ASM-based sewage treatment engineering general model intelligent control method and system

Abstract

The invention relates to the technical field of sewage engineering control, in particular to an ASM-based sewage treatment engineering general model intelligent control method and system. According to the method, the ASM model and the sludge growth kinetic model are fused, a general sewage treatment engineering model combining physical mechanism driving and data driving is constructed, the problem of insufficient applicability of a traditional model is solved, and the model has process universality and can dynamically adapt to specific sewage plant working conditions; the universality and scene adaptation capability of the model are remarkably enhanced, closed-loop adaptive optimization of the universal sewage treatment engineering model is achieved through the error monitoring and parameter correction feedback module, the problem that a traditional static model cannot conduct self-correction according to real-time operation deviation is solved, and the real-time operation deviation of the model is improved. The prediction accuracy and the control effect of the model can be continuously improved, and stable and efficient operation of the system under complex dynamic working conditions is ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of sewage engineering control, and particularly to an intelligent control method and system for a general model of sewage treatment engineering based on ASM. Background Art

[0002] Currently, the process control of the sewage biological treatment process in sewage treatment plants mainly relies on a series of control strategies based on the Activated Sludge Models (ASM). These models are proposed by the International Water Association (IWA), such as ASM1, ASM2, and ASM3. They depict the removal process of pollutants (such as COD, ammonia nitrogen, nitrate, total nitrogen, total phosphorus, etc.) by microbial populations through mathematical equations, and describe the biochemical reaction mechanisms of bacterial growth, substrate consumption, nitrification, denitrification, phosphorus release, etc. Existing control methods generally perform offline process optimization based on ASM simulation results when designing or reconstructing sewage treatment plants, or combine methods such as traditional PID control and threshold control based on empirical rules to adjust process parameters such as aeration volume, sludge return ratio, and carbon source dosing to achieve the purpose of meeting the effluent standards and energy conservation.

[0003] However, the problem with the existing technology is that the ASM model is essentially a simplified model developed for teaching and theoretical research. Its default parameters (such as maximum specific growth rate, half-saturation constant, oxygen limitation coefficient, etc.) are based on idealized experimental conditions and there are significant deviations from the actual working conditions of specific sewage treatment plants (including influent water quality fluctuations, temperature changes, sludge age fluctuations, etc.), resulting in a large number of parameter calibrations and adjustments required for the model in engineering applications. Secondly, existing solutions often regard the ASM model as a static prediction tool and lack the ability to link with real-time data, and cannot adaptively adjust to the nonlinear, time-varying, and disturbance characteristics in the process. Thirdly, traditional control methods (such as PID control) have a lag in response to complex coupled biochemical processes and insufficient adjustment accuracy, and it is difficult to cope with sudden load shocks or multi-objective (such as nitrogen removal, phosphorus removal, energy conservation) optimization requirements. Summary of the Invention

[0004] The present invention proposes an intelligent control method and system for a general model of sewage treatment engineering based on ASM, aiming to construct a general control method and control system applicable to sewage treatment engineering based on the ASM model by integrating the ASM model with the sludge growth kinetics model, introducing a nonlinear compensation algorithm and working condition parameter correction, and realizing multi-source control of sewage treatment engineering.

[0005] Among them, an intelligent control method for a general model of sewage treatment engineering based on ASM includes the following steps: S1. Obtain the multi-source operating condition data of the sewage treatment plant, preprocess the multi-source operating condition data, and construct a multi-source operating condition data matrix related to the sewage treatment plant; S2. According to the multi-source operating condition data matrix, through state variable mapping, respectively extract the reaction rate functions and state variable sets corresponding to the multi-source operating condition data matrix in the ASM model and the sludge growth kinetics model, construct a generalized sewage treatment engineering model, and output a state prediction matrix based on the multi-source operating condition data; S3. According to the state prediction matrix, construct a weighted multi-objective control model, and output a sewage treatment control quantity sequence through rolling horizon model predictive control; S4. According to the sewage treatment control quantity sequence, perform real-time control on the sewage treatment system to obtain the collected actual operating condition response data, construct an actual operating condition response data matrix, and calculate a prediction error matrix according to the state prediction matrix; S5. According to the prediction error matrix, correct the parameters of the generalized sewage treatment engineering model, output a parameter correction matrix and return to step S3 to perform closed-loop control on the generalized sewage treatment engineering model.

[0006] An intelligent control system for a generalized sewage treatment engineering model based on ASM, which is implemented based on any one of the above-mentioned intelligent control methods for a generalized sewage treatment engineering model based on ASM, and is characterized by including: A multi-source data acquisition module, which is used to collect the multi-source operating condition data of the sewage treatment plant, preprocess the multi-source operating condition data, and construct a multi-source operating condition data matrix related to the sewage treatment plant; An engineering model construction and state prediction module, which is used to respectively extract the reaction rate functions and state variable sets corresponding to the multi-source operating condition data matrix in the ASM model and the sludge growth kinetics model according to the multi-source operating condition data matrix through state variable mapping, construct a generalized sewage treatment engineering model, and output a state prediction matrix based on the multi-source operating condition data; A control strategy generation module, which is used to construct a weighted multi-objective control model according to the state prediction matrix, and output a sewage treatment control quantity sequence through rolling horizon model predictive control; A control strategy error monitoring module, which is used to perform real-time control on the sewage treatment system according to the sewage treatment control quantity sequence to obtain the collected actual operating condition response data, construct an actual operating condition response data matrix, and calculate a prediction error matrix according to the state prediction matrix; A parameter correction and feedback module, which is used to correct the parameters of the generalized sewage treatment engineering model according to the prediction error matrix, and output a parameter correction matrix to perform closed-loop control on the generalized sewage treatment engineering model.

[0007] A computer-readable storage medium for storing a computer program, which, when running on a computer, causes the computer to execute an intelligent control method for a general model of sewage treatment engineering based on ASM as described in any one of the above.

[0008] An electronic device, comprising: A memory for storing a computer program; A processor for executing the computer program to implement an intelligent control method for a general model of sewage treatment engineering based on ASM as described in any one of the above.

[0009] The beneficial effects of the present invention are: (1) By deeply integrating the ASM model with the sludge growth kinetics model, the present invention constructs a general sewage treatment engineering model combining physical mechanism-driven and data-driven, overcoming the problem of insufficient applicability caused by the traditional model's single dependence on empirical formulas or single data fitting, making the model have both process universality and the ability to dynamically adapt to the specific conditions of sewage treatment plants, and significantly enhancing the universality and scenario adaptation ability of the model.

[0010] (2) Through a hierarchical data preprocessing process such as multi-source data collection, missing value repair, outlier removal, and non-linear normalization, the present invention establishes a high-quality and highly consistent data input system, providing solid data support for model fusion, effectively avoiding prediction biases introduced by multi-source data spurs, noise, and anomalies, and laying a reliable foundation for subsequent prediction and control. BRIEF DESCRIPTION OF THE DRAWINGS

[0011] Figure 1 It is a flowchart of an intelligent control method for a general model of sewage treatment engineering based on ASM proposed by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0012] The technical solution of the present invention will be further described in detail below with reference to the drawings, but the protection scope of the present invention is not limited to the following.

[0013] In order to make the purpose, technical solution and advantages of the present invention clearer, the present invention will be further described in detail with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the described embodiments are only a part of the embodiments of the present invention, rather than all of the embodiments. Usually, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.

[0014] Accordingly, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely represents selected embodiments of the present invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without creative efforts fall within the scope of protection of the present invention. It should be noted that relational terms such as "first" and "second" are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations.

[0015] Moreover, the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, article or apparatus comprising a series of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article or apparatus. Without further limitation, an element defined by the statement "comprising a..." does not exclude the presence of additional identical elements in the process, method, article or apparatus comprising the element.

[0016] The features and performance of the present invention will be further described in detail below in conjunction with embodiments.

[0017] Embodiment 1 As Figure 1 shown, the embodiment of the present invention provides an intelligent control method for a general model of sewage treatment engineering based on ASM, including the following steps: S1. Obtain multi-source operating condition data of the sewage treatment plant, preprocess the multi-source operating condition data, and construct a multi-source operating condition data matrix related to the sewage treatment plant; S2. According to the multi-source operating condition data matrix, through state variable mapping, respectively extract the reaction rate functions and state variable sets corresponding to the multi-source operating condition data matrix in the ASM model and the sludge growth kinetics model, construct a general sewage treatment engineering model, and output a state prediction matrix based on the multi-source operating condition data; S3. According to the state prediction matrix, construct a weighted multi-objective control model, and output a sewage treatment control quantity sequence through rolling horizon model predictive control; S4. According to the sewage treatment control quantity sequence, perform real-time control on the sewage treatment system to obtain collected actual operating condition response data, construct an actual operating condition response data matrix, and calculate a prediction error matrix according to the state prediction matrix; S5. According to the prediction error matrix, correct the parameters of the general sewage treatment engineering model, output a parameter correction matrix and return to step S3 to perform closed-loop control on the general sewage treatment engineering model.

[0018] Further, step S1 specifically includes the following sub-steps: S101. Collect the influent pollutant concentration, effluent pollutant concentration, aeration volume, sludge retention time SRT, reflux ratio, PH value, and water temperature from the sewage treatment plant, and construct an original data matrix based on the collected data; S102. Extract spectral features by constructing a local linear trend model within a time window, keep the boundary observations unchanged, perform linear weighted calculation using the known data points on the left and right, and interpolate and repair the boundaries of the missing values in the original data matrix; S103. By setting a sliding time window with a fixed length, perform mean aggregation within the window for the original data with different sampling frequencies, and uniformly output a multi-source working condition data matrix with consistent time steps.

[0019] Specifically, the implementation principle processes of the above sub-steps are as follows: In step S101, through multi-source data collection, collect the influent water quality (COD, NH4⁺, TN, etc.), effluent indicators, aeration volume, reflux ratio, sludge concentration (MLSS), and working condition data such as air temperature and pH of the sewage treatment plant with the same sampling time length and variable dimension, and construct a multi-source working condition data matrix . To ensure the consistency of data sources, it is necessary to use the same sampling time length, and the variable dimensions of each working condition data are kept consistent. Exemplarily, collect data through the SCADA system of the sewage treatment plant, and combine it with an online water quality monitor with a unified data format standard and a historical record of manual experimental records in a template to construct a multi-source working condition data matrix of the sewage treatment plant .

[0020] In step S102, perform normalization processing on the missing items and the sampling frequency differences of different data sources in the multi-source working condition data matrix. By using the theory of boundary-preserving linear interpolation repair, repair the missing values in the multi-source data matrix, expressed as: , where represents the missing value interpolated at time point , , represent the time points corresponding to the left and right ends of the missing value, , represent the known values corresponding to the left and right ends of the time points of the missing value. For multi-source working condition data with inconsistent sampling periods, use a sliding window for resampling, expressed as: , where represents the resampled average value of variable i within the window period, represents the sliding window length, represents the observed value of variable i at time point k. Obtain a multi-source working condition data matrix with interpolation repair and sampling alignment.

[0021] In step S103, for the multi-source working condition data matrix with interpolation repair and sampling alignment, the sliding Z-score anomaly detection method is used to identify anomaly points. The data mean and standard deviation are calculated respectively. According to the Z-score anomaly detection, the outliers in the multi-source working condition data are extracted and the outliers are replaced to obtain a multi-source working condition data matrix with a consistent unified output time step.

[0022] Further, the step S2 specifically includes the following sub-steps: S201. According to the multi-source working condition data matrix, extract the organic matter degradation and sludge growth rate calculation sub-model and the ASM model state variable set in the ASM model, and according to the multi-source working condition data matrix, extract the environmental modulation function and the sludge kinetic model state variable set in the sludge kinetic model. By unifying the saturated substrate concentration term and the dissolved oxygen regulation function, the function-level fusion of the organic matter degradation and sludge growth rate calculation sub-model and the environmental modulation function is carried out to construct a total rate calculation model containing a dual regulation mechanism; S202. Introduce an environmental impact adjustment factor to adaptively compensate the total rate calculation model and dynamically adjust the biochemical reaction intensity; S203. Extract and construct a variable exchange mapping matrix based on the spectral characteristics of the multi-source working condition data matrix, and fuse the ASM model state variable set and the sludge kinetic model state variable set through the variable exchange mapping matrix to construct a unified state description space; S204. Through the error feedback backpropagation mechanism, use the collected error term to adaptively learn the unified state description space to obtain a generalized sewage treatment engineering model, and output a state prediction matrix based on the multi-source working condition data.

[0023] Specifically, the implementation principle processes of the above sub-steps are as follows: In step S201, first, according to the multi-source working condition data matrix, extract the ASM model reaction rate formula: , where represents the maximum specific growth rate of heterotrophic bacteria, represents the soluble substrate (BOD) concentration at time t, represents the half-saturation substrate constant, represents the dissolved oxygen concentration at time t, represents the oxygen half-saturation constant, represents the heterotrophic bacteria biomass concentration; extract the ASM model state variable set: , where, represents the soluble substrate concentration, represents the heterotrophic bacteria concentration, represents the autotrophic nitrifying bacteria concentration, represents the concentration of polyphosphate-accumulating organisms, represents the concentration of nitrate nitrogen. Then, extract the environmental regulation function in the sludge growth kinetic model: , where represents the maximum specific growth rate of theoretical sludge, represents the environmental sensitivity factor function; extract the state variable set of the sludge kinetic model: , where, represents the mixed liquor suspended solids concentration, represents the mixed liquor volatile suspended solids concentration, represents the sludge age, represents the food-to-microorganism load ratio.

[0024] Specifically, the ASM (Activated Sludge Model) proposed by the International Water Association (IWA) is a set of classic mathematical description systems for the activated sludge process. It originated from the ASM1 model launched by the IWA (then called IAWPRC) in 1987, and subsequent extended versions such as ASM2, ASM2d, and ASM3 have been developed. The ASM model is mainly constructed based on long-term operation data of sewage treatment. The core is to use a set of metabolic reaction processes (including organic matter degradation, nitrogen and phosphorus transformation, nitrification, denitrification, etc.) and related biokinetic rate equations to describe the behavior of the microbial population in the mixed liquor suspended solids (MLSS). The construction of this model requires defining a series of basic parameters, such as the concentration of soluble biodegradable organic matter, particulate organic matter concentration, dissolved oxygen concentration, ammonia nitrogen concentration, nitrite and nitrate concentration, concentration of active heterotrophic bacteria, concentration of active autotrophic bacteria, endogenous residue concentration, etc., as well as the corresponding kinetic parameters (such as the maximum specific growth rate μmax, half-saturation constant Ks, endogenous respiration coefficient bH, ammonia oxidation rate μA, denitrification rate μD, etc.). The core mechanism of the ASM model is to divide the material flow and energy flow in the sewage treatment process into a series of components and perform coupled simulation through mass conservation, rate equations, and stoichiometric relationships, which can comprehensively characterize complex biological reaction processes.

[0025] The sludge kinetic growth model is mainly derived from the classical Monod kinetics, bioreactor theory, and sludge retention time (SRT) control theory. By defining the growth rate, decay rate, endogenous respiration, and substrate consumption kinetics of microorganisms, this model describes the evolution laws of sludge volume, active components, and biomass over time. Usually, parameters such as sludge concentration (MLSS, MLVSS), substrate concentrations of influent and effluent (such as COD, BOD, ammonia nitrogen), sludge yield coefficient (Y), maximum specific growth rate (μmax), half-saturation constant (Ks), endogenous respiration coefficient (b), and sludge settling performance (such as SVI) need to be collected for model construction. By substituting these parameters into the Monod equation and mass balance equation, a dynamic sludge growth and decay equation can be established to predict and regulate the quantity and activity level of the microbial population in the bioreactor. This model is not only used to predict the treatment capacity of the activated sludge system but also often serves as an important tool for optimizing the sewage treatment process, aeration control, predicting the production of excess sludge, and energy consumption analysis. After being combined with the ASM model, it can achieve comprehensive coupling analysis of substrate-microorganism-environment in the engineering model.

[0026] In step S202, an environmental impact adjustment factor is introduced , to perform adaptive compensation on the total rate calculation model. The environmental sensitivity factor function is used to compound the temperature adjustment, dissolved oxygen adjustment, and sludge adjustment terms and embed them as a scaling regulator of the biochemical rate into the total rate calculation model. In step S203, an ASM model state variable set and a sludge kinetic model state variable set are constructed respectively according to the ASM model and the sludge growth kinetic model, and variable interchange mapping is performed. According to the mapping relationship, the two sets are constructed into a unified state description space. In step S204, through the error feedback backpropagation mechanism, the unified state description space is adaptively learned using the collected error terms to obtain a generalized sewage treatment engineering model, and a state prediction matrix is output based on multi-source operating condition data.

[0027] Specifically, the specific implementation principle process of step S204 is as follows: First, based on the state variable parameters obtained by fusing the ASM model state variable set and the sludge kinetic model state variable set in the unified state description space, at the beginning of each control period (such as every 30 minutes) according to the control period of the sewage treatment plant, the latest operating condition data of the corresponding state variable parameters of the sewage treatment plant are collected through multi-source data collection , and at the same time, combined with the actual operating condition response data in the previous control period , perform a forward propagation based on the generalized sewage treatment engineering model of the current control cycle version to obtain the state prediction value within the current control cycle ; according to the state prediction value within the control cycle at the current moment t and the actual working condition response data of the control cycle at the current moment t are compared to calculate the prediction error value of the current control cycle, which is expressed as , where represents the prediction error value within the control cycle at the current moment t.

[0028] Subsequently, based on the prediction error value, construct an error loss function , where represents the state variable parameter in the unified state description space, represents the -th weight factor of the state variable parameter, represents the -th prediction error value of the state variable parameter within the current control cycle; through the backpropagation mechanism, is differentiated with respect to to obtain the gradient, and an optimization algorithm (such as the Adam algorithm or the L-BFGS algorithm) is used to perform gradient update to obtain the updated state variable parameter .

[0029] Then, according to the updated state variable parameter, combine the unified state description space, the total rate calculation model, and the error loss function to obtain the generalized sewage treatment engineering model, and update the state variable parameter of the generalized sewage treatment engineering model to output the state prediction matrix of the next control cycle .

[0030] Finally, according to the state prediction matrix of the next control cycle , as the data input for step S3, provide an accurate state prior prediction for generating the sewage treatment control quantity sequence. In this way, an error-driven adaptive model construction, self-optimizing prediction, and feedback closed-loop mechanism are realized, and the system has good generalization ability, fast adaptation ability to sudden working conditions, as well as stability and convergence in long-term operation.

[0031] Further, the total rate calculation model is specifically expressed as: where represents the actual biological reaction rate at time t, represents the maximum specific growth rate, represents the soluble substrate concentration, represents the half-saturation substrate constant, represents the dissolved oxygen concentration at time t, represents the oxygen half-saturation constant, represents the environmental sensitivity factor function.

[0032] Furthermore, in the step S203, the unified state description space is specifically expressed as: ; wherein, represents the target variable vector, represents the variable mapping matrix, represents the ASM model state variable set, represents the sludge kinetics model state variable set.

[0033] Specifically, the construction principle of the variable mapping matrix is as follows: First, based on the multi-source operating condition data matrix , perform time series spectrum analysis to obtain the response frequency band and collaborative characteristics between the ASM model state variable set and the sludge kinetics model state variable set . Determine the high-correlation mapping channels between the two state variable sets through mutual information analysis and principal component correlation ranking; subsequently, based on the physical meanings and time series coupling behaviors of the ASM model state variable set and the sludge kinetics model state variable set , construct an initial variable exchange rule table, where the physical meaning refers to that a certain state variable in the two state variable sets is either a variable reflecting the organic matter concentration, or a variable reflecting the biomass, or a variable reflecting the oxygen concentration; construct a set of candidate mapping function families using logical mapping and unit conversion factors, including at least linear mapping functions, exponential mapping functions, and piecewise mapping functions. Then, through the minimum state prediction error criterion, use the quasi-Newton type gradient optimization algorithm to train and calibrate the parameters in these function families, and finally converge to obtain the optimal variable mapping structure, that is, form the variable mapping matrix ; the variable mapping matrix is essentially a weight association matrix, and its element values represent the mapping intensity or proportion of a certain state variable in the ASM model state variable set to the sludge kinetics model state variable set for the target unified variable. Finally, the variable mapping matrix serves as the basic structure of variable mapping, and performs linear combination and transformation on the ASM model state variable set and the sludge kinetics model state variable set during the model fusion operation process, so as to construct a unified state description space and achieve the structural consistency and state controllability of the ASM model and the sludge kinetics model.

[0034] By introducing a dual regulation rate mechanism, multiple factors such as the heterotrophic bacteria biomass concentration, soluble substrate concentration, dissolved oxygen concentration, and the maximum specific growth rate of heterotrophic bacteria are integrated into the rate calculation model, and the unification of the state set is achieved through a variable mapping matrix, which can finely depict the complex multiphase reaction mechanism and overcome the technical bottleneck that a single model is difficult to capture the internal multivariable coupling relationship of the system.

[0035] Furthermore, the specific steps of step S3 include the following sub-steps: S301. According to the state prediction matrix, a weighted multi-objective control model is constructed with the goals of minimizing the effluent pollution concentration, minimizing the operating energy consumption, and balancing the sludge load stability respectively; S302. Set the initial variable sequence, use the initial variable sequence as the minimum feasible solution space, and adopt the rolling horizon model predictive control to calculate the optimal control variable sequence within each control period as the sewage treatment control quantity sequence.

[0036] Furthermore, in step S301, constructing a weighted multi-objective control model with the goals of minimizing the effluent pollution concentration, minimizing the operating energy consumption, and balancing the sludge load stability respectively includes the following sub-steps: S3011. With the goal of minimizing the effluent pollution concentration, calculate the time cumulative value of the L2 norm error between the predicted effluent pollutant value and the target emission value, which is specifically expressed as: ; Among them, represents the pollutant emission deviation from the target penalty term, represents the predicted effluent pollutant concentration vector at time t, represents the target effluent discharge standard concentration vector, represents the square of the Euclidean norm; S3012. With the goal of minimizing the operating energy consumption, calculate the cumulative value of the power consumption of the aeration equipment and the sludge pump, which is specifically expressed as: ; Among them, represents the operating energy consumption deviation from the target penalty term, represents the power consumption of the aeration equipment at time t at time t, represents the power consumption of the sludge pump at time t at time t; S3013. With the goal of balancing the sludge load stability, calculate the approximation value of the unit sludge load and the design load target of the sewage treatment plant, which is specifically expressed as: ; Among them, represents the penalty term for sludge load stability deviation from the target, represents the suspended solid concentration of the mixed solution at time t, represents the average sludge residence time at time t, represents the target ratio; S3014. According to the calculated time accumulated value, power consumption accumulated value and approach value, a weighted multi-objective control model is constructed, which is specifically expressed as: ; in, represents the weighted multi-objective control model, , , They respectively represent the contribution of effluent pollution concentration to the weighted multi-objective control model, the contribution of operating energy consumption to the weighted multi-objective control model, and the contribution of sludge load stability to the weighted multi-objective control model. Based on the weighted multi-objective control model, it is possible to take into account both energy consumption minimization and system load stability while ensuring that the effluent quality meets the standard, overcoming the shortcomings of traditional control methods that focus on a single objective and lack comprehensive trade-offs, and providing a multi-objective collaborative optimization solution for the sewage treatment process.

[0037] Furthermore, the step S4 specifically includes the following sub-steps: S401. According to the sewage treatment control quantity sequence, the multi-source operating condition data are respectively controlled in real time, and the actual operating condition response data are collected in real time according to the fixed sampling interval to construct the actual operating condition response data matrix; S402. Through timestamp matching and variational dynamic time warping algorithm, the actual working condition response data matrix and the state prediction matrix are aligned to calculate the prediction error matrix.

[0038] Furthermore, the step S5 specifically includes the following sub-steps: S501. According to the prediction error matrix, sensitivity analysis is performed on the parameters of the weighted multi-objective control model, and a priority update parameter subset is obtained by calculating the partial derivatives of the error matrix and the corresponding parameters; S502. According to the priority update parameter subset, a multi-scale coordination factor is introduced, and the priority update parameter subset is updated in combination with the model complexity constraint; S503. Output a parameter correction matrix according to the updated priority update parameter subset, and feed the parameter correction matrix back to step S3.

[0039] Furthermore, the step S501 specifically includes the following sub-steps: S5011. According to the prediction error matrix, calculate the norm of the prediction error matrix at the current moment; S5012. Calculate the sensitivity partial derivatives of all parameters in the weighted multi-objective control model respectively according to the prediction error matrix norm; S5013. Set a sensitivity threshold based on the historical multi-source operating condition data of the sewage treatment plant, sort the sensitivity partial derivatives of all parameters, and when the sensitivity partial derivative is higher than the sensitivity threshold, it is used as a high-sensitivity parameter to obtain a subset of parameters to be updated preferentially.

[0040] Further, the step S502 specifically includes the following sub-steps: S5021. Set a basic learning rate according to the dynamic learning strategy; S5022. Perform spectral decomposition on the prediction error matrix, and calculate the frequency error components and normalized energy weights of all parameters in the prediction error matrix respectively; S5023. According to the subset of parameters to be updated preferentially, calculate the correction amount of each parameter in the subset under the frequency error component, and merge the correction amounts of each parameter into the final update amount.

[0041] Through error monitoring and parameter correction feedback, the closed-loop adaptive optimization of the general sewage treatment engineering model is realized, overcoming the problem that the traditional static model cannot self-correct according to real-time operation deviations, and being able to continuously improve the prediction accuracy and control effect of the model, ensuring the stable and efficient operation of the system under complex dynamic operating conditions.

[0042] Embodiment 2 As a preferred implementation manner of the above embodiment, an intelligent control system for a general sewage treatment engineering model based on ASM is proposed. This system is implemented based on the intelligent control method for a general sewage treatment engineering model based on ASM described in any one of the above, and includes: A multi-source data acquisition module, which is used to acquire the multi-source operating condition data of the sewage treatment plant, preprocess the multi-source operating condition data, and construct a multi-source operating condition data matrix related to the sewage treatment plant; An engineering model construction and state prediction module, which is used to respectively extract the reaction rate function and state variable set corresponding to the multi-source operating condition data matrix in the ASM model and the sludge growth kinetics model according to the multi-source operating condition data matrix through state variable mapping, construct a general sewage treatment engineering model, and output a state prediction matrix based on the multi-source operating condition data; A control strategy generation module, which is used to construct a weighted multi-objective control model according to the state prediction matrix, and output a sewage treatment control quantity sequence through rolling horizon model predictive control; A control strategy error monitoring module, which is used to perform real-time control on the sewage treatment system according to the sewage treatment control quantity sequence, obtain the collected actual operating condition response data, construct an actual operating condition response data matrix, and calculate a prediction error matrix according to the state prediction matrix; A parameter correction and feedback module, which is used to correct the parameters of the generalized sewage treatment engineering model according to the prediction error matrix, and output a parameter correction matrix to perform closed-loop control on the generalized sewage treatment engineering model.

[0043] Furthermore, the multi-source working condition data includes influent pollutant concentration, effluent pollutant concentration, aeration volume, sludge retention time (SRT), reflux ratio, pH value, and water temperature.

[0044] Furthermore, the engineering model construction and state prediction module specifically further includes: A dual-regulation rate model construction module, which is used to extract the organic matter degradation and sludge growth rate calculation sub-model and the ASM model state variable set in the ASM model according to the multi-source working condition data matrix, and extract the environmental modulation function and the sludge kinetics model state variable set in the sludge kinetics model according to the multi-source working condition data matrix. By unifying the saturated substrate concentration term and the dissolved oxygen regulation function, the function-level fusion of the organic matter degradation and sludge growth rate calculation sub-model and the environmental modulation function is carried out to construct a total rate calculation model including a dual-regulation mechanism; An environmental factor regulation module, which is used to introduce an environmental impact regulation factor to perform adaptive compensation on the total rate calculation model and dynamically regulate the biochemical reaction intensity; A variable fusion mapping module, which extracts and constructs a variable exchange mapping matrix based on the spectral characteristics of the multi-source working condition data matrix, and fuses the ASM model state variable set and the sludge kinetics model state variable set through the variable exchange mapping matrix to construct a unified state description space; A model solution and state response module, which is used to perform adaptive learning on the unified state description space through an error feedback backpropagation mechanism using the collected error terms to obtain a generalized sewage treatment engineering model, and output a state prediction matrix based on the multi-source working condition data.

[0045] Furthermore, the control strategy generation module specifically further includes: A multi-objective fusion constraint integration module, which is used to construct a weighted multi-objective control model with the minimization of effluent pollution concentration, the minimization of operating energy consumption, and the balance of sludge load stability as objectives according to the state prediction matrix; A control sequence generation module, which is used to set an initial variable sequence, use the initial variable sequence as the minimum feasible solution space, and adopt a rolling horizon model predictive control to calculate an optimal control variable sequence in each control period as the sewage treatment control quantity sequence.

[0046] Specifically, the implementation principle process of the above system is as follows: First, through the multi-source data acquisition module, multi-source operating condition data including influent pollutant concentration, effluent pollutant concentration, aeration volume, sludge retention time (SRT), reflux ratio, pH value, water temperature, etc. are obtained from multiple sources such as the SCADA system of the sewage treatment plant, on-line water quality monitoring instruments, and manual detection platforms. Through preprocessing such as missing value repair, time alignment, outlier removal, and normalization, a multi-source operating condition data matrix adapted to the model input is finally constructed. This data matrix serves as the core input for the subsequent engineering modeling and prediction module.

[0047] Subsequently, through the dual-regulation rate model construction module, the organic matter degradation rate and sludge growth rate formulas in the ASM model are extracted, and the environmental modulation function is introduced from the sludge kinetics model. By unifying the saturated substrate concentration term and the dissolved oxygen regulation function, the rate calculation model is fused to form a total rate calculation function that can represent carbon and nitrogen removal and sludge dynamics. At the same time, environmental regulation factors including parameters such as temperature and pH are introduced through the environmental factor regulation module to perform adaptive compensation on the total rate model and dynamically adjust the biochemical reaction intensity. Then, by analyzing the spectral characteristics of the multi-source operating condition data matrix, a state variable interchange mapping matrix is constructed to uniformly map the state variable sets of the ASM model and the sludge kinetics model, and a unified state description space is established. Further combined with the error feedback backpropagation mechanism, the unified state space is adaptively learned using the actual operating condition acquisition error term to obtain a generalized sewage treatment engineering model, and a state prediction matrix is output in real time based on multi-source data to form a prediction result of the future treatment state.

[0048] During control, taking the state prediction matrix as the input, first, through the multi-objective fusion constraint integration module, a weighted multi-objective control model is constructed around the multi-objective control requirements of minimizing effluent pollution concentration, minimizing operating energy consumption, and sludge load stability. Then, the control sequence generation module sets an initial variable sequence as the minimum feasible solution space, and uses the model predictive control (MPC) strategy with a rolling time domain to calculate the optimal control variable sequence (such as aeration volume, reflux ratio, sludge discharge volume, etc.) in real time during each control period, and the output is used as the sewage treatment control quantity sequence to drive the engineering control of the sewage treatment plant.

[0049] After the control is implemented, through the multi-source data acquisition module, based on the real-time operating condition response after being controlled by the sewage treatment control quantity sequence, real-time operating condition response parameters such as actual effluent indicators, energy consumption level, and sludge concentration are collected to construct a real-time operating condition response data matrix, which is compared with the state prediction matrix, and a prediction error matrix is calculated. Through the prediction error matrix, the sensitive parameters (such as reaction rate constants, regulation factors, saturation coefficients, etc.) in the generalized sewage treatment engineering model are corrected using the error backpropagation mechanism, and a parameter correction matrix is output, thereby realizing the closed-loop optimization of the entire modeling and control link.

[0050] Example 3 Based on Example 1, there is a multi-objective control application scenario for a sewage treatment plant, which adopts the intelligent control method and system of the general model for sewage treatment engineering based on ASM described in the above example.

[0051] The specific implementation method is as follows: In practical applications, this example takes a large-scale sewage treatment plant in a certain city as the test scenario, and implements it using the intelligent control method and system of the general model for sewage treatment engineering based on ASM, so as to realize the multi-objective optimization control of meeting the effluent water quality standards, reducing energy consumption, and stabilizing the sludge load. The SCADA system, on-line water quality monitoring instruments and artificial experiment platform of the sewage treatment plant collect multi-source operating condition data in real time, including the concentrations of influent COD, NH4⁺, TN, TP, effluent indexes, aeration volume, sludge retention time SRT, reflux ratio, mixed liquor suspended solids MLSS, pH value and water temperature, etc., and construct an original operating condition data matrix. Through missing value repair, sliding window resampling, sliding Z-score outlier removal and non-linear piecewise normalization processing, the pre-processed data matrix of the system is standardized and encapsulated as the input of the fusion model.

[0052] In the process of engineering model construction and state prediction, the system integrates the organic matter degradation and nitrification-denitrification rate formulas of the ASM model with the environmental modulation function of the sludge kinetic model at the function level through the dual-regulation rate model construction module to form a total rate calculation model, and introduces temperature and pH dynamic regulation factors through the environmental factor regulation module to enhance the adaptability of the model to different climate conditions and influent shocks. The variable fusion mapping module uses the spectral characteristics of multi-source operating condition data to establish a variable interchange mapping relationship between the ASM and the sludge model, and forms a state prediction matrix containing each state variable in the unified state description space. Through the model solving and state response module, the system uses the historical and real-time operating condition data for adaptive iterative learning based on the error feedback backpropagation mechanism, and predicts the change trends of future effluent COD, NH4⁺, TN and other indexes in real time.

[0053] The control strategy generation module constructs a weighted multi-objective control model aiming at minimizing the effluent pollution concentration, minimizing the total aeration energy consumption, and minimizing the sludge load fluctuation according to the state prediction matrix, sets the initial control variable sequence (such as initial aeration volume, reflux ratio, sludge discharge amount, etc.), and uses the moving horizon model predictive control (MPC) algorithm to calculate the optimal control variable sequence in each control period and output it to the actuator of the sewage treatment plant to drive real-time adjustment. The control strategy error monitoring module, based on the actual sewage treatment working condition response, collects the effluent indexes, energy consumption, and sludge concentration, constructs the actual working condition response data matrix, compares it with the predicted value, and calculates the prediction error matrix. The parameter correction and feedback module accordingly corrects the reaction rate constant, adjustment factor, etc. of the generalized sewage treatment engineering model to achieve full-link closed-loop control.

[0054] The experimental test period was 30 consecutive days. The results showed that after adopting this intelligent control system, the average effluent COD decreased from 35 mg / L of the original system to 28 mg / L, the average NH4⁺ decreased from 2.5 mg / L to 1.8 mg / L, the average TN decreased from 13 mg / L to 10.5 mg / L, and the effluent compliance rate increased from 92% to 98%; the energy consumption of the aeration system decreased by about 12.3% compared with the original control strategy based on the empirical method, and the sludge load fluctuation (measured by the standard deviation of MLSS) decreased by about 18.7%. The system showed stronger adaptability and stability under complex working conditions such as peak load, low temperature, and influent shock. Through dynamic error closed-loop adjustment, the model prediction error converged from an initial average of 12% to within 5% under long-term operation, significantly improving the multi-objective comprehensive control ability and system operation efficiency of the sewage treatment process.

[0055] Embodiment 4 Based on Embodiment 1, this embodiment proposes a terminal device for intelligent control of a generalized sewage treatment engineering model based on ASM. The terminal device includes at least one memory, at least one processor, and a bus connecting different platform systems.

[0056] The memory may include a readable medium in the form of volatile memory, such as RAM211 and / or cache memory, and may further include ROM213.

[0057] Among them, the memory also stores a computer program that can be executed by the processor, enabling the processor to execute any one of the above-mentioned intelligent control methods for an ASM-based sewage treatment engineering general model in the embodiments of the present application. The specific implementation manner is consistent with the implementation manners and the achieved technical effects described in the embodiments of the above method, and some contents will not be elaborated. The memory may also include a program / utilities having a set (at least one) of program modules. Such program modules include, but are not limited to: an operating system, one or more application programs, other program modules, and program data. Each or some combination of these examples may include the implementation of a network environment.

[0058] Correspondingly, the processor can execute the above computer program and can also execute the program / utilities.

[0059] The bus can represent one or more of several types of bus structures, including a memory bus or a memory controller, a peripheral bus, a graphics acceleration port, a processor, or a local bus using any bus structure in a variety of bus structures.

[0060] The terminal device can also communicate with one or more external devices such as a keyboard, a pointing device, a Bluetooth device, etc., and can also communicate with one or more devices capable of interacting with the terminal device, and / or communicate with any device (such as a router, a modem, etc.) that enables the terminal device to communicate with one or more other computing devices. Such communication can be carried out through the I / O interface. Moreover, the terminal device can also communicate with one or more networks (such as a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) through a network adapter. The network adapter can communicate with other modules of the terminal device through the bus. It should be understood that although not shown in the figure, other hardware and / or software modules can be used in combination with the terminal device, including but not limited to: microcode, device drivers, redundant processors, external disk drive arrays, RAID systems, tape drives, and data backup storage platforms, etc.

[0061] Embodiment Five Based on Embodiment One, this embodiment proposes a computer-readable storage medium for intelligent control of an ASM-based sewage treatment engineering general model. Instructions are stored on the computer-readable storage medium, and when the instructions are executed by the processor, the above-mentioned intelligent control method for an ASM-based sewage treatment engineering general model is implemented. The specific implementation manner is consistent with the implementation manners and the achieved technical effects described in the embodiments of the above method, and some contents will not be elaborated.

[0062] This embodiment provides a program product for implementing the above method. It can be a portable compact disc read-only memory (CD-ROM) and includes program code, and can run on a terminal device, such as a personal computer. However, the program product of the present invention is not limited to this. In this embodiment, the readable storage medium can be any tangible medium that contains or stores a program, and this program can be used by or in conjunction with an instruction execution system, device, or apparatus. The program product can adopt any combination of one or more readable media. The readable media can be a readable signal medium or a readable storage medium. The readable storage medium can, for example, but not be limited to, be an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or apparatus, or any combination of the above. More specific examples (non-exhaustive list) of the readable storage medium include: an electrical connection having one or more wires, a portable disc, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above.

[0063] The computer-readable storage medium can include a data signal propagated in a baseband or as part of a carrier wave, in which the readable program code is carried. Such a propagated data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The readable storage medium can also be any readable medium other than the readable storage medium, and this readable medium can send, propagate, or transmit a program for use by or in conjunction with an instruction execution system, device, or apparatus. The program code contained on the readable storage medium can be transmitted using any appropriate medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination of the above. The program code for performing the operations of the present invention can be written in any combination of one or more programming languages. The programming languages include object-oriented programming languages such as Java, C++, etc., and also include conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user computing device, partially on the user device, executed as an independent software package, partially on the user computing device and partially on a remote computing device, or entirely on a remote computing device or server. In the case of a remote computing device, the remote computing device can be connected to the user computing device through any type of network, including a local area network (LAN) or a wide area network (WAN), or can be connected to an external computing device (for example, through an Internet service provider via the Internet).

[0064] The above are only the preferred embodiments of the present invention. It should be understood that the present invention is not limited to the forms disclosed herein, and should not be regarded as excluding other embodiments. Instead, it can be used in various other combinations, modifications, and environments, and can be changed within the scope of the concept described herein through the above teachings or the techniques or knowledge in the relevant field. Any changes and modifications made by those skilled in the art without departing from the spirit and scope of the present invention shall fall within the protection scope of the appended claims of the present invention.

Claims

1. An intelligent control method for an engineering general model of sewage treatment based on ASM, characterized in that It includes the following steps: S1. Obtain the multi-source operating condition data of the sewage treatment plant, preprocess the multi-source operating condition data, and construct a multi-source operating condition data matrix related to the sewage treatment plant; S2. According to the multi-source operating condition data matrix, through state variable mapping, respectively extract the reaction rate functions and state variable sets corresponding to the multi-source operating condition data matrix in the ASM model and the sludge growth kinetics model, construct a generalized sewage treatment engineering model, and output a state prediction matrix based on the multi-source operating condition data; S3. According to the state prediction matrix, construct a weighted multi-objective control model, and output a sewage treatment control quantity sequence through model predictive control with a rolling time domain; S4. According to the sewage treatment control quantity sequence, perform real-time control on the sewage treatment system to obtain the collected actual operating condition response data, construct an actual operating condition response data matrix, and calculate a prediction error matrix according to the state prediction matrix; S5. According to the prediction error matrix, correct the parameters of the generalized sewage treatment engineering model, output a parameter correction matrix and return to step S3 for closed-loop control of the generalized sewage treatment engineering model.

2. The intelligent control method for the general model of sewage treatment engineering based on ASM according to claim 1, wherein, Step S1 includes the following steps: S101. Collect from the sewage treatment plant the influent pollutant concentration, effluent pollutant concentration, aeration volume, sludge retention time SRT, reflux ratio, pH value, and water temperature, and construct an original data matrix; S102. Extract spectral features by constructing a local linear trend model within a time window, keep the boundary observation values unchanged, and perform linear weighted calculation using the known data points on the left and right to interpolate and repair the boundaries of the missing values in the original data matrix; S103. Through a sliding time window with a fixed length, perform mean aggregation within the window for the original data with different sampling frequencies, and uniformly output a multi-source operating condition data matrix with a consistent time step.

3. The intelligent control method of a general model for sewage treatment engineering based on ASM according to claim 1, characterized in that Step S2 includes the following steps: S201. According to the multi-source operating condition data matrix, extract the organic matter degradation and sludge growth rate calculation sub-model and the ASM model state variable set in the ASM model, and according to the multi-source operating condition data matrix, extract the environmental modulation function and the sludge kinetics model state variable set in the sludge kinetics model. Through unifying the saturated substrate concentration term and the dissolved oxygen regulation function, perform function-level fusion on the organic matter degradation and sludge growth rate calculation sub-model and the environmental modulation function to construct a total rate calculation model with a dual regulation mechanism; S202. Introduce an environmental impact adjustment factor to perform adaptive compensation on the total rate calculation model and dynamically adjust the biochemical reaction intensity; S203. Extract and construct a variable swap mapping matrix based on the spectral features of the multi-source operating condition data matrix, and fuse the ASM model state variable set and the sludge kinetics model state variable set through the variable swap mapping matrix to construct a unified state description space; S204. Through an error feedback backpropagation mechanism, use the collected error terms to perform adaptive learning on the unified state description space to obtain a generalized sewage treatment engineering model, and output a state prediction matrix based on the multi-source operating condition data.

4. An intelligent control method for an engineering general model of sewage treatment based on ASM according to claim 3, characterized in that In step S201, the total rate calculation model is specifically expressed as: ； Among them, represents the actual biological reaction rate at time t, represents the maximum specific growth rate, represents the soluble substrate concentration at time t, represents the half-saturation substrate constant, represents the dissolved oxygen concentration at time t, represents the oxygen half-saturation constant, represents the environmental sensitivity factor function.

5. The intelligent control method of a general model for sewage treatment engineering based on ASM according to claim 3, characterized in that, In step S203, the unified state description space is specifically expressed as: ； Among them, represents the target variable vector, represents the variable mapping matrix, represents the set of ASM model state variables, represents the set of sludge kinetic model state variables.

6. The intelligent control method of a general model for sewage treatment engineering based on ASM according to claim 1, characterized in that, Step S3 includes the following steps: S301. According to the state prediction matrix, construct a weighted multi-objective control model with the goals of minimizing the effluent pollution concentration, minimizing the operating energy consumption, and balancing the sludge load stability respectively; S302. Set the initial variable sequence, use the initial variable sequence as the minimum feasible solution space, and adopt the rolling horizon model predictive control to calculate the optimal control variable sequence within each control period as the sewage treatment control quantity sequence.

7. An intelligent control method for a general model of sewage treatment engineering based on ASM according to claim 6, characterized in that, In step S301, constructing a weighted multi-objective control model with the goals of minimizing the effluent pollution concentration, minimizing the operating energy consumption, and balancing the sludge load stability respectively includes the following sub-steps: S3011. With the goal of minimizing the effluent pollution concentration, calculate the time accumulation value of the L2 norm error between the predicted value of the effluent pollutant and the target emission value, specifically expressed as: ； Among them, represents the pollutant emission deviation from the target penalty term, represents the predicted effluent pollutant concentration vector at time t, represents the target effluent discharge standard concentration vector, represents the square of the Euclidean norm; S3012. With the goal of minimizing the operating energy consumption, calculate the cumulative consumption power of the aeration equipment and the sludge pump, specifically expressed as: ； Among them, represents the penalty term for the deviation of the operating energy consumption from the target, represents the power consumption of the aeration equipment at time t at time t, represents the power consumption of the sludge pump at time t at time t; S3013. With the goal of balancing the sludge load stability, calculate the approaching value of the unit sludge load and the design load target of the sewage treatment plant, specifically expressed as: ； Among them, represents the penalty term for the deviation of the sludge load stability from the target, represents the mixed liquor suspended solids concentration at time t, represents the average sludge retention time at time t, represents the target ratio; S3014. According to the calculated time accumulation value, cumulative consumption power value and approaching value, construct a weighted multi-objective control model, specifically expressed as: ； Among them, represents the weighted multi-objective control model, , , respectively represent the contribution of the effluent water pollution concentration to the weighted multi-objective control model, the contribution of the operating energy consumption to the weighted multi-objective control model, and the contribution of the sludge load stability to the weighted multi-objective control model.

8. An intelligent control method for an engineering general model of sewage treatment based on ASM according to claim 1, characterized in that Step S4 includes the following steps: S401. According to the sewage treatment control quantity sequence, perform real-time control on the multi-source working condition data respectively, and collect the actual working condition response data in real time according to the fixed sampling interval to construct the actual working condition response data matrix; S402. Through timestamp matching and the variational dynamic time warping algorithm, align the data of the actual working condition response data matrix and the state prediction matrix, and calculate the prediction error matrix.

9. An intelligent control method for an engineering general model of sewage treatment based on ASM according to claim 1, characterized in that Step S5 includes the following steps: S501. According to the prediction error matrix, perform sensitivity analysis on the parameters of the weighted multi-objective control model, and obtain the priority update parameter subset by calculating the partial derivatives of the error matrix and the corresponding parameters; S502. According to the priority update parameter subset, introduce the multi-scale coordination factor, and update the parameters of the priority update parameter subset in combination with the model complexity constraint; S503. According to the updated priority update parameter subset, output the parameter correction matrix, and feedback the parameter correction matrix to step S3.

10. The intelligent control method of a general model for sewage treatment engineering based on ASM according to claim 9, characterized in that, The said step S501 includes the following sub-steps: S5011. According to the prediction error matrix, calculate the norm of the prediction error matrix at the current moment; S5012. According to the norm of the prediction error matrix, calculate the sensitivity partial derivatives of all parameters in the weighted multi-objective control model respectively; S5013. Based on the historical multi-source working condition data of the sewage treatment plant, set the sensitivity threshold, sort the sensitivity partial derivatives of all parameters, and when the sensitivity partial derivative is higher than the sensitivity threshold, it is used as a high-sensitivity parameter to obtain the priority update parameter subset.

11. An intelligent control method for an engineering general model of sewage treatment based on ASM according to claim 9, characterized in that, The said step S502 includes the following sub-steps: S5021. According to the dynamic learning strategy, set the base learning rate; S5022. Perform spectral decomposition on the prediction error matrix, and calculate the frequency error component and the normalized energy weight of all parameters in the prediction error matrix respectively; S5023. Calculate the correction amount of each parameter under the frequency error component according to the priority update parameter subset, and merge the correction amounts of each parameter into the final update amount.

12. An intelligent control system for an engineering general model of sewage treatment based on ASM, which is implemented based on the intelligent control method for an engineering general model of sewage treatment based on ASM described in any one of claims 1-11, and is characterized in that, Including: A multi-source data acquisition module, which is used to acquire the multi-source operating condition data of the sewage treatment plant, preprocess the multi-source operating condition data, and construct a multi-source operating condition data matrix related to the sewage treatment plant; An engineering model construction and state prediction module, which is used to extract the reaction rate function and the state variable set corresponding to the multi-source operating condition data matrix in the ASM model and the sludge growth kinetics model respectively through state variable mapping according to the multi-source operating condition data matrix, construct a generalized sewage treatment engineering model, and output a state prediction matrix based on the multi-source operating condition data; A control strategy generation module, which is used to construct a weighted multi-objective control model according to the state prediction matrix, and output a sewage treatment control quantity sequence through model predictive control of the rolling time domain; A control strategy error monitoring module, which is used to perform real-time control on the sewage treatment system according to the sewage treatment control quantity sequence, obtain the collected actual operating condition response data, construct an actual operating condition response data matrix, and calculate a prediction error matrix according to the state prediction matrix; A parameter correction and feedback module, which is used to correct the parameters of the generalized sewage treatment engineering model according to the prediction error matrix, and output a parameter correction matrix for closed-loop control of the generalized sewage treatment engineering model.

13. An intelligent control system for an engineered general model of sewage treatment based on ASM according to claim 12, characterized in that, The multi-source operating condition data includes influent pollutant concentration, effluent pollutant concentration, aeration volume, sludge age SRT, reflux ratio, PH value, and water temperature.

14. An intelligent control system for a general model of sewage treatment engineering based on ASM according to claim 12, characterized in that, The engineering model construction and state prediction module specifically further includes: A dual-regulation rate model construction module, which is used to extract the organic matter degradation and sludge growth rate calculation sub-model and the ASM model state variable set in the ASM model according to the multi-source operating condition data matrix, and extract the environmental modulation function and the sludge kinetics model state variable set in the sludge kinetics model according to the multi-source operating condition data matrix. Through unifying the saturated substrate concentration term and the dissolved oxygen regulation function, perform function-level fusion on the organic matter degradation and sludge growth rate calculation sub-model and the environmental modulation function, and construct a total rate calculation model including a dual-regulation mechanism; An environmental factor adjustment module, which is used to introduce an environmental impact adjustment factor, perform adaptive compensation on the total rate calculation model, and dynamically adjust the biochemical reaction intensity; A variable fusion mapping module, which is used to extract and construct a variable exchange mapping matrix based on the spectral characteristics of the multi-source operating condition data matrix, and fuse the ASM model state variable set and the sludge kinetics model state variable set through the variable exchange mapping matrix to construct a unified state description space; A model solution and state response module, which is used to perform adaptive learning on the unified state description space through an error feedback backpropagation mechanism using the collected error term, obtain a generalized sewage treatment engineering model, and output a state prediction matrix based on the multi-source operating condition data.

15. An intelligent control system for a general model of sewage treatment engineering based on ASM according to claim 12, characterized in that, The control strategy generation module specifically further includes: The multi-objective fusion constraint integration module is used to construct a weighted multi-objective control model with the goals of minimizing the effluent pollution concentration, minimizing the operating energy consumption, and balancing the sludge load stability respectively according to the state prediction matrix. The control sequence generation module is used to set the initial variable sequence, take the initial variable sequence as the minimum feasible solution space, and use the rolling horizon model predictive control to calculate the optimal control variable sequence within each control period as the sewage treatment control quantity sequence.

Images