as the number of times the identity of the dominant genus changed throughout the entire simulation period. A change in the dominant genus identity refers to a difference in the identifier code of the dominant genus between two adjacent simulation time steps. Each occurrence of a different identifier code is counted as one replacement. Dividing the total number of replacements by the total duration of the simulation period yields the frequency of dominant genus replacement per unit time, measured in times per hour. When calculating the risk value of interspecific competition imbalance, it is necessary to set a high-risk threshold for the magnitude of the decline in the diversity index and a high-risk threshold for the frequency of dominant genus replacement.The high-risk threshold for the decline in diversity index is obtained by performing digital twin model inversion simulation analysis on the operating data of historical periods with long-term stable operation and good solid-liquid separation effect. Specifically, the operating data of these historical stable periods are input into the digital twin model for inversion simulation. The decline in diversity index in each inversion simulation period is calculated, and the distribution of the decline values ​​of diversity index obtained in all inversion simulation periods is statistically analyzed. The high percentile value of this distribution is taken as the basis for setting the high-risk threshold for the decline in diversity index. For example, the 95th percentile value is taken as the high-risk threshold for the decline in diversity index, and this threshold may be 15%. The high-risk threshold for dominant genus turnover frequency was obtained through digital twin model inversion simulation analysis of historical operational data from a long-term stable operation with good solid-liquid separation performance. Specifically, the operational data from these historical stable periods were input into the digital twin model for inversion simulation. The dominant genus turnover frequency within each simulation period was calculated, and the distribution of the dominant genus turnover frequencies obtained from all simulation periods was statistically analyzed. A high percentile value of this distribution was taken as the basis for setting the high-risk threshold for dominant genus turnover frequency; for example, the 95th percentile value was used as the high-risk threshold for dominant genus turnover frequency, which could be 0.5 times per hour. The calculation rule for the risk value of interspecific competition imbalance is to assess the risk of both the decline in diversity index and the dominant genus turnover frequency and assign scores accordingly. For the magnitude of the decline in the diversity index, if it is greater than or equal to the high-risk threshold for the magnitude of the decline, the risk score is 2; if it is greater than or equal to but less than the high-risk threshold, the risk score is 1; and if it is less than the high-risk threshold, the risk score is 0. For the frequency of dominant genus turnover, if it is greater than or equal to the high-risk threshold, the risk score is 2; if it is greater than or equal to but less than the high-risk threshold, the risk score is 1; and if it is less than the high-risk threshold, the risk score is 0. The magnitude of the decline risk score and the frequency of turnover risk score are added together to obtain the interspecific competition imbalance risk value corresponding to the candidate carbon source addition strategy. The interspecific competition imbalance risk value is an integer ranging from 0 to 4.

[0087] The stability indices of the internal distribution structure of denitrifying bacteria, used to calculate the risk value of interspecific competition imbalance, include the decline in the denitrifying bacteria genus diversity index and the frequency of dominant genus turnover during the simulation period. The decline in the denitrifying bacteria genus diversity index specifically characterizes the degree to which species richness and evenness within the denitrifying bacteria community decreases under the influence of candidate carbon source addition strategies. The frequency of dominant genus turnover specifically characterizes how frequently the dominant genus within the denitrifying bacteria community changes under the influence of candidate carbon source addition strategies. These two indicators quantify the dynamic stability of the community structure from different dimensions.

[0088] S5: Based on the mapping relationship and the risk value of interspecific competition imbalance, select the optimal carbon source addition strategy from the candidate carbon source addition strategies that meets the preset denitrification efficiency requirements and has the best comprehensive risk value of denitrification foam formation and interspecific competition imbalance. The implementation is as follows:

[0089] The execution process involves selecting candidate carbon source addition strategies from all available options based on a mapping relationship, choosing those that achieve the preset denitrification efficiency requirement. The mapping relationship is a data table recording the identifier, dosage, acceleration rate, denitrification efficiency value, and denitrification foam formation risk value for each candidate carbon source addition strategy. The preset denitrification efficiency requirement is a pre-defined percentage value. The specific steps for setting this requirement are: obtaining the effluent total nitrogen concentration discharge limit that the soybean product wastewater treatment system must adhere to, for example, 15 mg / L; reading the influent total nitrogen concentration from real-time influent parameters; calculating the theoretical minimum denitrification efficiency percentage by subtracting the effluent total nitrogen concentration discharge limit from the influent total nitrogen concentration, obtaining a difference concentration, dividing this difference concentration by the influent total nitrogen concentration, and then multiplying the result by 100. A safety margin percentage is added to the theoretical minimum nitrogen removal efficiency percentage. This safety margin percentage is used to account for fluctuations in influent water quality and errors in model simulation; for example, the safety margin percentage is 5%. The theoretical minimum nitrogen removal efficiency percentage is added to the safety margin percentage, and the result is used as the preset nitrogen removal efficiency requirement. The screening process involves traversing each record in the mapping relationship data table. For each record in the mapping relationship data table, the value stored in the nitrogen removal efficiency field of that record is read. This nitrogen removal efficiency value is compared with the preset nitrogen removal efficiency requirement. If the nitrogen removal efficiency value is greater than or equal to the preset nitrogen removal efficiency requirement, the identifier of the candidate carbon source addition strategy corresponding to that record is added to a new set, which is called the qualified strategy set. If the nitrogen removal efficiency value is less than the preset nitrogen removal efficiency requirement, the candidate carbon source addition strategy corresponding to that record is ignored. After traversing all records, the qualified strategy set contains all candidate carbon source addition strategies whose nitrogen removal efficiency values ​​meet the preset nitrogen removal efficiency requirement.

[0090] The process involves performing a weighted calculation of the denitrification foam formation risk value and the interspecific competition imbalance risk value for each selected candidate carbon source addition strategy, to obtain a comprehensive risk assessment value. This step is performed for each candidate carbon source addition strategy in the qualified strategy set. For a candidate carbon source addition strategy in the qualified strategy set, firstly, the denitrification foam formation risk value corresponding to that strategy is obtained. The denitrification foam formation risk value is read from the record corresponding to that strategy in the mapping relationship data table. Then, the interspecific competition imbalance risk value corresponding to that strategy is obtained. The interspecific competition imbalance risk value is read from another independently stored data list indexed by the candidate carbon source addition strategy identifier. This data list stores the calculated interspecific competition imbalance risk value for each candidate carbon source addition strategy. Before performing the weighted calculation, the denitrification foam formation risk value and the interspecific competition imbalance risk value are normalized respectively. The weighted calculation requires assigning a weight coefficient to the denitrification foam formation risk value and another weight coefficient to the interspecific competition imbalance risk value. The weighting coefficients are set based on the analysis of historical operational events of the soybean product wastewater treatment system. The analysis process involves collecting historical records of events causing denitrification foam leading to sludge buildup, and historical records of events causing fluctuations in effluent quality due to microbial community instability. The extent of damage caused by these two types of events to the normal operation of the system, the recovery time required, and the treatment costs are assessed. Based on the assessment results, the relative importance of the two types of risks is quantified. For example, the analysis may indicate that denitrification foam problems occur more rapidly and have a more direct impact on effluent suspended solids, thus assigning it a higher weight; while microbial community competition imbalance problems have a more profound but slower impact, thus assigning it a relatively lower weight. Based on this quantitative analysis, a weighting coefficient is assigned to the risk value of denitrification foam formation, for example, 0.6; and a weighting coefficient is assigned to the risk value of interspecific competition imbalance, for example, 0.4. The sum of the two weighting coefficients is 1. The specific execution process of the weighted calculation is to multiply the risk value of denitrification foam formation by the weighting coefficient of the risk value of denitrification foam formation, obtaining the first product. Multiply the risk value of interspecific competition imbalance by its weighting coefficient to obtain the second product. Add the first product to the second product; the sum is the comprehensive risk assessment value of the candidate carbon source addition strategy. The comprehensive risk assessment value is a real number. For each candidate carbon source addition strategy in the qualified strategy set, repeat the above process of obtaining the risk value and weighting calculation to calculate a comprehensive risk assessment value for each strategy.

[0091] The process involves selecting the candidate carbon source addition strategy with the lowest comprehensive risk assessment value from the selected candidate carbon source addition strategies as the optimized carbon source addition strategy. This step is performed after the comprehensive risk assessment values ​​of all candidate carbon source addition strategies in the qualified strategy set have been calculated. First, the comprehensive risk assessment value of each candidate carbon source addition strategy in the qualified strategy set is collected. Then, the minimum value among all these comprehensive risk assessment values ​​is identified. Next, it is checked which candidate carbon source addition strategies have a comprehensive risk assessment value equal to this minimum value. If only one candidate carbon source addition strategy has a comprehensive risk assessment value equal to the minimum value, then that candidate carbon source addition strategy is directly determined as the optimized carbon source addition strategy. If two or more candidate carbon source addition strategies have comprehensive risk assessment values ​​that are simultaneously equal to the minimum value, an additional rule is required for decision-making. The additional rule is to compare the denitrification efficiency values ​​of these candidate carbon source addition strategies with the same lowest comprehensive risk assessment value. The denitrification efficiency values ​​of these tied strategies are read from the mapping relationship data table. The candidate carbon source addition strategy with the highest denitrification efficiency value is selected as the optimized carbon source addition strategy. If multiple strategies have the same denitrification efficiency value, the candidate carbon source addition amounts of these strategies are further compared. The candidate carbon source addition strategy with the smallest candidate carbon source addition amount value is selected as the optimized carbon source addition strategy. If a distinction still cannot be made at this point, one candidate carbon source addition strategy is randomly selected as the optimized carbon source addition strategy. The optimized carbon source addition amount and optimized carbon source addition acceleration rate mode of the scheme finally selected as the optimized carbon source addition strategy will be extracted to guide subsequent actual carbon source addition operations.

[0092] S6: Execute the carbon source addition operation according to the optimized carbon source addition strategy, as follows:

[0093] The process involves extracting the optimized carbon source dosage and the optimized carbon source acceleration rate pattern from the optimized carbon source dosage strategy. The optimized carbon source dosage strategy is a data record containing two data items: the optimized carbon source dosage amount and the optimized carbon source acceleration rate pattern. The extraction operation reads the optimized carbon source dosage amount data item from the optimized carbon source dosage strategy data record; this data item stores a value in kilograms. Simultaneously, the extraction operation reads the optimized carbon source acceleration rate pattern data item from the optimized carbon source dosage strategy data record; this data item stores either a text string or a numeric code, which corresponds to a predefined carbon source dosage time distribution rule.

[0094] The steps involve setting the total dosage control parameters for the carbon source dosing equipment based on the optimized carbon source dosage. The carbon source dosing equipment is a physical device used to add carbon sources to a soybean product wastewater treatment system. The total dosage control parameter is the target total dosage that the carbon source dosing equipment needs to complete within one dosing cycle. The process of setting the total dosage control parameter involves using the value read from the optimized carbon source dosage data item as the target total mass. Necessary parameter conversions are performed based on the actual form of the carbon source used by the carbon source dosing equipment. If the carbon source dosing equipment uses a liquid carbon source, and the equipment control system uses volume liters as the total dosage setting unit, then the target total mass needs to be converted to the target total volume. The conversion method is to obtain the density value of the liquid carbon source used. The density value is obtained by measuring the batch of carbon source solution, for example, using a hydrometer, and the density value is stored in kilograms per liter. The target total volume equals the target total mass divided by the density value. If the carbon source dosing equipment directly supports setting the unit as mass, then no conversion is required. The calculated or directly obtained target total amount is input into the control unit of the carbon source dosing equipment via a data communication interface or human-machine interface, and written into a register or variable specifically storing the target total dosing amount in the control unit. The value stored in this register or variable becomes the total dosing amount control parameter for this dosing cycle.

[0095] The process involves generating rate control commands for the carbon source dosing equipment during the dosing cycle based on the optimized carbon source dosing acceleration rate mode. The generation of rate control commands is based on the optimized carbon source dosing acceleration rate mode data item extracted from the optimized carbon source dosing strategy. Rate control commands are a set of commands that control the operating status of the carbon source dosing equipment at different points in time during the dosing cycle. The generation process first identifies the mode type represented by the optimized carbon source dosing acceleration rate mode data item. If the optimized carbon source dosing acceleration rate mode data item indicates a constant rate mode, one rate control command is generated. This rate control command specifies that the carbon source dosing equipment maintains a constant dosing acceleration rate throughout the entire dosing cycle. The constant dosing acceleration rate value is calculated by dividing the optimized carbon source dosing amount by the total duration of the dosing cycle, which is a preset value, such as 6 hours. If the optimized carbon source dosing acceleration rate mode data item indicates an early-stage enhancement mode, two rate control commands are generated. The early-stage enhancement mode requires dividing the dosing cycle into two sub-periods. Based on the definition of the early-stage enhancement mode, the proportion of the first sub-period to the total duration is determined, for example, one-third. Determine the acceleration ratio between the preceding and following sub-periods, for example, a ratio of 2:1. Calculate the duration of the preceding sub-period as equal to the total duration of the feeding cycle multiplied by the proportion of the preceding sub-period. Calculate the duration of the following sub-period as equal to the total duration of the feeding cycle minus the duration of the preceding sub-period. Based on the optimized carbon source dosage and acceleration ratio, solve the simultaneous equations to obtain the acceleration values ​​for the preceding and following sub-periods. The first generated rate control command specifies that, from the start of the feeding cycle, for the duration of the preceding sub-period, the carbon source feeding equipment operates at the acceleration value of the preceding sub-period. The second generated rate control command specifies that, from the end of the preceding sub-period until the end of the feeding cycle, the carbon source feeding equipment operates at the acceleration value of the following sub-period. If the optimized carbon source acceleration mode data item indicates an intermittent pulse mode, a series of repeating rate control command pairs are generated. The intermittent pulse mode requires defining the pulse period length and duty cycle. The pulse period length is a time value, for example, 1 hour. Duty cycle is the proportion of device operating time within each pulse cycle, for example, a duty cycle of 30%. The runtime within each pulse cycle is calculated as the pulse cycle length multiplied by the duty cycle. The pause time within each pulse cycle is calculated as the pulse cycle length minus the runtime. The required pulse acceleration rate is calculated, which must ensure the optimal carbon source dosage is completed within the total runtime. The pulse acceleration rate is equal to the optimal carbon source dosage divided by the total runtime. The total runtime is equal to the number of pulse cycles within the dosage cycle multiplied by the runtime of each cycle. The generated rate control command sequence consists of multiple pairs of commands, each pair containing an on command and a off command. The on command instructs the carbon source dosage device to start operating at the pulse acceleration rate value and sets the runtime to the runtime of that pulse cycle. The off command instructs the carbon source dosage device to stop operating and sets the pause time to the pause time of that pulse cycle.These instructions are arranged chronologically, covering the entire dosing cycle. All generated rate control instructions are converted into a specific instruction format that the carbon source dosing device control unit can directly parse and execute, such as a data packet sequence containing timestamps and control values.

[0096] The process involves driving a carbon source dosing device to add carbon source to the anoxic tank of a soybean product wastewater treatment system according to the total dosing control parameters and rate control commands. Driving the carbon source dosing device is the process of physically starting its operation and initiating the dosing. Specifically, the control unit of the carbon source dosing device is activated. The control unit reads the preset total dosing control parameters and the loaded rate control command sequence from memory. The control unit initializes the cumulative dosing counter, which records the total amount of carbon source added during the current dosing cycle. The control unit enters a real-time control loop. In the real-time control loop, the control unit first checks whether the current time has reached the trigger time of a certain command in the rate control command sequence. When the trigger time of a certain start command is reached, the control unit sends a start signal and a rate setpoint to the actuator of the carbon source dosing device, for example, sending an analog signal of the corresponding frequency to the frequency converter of the metering pump, causing the device to start delivering carbon source to the dosing pipeline according to the set dosing rate. When the trigger time of a certain stop command is reached, the control unit sends a stop signal to the actuator, interrupting the carbon source delivery. Meanwhile, during equipment operation, the control unit monitors the actual added flow rate in real time through sensors, such as reading the pulse signal from the flow meter, and integrates to calculate the instantaneous cumulative addition. The control unit compares the instantaneous cumulative addition with the total addition control parameter in real time. Once the instantaneous cumulative addition reaches or exceeds the total addition control parameter, the control unit will immediately override all rate control commands, forcibly send an emergency stop signal to the actuator, and terminate the addition process to ensure that the total addition does not exceed the set value. The carbon source is transported through the addition pipeline to the designated addition point in the anoxic tank of the soybean product wastewater treatment system, located in the influent mixing area of ​​the anoxic tank. The entire addition process continues until the end of the addition cycle or the cumulative addition reaches the total addition control parameter and is terminated early. After the addition operation is completed, the control unit records the actual cumulative addition and the actual addition time, and stores this data as part of the carbon source addition history data for use in subsequent optimization cycles.

[0097] To verify the effectiveness of this method in actual production, a 30-day continuous comparative experiment was conducted at a soybean product wastewater treatment plant with a treatment capacity of 500 cubic meters per day. The plant employed the A² / O biological denitrification process and was equipped with a complete online monitoring system. During the experiment, all data used for analysis were strictly collected and recorded in real time during the production process. Online ammonia nitrogen analyzers, total nitrogen analyzers, and chemical oxygen demand analyzers at the influent end collected and recorded instantaneous values ​​every 2 hours. The daily influent concentration was represented by the arithmetic mean of all instantaneous values ​​for that day. The effluent total nitrogen concentration was also collected and recorded daily averages at the same frequency by the online total nitrogen analyzer at the effluent outlet. The experiment was divided into two phases, each lasting strictly continuous for 15 days.

[0098] The first phase operates using the plant's existing control strategy. The control logic of this strategy is as follows: based on the readings of the online nitrate nitrogen analyzer installed at the inlet of the anoxic tank, the frequency of the sodium acetate dosing pump is adjusted in real time using a PID control algorithm within the PLC. The control objective is to stabilize the nitrate nitrogen concentration at the end of the anoxic tank at 2.5 mg / L (set value). During this phase, the daily denitrification efficiency is calculated daily based on the actual daily average total nitrogen in the influent (denoted as TN_in_day) and the daily average total nitrogen in the effluent (denoted as TN_out_day), using the unified formula: Denitrification efficiency = (TN_in_day - TN_out_day) / TN_in_day × 100%. The daily operating data for this phase is shown in Table 1, the daily operating data table for the first phase (existing control strategy).

[0099] Table 1 Daily Operational Data for Phase 1 (Existing Control Strategy)

[0100] Days Daily average total nitrogen in influent (mg / L) Daily average total nitrogen in effluent (mg / L) Nitrogen removal efficiency (%) 1 152 33.1 78.2 2 138 20.7 85.0 3 165 32.3 80.4 4 121 19.7 83.7 5 178 36.7 79.4 6 142 19.3 86.4 7 158 28.6 81.9 8 135 21.1 84.4 9 148 25.8 82.6 10 167 36.1 78.4 11 128 23.0 82.0 12 172 34.0 80.2 13 141 23.0 83.7 14 155 23.3 85.0 15 130 24.2 81.4

[0101] According to statistics, the arithmetic mean of the denitrification efficiency in the first stage over 15 days was 82.5%, and the standard deviation was 6.8%. The carbon source consumption in this stage was calculated by converting the cumulative flow meter of the calibrated metering pump with the density of the sodium acetate solution. The cumulative amount of sodium acetate added over 15 days (calculated as 100%) was 4650 kg, with an average of 310 kg per day.

[0102] The second phase immediately switched to using this method for carbon source addition decision-making and control. Prior to the switch, all historical influent water quality, process operating parameters (dissolved oxygen, sludge concentration, etc.) and corresponding carbon source addition records from the past six months were imported into the digital twin model platform of this method, completing the training and calibration of the model parameters. The water quality transformation kinetics sub-model in the model is based on the ASM2d framework, and the calibrated maximum specific rate parameter for denitrification is 0.15 days. -1(20℃); Based on the abundance distribution of denitrification-related bacteria genera in the 16S rRNA gene sequencing results of historical sludge samples from the plant, the microbial community succession model established three virtual functional groups, with their maximum specific growth rates set at 1.8 days. -1 1.3 days -1 and 2.1 days -1 The risk assessment thresholds were set based on statistical analysis of long-term operating data: the threshold for the proportion of hydrophobic components was 35%, the threshold for the extracellular polymer secretion rate was 18 mg / L / hour, the threshold for the decline in the Shannon diversity index was 12%, and the threshold for the frequency of dominant genus turnover was 0.4 times per day.

[0103] In the second stage, the system automatically initiates the optimization process of this method at 00:00 every day: It collects real-time influent sequences from the previous day to the present, dissolved oxygen and sludge concentration data for each biological treatment tank, and inputs them into the calibrated digital twin model; based on the current state, it generates multiple candidate carbon source addition strategies, including different total carbon source addition amounts and a defined addition rate control sequence based on discrete time points and corresponding addition flow commands, driving the model to iteratively simulate the process state for the next 24 hours at a preset time step (e.g., 1 minute); the model outputs, under each strategy, the effluent total nitrogen concentration, microbial metabolic index quantification, and parameter values ​​reflecting the microbial community structure generated in a time series throughout the simulation cycle; based on these complete time-series data, the system calculates the predicted denitrification efficiency, denitrification foam formation risk, and interspecies competition imbalance risk for each strategy, and selects the optimal strategy according to preset rules, converting it into a control command sequence for the metering pump. The daily operating data for this stage is shown in Table 2, the daily operating data table for the second stage (method of this embodiment):

[0104] Table 2 Daily Operational Data for the Second Stage (Method of this Embodiment)

[0105] Days Daily average total nitrogen in influent (mg / L) Daily average total nitrogen in effluent (mg / L) Nitrogen removal efficiency (%) 1 159 21.4 86.5 2 131 14.2 89.2 3 170 21.8 87.2 4 126 13.1 89.6 5 145 19.9 86.3 6 163 18.3 88.8 7 134 16.1 88.0 8 176 21.9 87.6 9 139 18.6 86.6 10 168 18.5 89.0 11 122 15.6 87.2 12 154 16.9 89.0 13 141 18.8 86.7 14 166 16.6 90.0 15 129 16.1 87.5

[0106] According to statistics, the arithmetic mean of the denitrification efficiency in the second stage over 15 days was 88.3%, with a standard deviation of 3.1%. The amount of carbon source added in this stage was accurately counted by the cumulative flow of the metering pump according to the instructions executed by the daily optimization strategy. The cumulative addition over 15 days was 4275 kg, with an average of 285 kg per day.

[0107] The operational stability of the two stages is further reflected by the effluent suspended solids concentration and simulated risk value. Table 3 compares the effluent suspended solids concentrations of the two stages.

[0108] Table 3 Comparison of suspended solids concentration in effluent from the two stages

[0109] Days First stage effluent suspended solids concentration (mg / L) Second stage effluent suspended solids concentration (mg / L) 1 31.2 18.5 2 22.5 14.2 3 35.1 20.1 4 18.9 12.8 5 28.4 17.3 6 24.0 15.9 7 26.7 16.8 8 23.8 19.2 9 29.5 17.0 10 32.8 14.5 11 21.3 15.1 12 27.6 18.0 13 22.1 16.2 14 25.0 13.7 15 24.8 15.5

[0110] According to statistics, the average concentration of suspended solids in the effluent during the first stage was 25.6 mg / L, and the average concentration of suspended solids in the effluent during the second stage was 16.2 mg / L.

[0111] During the daily optimization run in the second phase, the digital twin model simulated and calculated the risk values ​​for denitrification foam formation and interspecific competition imbalance for the selected optimal strategy. Refer to Table 4 for the daily monitoring data of risk values ​​in the second phase. Both risk values ​​are calculated using an integer scale of 0-4.

[0112] Table 4 Daily Monitoring Data of Risk Values ​​in Phase II

[0113] Days Risk value of denitrification foam formation Risk value of interspecific competition imbalance 1 1 1 2 2 1 3 1 0 4 0 1 5 1 2 6 1 0 7 0 1 8 2 1 9 1 0 10 1 1 11 1 1 12 0 1 13 2 2 14 1 0 15 1 1

[0114] According to statistics, the average risk value for the formation of denitrification foam in the second stage was 1.1, and the average risk value for the imbalance of interspecific competition was 0.9. The on-site operation log also recorded that the visual observation level of foam on the surface of the biological treatment tank in the second stage was significantly reduced, and the phenomenon of sludge floating in the secondary sedimentation tank basically disappeared.

[0115] Specifically, on day 22 of the experiment (the second phase), the influent chemical oxygen demand (COD) experienced a pulsed increase between 8:00 and 10:00 AM, with a peak instantaneous increase of 2050 mg / L. The digital twin model of this method, when generating the optimized strategy in the early morning, had already predicted the change in carbon source demand patterns based on the influent water quality trend. The generated strategy was "continuous low-flow dosing in the early stage, followed by enhanced dosing based on simulated nitrate accumulation in the mid-term." In actual operation, the effluent total nitrogen concentration slowly increased from 11.5 mg / L after the shock, reaching a peak of 16.1 mg / L at 4:00 PM, which did not exceed the real-time monitoring alarm limit of 15 mg / L. It then gradually decreased, returning to a stable level below 12.0 mg / L by 4:00 AM the following day. Compared with a similar shock event recorded in the first phase historical log (day 7), the total nitrogen in the effluent exceeded the standard (>15 mg / L) 4 hours after the shock and continued to exceed the standard for about 36 hours, reaching a maximum of 22.5 mg / L, accompanied by a large amount of foam and visible floating sludge.

[0116] This comparative experiment fully presents the entire process of calculating denitrification efficiency, statistically analyzing material consumption, and recording risk values ​​based on actual production monitoring data. The data shows that, under the same influent water quality conditions, the mean denitrification efficiency calculated from actual monitoring data using this method is improved, and its volatility (standard deviation) is significantly reduced; while achieving better effluent water quality, carbon source consumption is reduced; both model evaluation and on-site observation confirm that secondary process risks are effectively controlled; and the system's resilience to water quality shocks is enhanced. All of the above data and results are derived from data collection and calculations during the actual production process.

[0117] Figure 3The following graphs illustrate the changes in denitrification efficiency, foam risk, and competition imbalance indices over operating time in embodiments of the present invention. These graphs correspond to process monitoring data from a continuous 12-hour stable treatment of soybean product wastewater using the process optimization method of the present invention. The blue curve represents denitrification efficiency, the red curve represents foam risk values ​​(grades 0-4, with 0 representing no risk), and the green dashed line represents the threshold line for bacterial community competition imbalance (grade 1, corresponding to the critical value for bacterial community balance). During the 12-hour operating cycle, the denitrification efficiency steadily increased with operating time, initially reaching approximately 71%, and then remaining stable in the 75%-77% range during the later stages of operation without significant fluctuations. The foam risk value remained at the risk-free level (grade 0) throughout the process, and the bacterial community competition imbalance also remained stable near the threshold line without exceeding the safe range.

[0118] The above monitoring results verified the implementation effect. Through precise carbon source addition and dynamic process control strategy, the denitrification efficiency can be steadily improved during continuous operation. At the same time, it effectively suppresses the risk of foam growth and maintains the balance of nitrifying and denitrifying bacteria. It solves the problems of difficulty in improving denitrification efficiency and poor process stability in the existing technology, and provides a reliable guarantee for the long-term efficient operation of the soybean product wastewater treatment system, fully demonstrating its practicality and technical advantages.

[0119] Example 2: Figure 2 A schematic diagram of a digital twin-based wastewater treatment process optimization system for soybean products is provided. This system includes the following modules:

[0120] The data acquisition module is used to acquire real-time operating data of the soybean product wastewater treatment system;

[0121] The model input module is used to input real-time running data into the digital twin model;

[0122] The strategy simulation module is used to set several candidate carbon source addition strategies based on real-time operation data. Each candidate carbon source addition strategy includes the candidate carbon source addition amount and the candidate carbon source addition acceleration rate mode. The digital twin model is used to simulate the denitrification efficiency value and the denitrification foam formation risk value under each strategy, and to establish the mapping relationship between the candidate carbon source addition strategy and the denitrification efficiency value and the denitrification foam formation risk value.

[0123] The risk assessment module is used to assess the risk value of interspecies competition imbalance caused by changes in the internal distribution structure of denitrifying bacteria due to the candidate carbon source addition rate mode, based on the simulation process of the digital twin model for each candidate carbon source addition strategy.

[0124] The strategy optimization module is used to select the optimal carbon source addition strategy from the candidate carbon source addition strategies based on the mapping relationship and the risk value of interspecific competition imbalance. The strategy meets the preset denitrification efficiency requirements and has the best comprehensive risk value of denitrification foam formation and interspecific competition imbalance.

[0125] The carbon source addition module is used to execute carbon source addition operations according to the optimized carbon source addition strategy.

[0126] All calculations involved in the embodiments are dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0127] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, in the form of a computer program product.

[0128] Those skilled in the art will recognize that the modules and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and inventive constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0129] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0130] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0131] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application.

[0132] In conclusion, the above are merely preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A method for optimizing the treatment process of soybean product wastewater based on digital twins, characterized in that, Includes the following steps: S1: Obtain real-time operating data of the soybean product wastewater treatment system; S2: Input real-time operational data into the digital twin model; S3: Based on real-time operational data, several candidate carbon source addition strategies are set. Each candidate carbon source addition strategy includes the candidate carbon source addition amount and the candidate carbon source addition acceleration rate mode. A digital twin model is used to simulate the denitrification efficiency value and the denitrification foam formation risk value under each strategy, and a mapping relationship is established between the candidate carbon source addition strategy and the denitrification efficiency value and the denitrification foam formation risk value, including: Based on real-time influent parameters and historical carbon source addition data, several candidate carbon source addition strategies are generated, including different candidate carbon source addition amounts and different candidate carbon source addition acceleration rate modes. Each candidate carbon source addition strategy is input into the digital twin model, which then drives the digital twin model to simulate and execute the biochemical reaction process under the candidate carbon source addition strategy. The simulated total nitrogen concentration of the effluent output by the digital twin model is obtained to calculate the denitrification efficiency value, and the simulated microbial metabolic indicators are obtained to assess the risk value of denitrification foam formation. Each candidate carbon source addition strategy is associated with its corresponding denitrification efficiency value and denitrification foam formation risk value to form a mapping relationship. S4: For each candidate carbon source addition strategy, based on the simulation process of the digital twin model, evaluate the risk value of interspecies competition imbalance corresponding to the changes in the internal distribution structure of the denitrifying microbial community caused by the candidate carbon source addition acceleration mode, including: For each candidate carbon source addition strategy, obtain the denitrifying functional microbial community simulation data output by the digital twin model during the simulation of the biochemical reaction under that candidate carbon source addition strategy; Based on simulation data of denitrifying functional bacteria, we analyzed the relative abundance and metabolic activity trends of different denitrifying bacteria genera during the simulation period. Based on the trends of relative abundance and metabolic activity, an index characterizing the stability of the internal distribution structure of denitrifying bacteria was calculated as a risk value for interspecific competition imbalance. The stability indicators of the internal distribution structure of denitrifying bacteria, on which the risk value of interspecific competition imbalance is based, include the decline of the denitrifying bacteria genus diversity index and the frequency of dominant genus replacement during the simulation period. S5: Based on the mapping relationship and the risk value of interspecific competition imbalance, select the optimal carbon source addition strategy from the candidate carbon source addition strategies that meets the preset denitrification efficiency requirements and has the best comprehensive risk value of denitrification foam formation and interspecific competition imbalance. S6: Perform carbon source addition operation according to the optimized carbon source addition strategy.

2. The method for optimizing the treatment process of soybean product wastewater based on digital twins according to claim 1, characterized in that, S1 includes: Obtain real-time influent parameters at the inlet of the soybean product wastewater treatment system; Obtain real-time status parameters of the biochemical reaction tank within the soybean product wastewater treatment system; Obtain historical data on carbon source addition for the soybean product wastewater treatment system.

3. The method for optimizing the treatment process of soybean product wastewater based on digital twins according to claim 1, characterized in that, S2 include: Input the real-time influent parameters into the influent water quality parameter interface of the digital twin model; Input real-time status parameters into the biochemical reaction process parameter interface of the digital twin model; Input historical carbon source addition data into the carbon source addition record interface of the digital twin model.

4. The method for optimizing the treatment process of soybean product wastewater based on digital twins according to claim 3, characterized in that, Digital twin models include: A water quality transformation kinetics sub-model is used to simulate the ammonia nitrogen transformation, nitrification and denitrification reaction processes based on real-time influent parameters and historical carbon source addition data. Microbial community succession model is used to simulate the growth, metabolism, and population structure dynamics of denitrifying functional microbial communities based on real-time state parameters; The water quality transformation kinetics sub-model and the microbial community succession sub-model are coupled together. The microbial metabolic activity parameters output by the microbial community succession sub-model serve as the regulatory factors for the denitrification reaction rate in the water quality transformation kinetics sub-model.

5. The method for optimizing the treatment process of soybean product wastewater based on digital twins according to claim 1, characterized in that, The microbial metabolic indicators on which the risk value of denitrification foam formation is assessed include the proportion of hydrophobic components and the secretion rate in the simulated extracellular polymer.

6. The method for optimizing the treatment process of soybean product wastewater based on digital twins according to claim 1, characterized in that, S5 include: Based on the mapping relationship, candidate carbon source addition strategies that meet the preset denitrification efficiency requirements are selected from all candidate carbon source addition strategies; For each candidate carbon source addition strategy selected, the corresponding risk value of denitrification foam formation and the risk value of interspecific competition imbalance are weighted and calculated to obtain a comprehensive risk assessment value. From the selected candidate carbon source addition strategies, the one with the lowest comprehensive risk assessment value is chosen as the optimized carbon source addition strategy.

7. The method for optimizing the treatment process of soybean product wastewater based on digital twins according to claim 1, characterized in that, S6 include: Extract the optimal carbon source addition amount and optimal carbon source addition rate mode from the optimized carbon source addition strategy; Set the total carbon source addition control parameters for the carbon source addition equipment based on the optimized carbon source addition amount; Based on the optimized carbon source dosing rate mode, a rate control command is generated for the carbon source dosing equipment during the dosing cycle. According to the total amount control parameters and rate control instructions, the carbon source dosing equipment is driven to add carbon source to the anoxic tank of the soybean product wastewater treatment system.

8. A digital twin-based wastewater treatment process optimization system for soybean products, used to implement the digital twin-based wastewater treatment process optimization method for soybean products as described in any one of claims 1-7, characterized in that, Includes the following modules: The data acquisition module is used to acquire real-time operating data of the soybean product wastewater treatment system; The model input module is used to input real-time running data into the digital twin model; The strategy simulation module is used to set several candidate carbon source addition strategies based on real-time operation data. Each candidate carbon source addition strategy includes the candidate carbon source addition amount and the candidate carbon source addition acceleration rate mode. The digital twin model is used to simulate the denitrification efficiency value and the denitrification foam formation risk value under each strategy, and to establish the mapping relationship between the candidate carbon source addition strategy and the denitrification efficiency value and the denitrification foam formation risk value. The risk assessment module is used to assess the risk value of interspecies competition imbalance corresponding to the changes in the internal distribution structure of denitrifying bacteria caused by the candidate carbon source addition rate mode, based on the simulation process of the digital twin model for each candidate carbon source addition strategy. The strategy optimization module is used to select the optimal carbon source addition strategy from the candidate carbon source addition strategies based on the mapping relationship and the risk value of interspecific competition imbalance. The strategy meets the preset denitrification efficiency requirements and has the best comprehensive risk value of denitrification foam formation and interspecific competition imbalance. The carbon source addition module is used to execute carbon source addition operations according to the optimized carbon source addition strategy.

Images