Multi-parameter intelligent regulation method and system for industrial wastewater treatment

Abstract

The present application relates to wastewater treatment technical field, especially to the multi-parameter intelligent control method and system for industrial wastewater treatment, the method comprises the apparent viscosity change trend, the reaction rate change trend and the dosing quantity change trend in the reaction unit are acquired, and the reaction control mechanism evolution track curve is constructed based on the time sequence correlation among the three, the effective response interval of the reaction rate to the dosing change is extracted, the critical interval of the dosing behavior starting to fail is identified, whether the reaction process enters the limited operation state dominated by the mass transfer condition is judged, the control focus is changed from the reaction rate improvement to the reaction interface update ability coefficient recovery, and the matched control constraint condition matrix is formed.The present application synchronously acquires the key operation parameters such as apparent viscosity, reaction rate and dosing quantity in the continuous operation process of wastewater treatment, constructs the reaction control mechanism evolution track reflecting the internal coupling relationship of the reaction process, and realizes the dynamic identification of the actual action boundary of the control behavior.

Description

Technical Field

[0001] This invention relates to the field of wastewater treatment technology, specifically to a multi-parameter intelligent control method and system for industrial wastewater treatment. Background Technology

[0002] In existing industrial wastewater treatment processes, the dosing efficiency is generally adjusted by monitoring the changing trends of the reaction rate in the wastewater. However, as the wastewater treatment process progresses, the content of high-molecular-weight organic matter, emulsions, and suspended colloids in the wastewater gradually increases. The flow state in the reaction tank gradually changes from fully turbulent to weakly turbulent, and low shear strength zones appear in local areas. Under high viscosity conditions, the diffusion coefficient of pollutant molecules in the liquid phase decreases significantly, forming a stable concentration boundary layer near the reaction interface. Although the control system continuously increases the dosage of reagents to improve the reaction rate, the actual reaction process gradually shifts from chemical reaction kinetic control to mass transfer control. The sensitivity of the reaction rate to changes in reagent concentration decreases significantly. Under the condition of persistent mass transfer limitation, the apparent removal rate of the system decreases slowly, but no drastic anomalies appear in a short period of time, which can easily be misjudged as normal fluctuations.

[0003] To maintain the effluent quality standards, the control system continuously increases the dosage of chemicals, resulting in a significant increase in operating costs. At the same time, reaction byproducts accumulate in the system. After long-term operation, the effective treatment capacity of the reaction unit is irreversibly reduced. Once the influent load increases again, the system will quickly become unstable, the recovery period will be significantly prolonged, and there is a risk of exceeding the emission standards. Summary of the Invention

[0004] To address the aforementioned problems, this invention provides a multi-parameter intelligent control method and system for industrial wastewater treatment.

[0005] This invention employs the following technical solution: a multi-parameter intelligent control method for industrial wastewater treatment, comprising:

[0006] During the continuous operation of wastewater treatment, the trends of apparent viscosity, reaction rate, and dosage within the reaction unit are acquired simultaneously, and the evolution trajectory curve of the reaction control mechanism is constructed based on the temporal correlation between the three.

[0007] Based on the evolution trajectory curve of the reaction control mechanism, the effective response range of the reaction rate to the change in drug dosage is extracted, dynamic response characteristics reflecting the actual boundary of the regulatory action are constructed, and the critical interval at which the drug dosage behavior begins to fail is identified.

[0008] After identifying the critical interval, the reaction interface renewal capacity coefficient is quantitatively characterized by combining the apparent viscosity change trend to determine whether the reaction process has entered a restricted operating state dominated by mass transfer conditions.

[0009] After confirming the restricted operating state, the focus of regulation is shifted from increasing the reaction rate to restoring the reaction interface update capability coefficient, and a matching regulation constraint matrix is ​​formed.

[0010] The decay rate per unit time of the reaction interface renewal capacity coefficient, the reaction interface renewal capacity coefficient, the control parameters, and the control constraint matrix are input into the pre-constructed wastewater treatment control and prediction model, and the control parameters are output. The wastewater treatment unit is automatically controlled based on the output control parameters.

[0011] As a further description of the above technical solution: the method for constructing the evolution trajectory curve of the reaction control mechanism includes:

[0012] Apparent viscosity time series, reaction rate time series, and dosage time series are collected synchronously according to a unified sampling period and time calibration is performed to ensure that each parameter has a one-to-one correspondence on the same time axis.

[0013] With a preset sliding window length, the apparent viscosity time series, reaction rate time series, and dosage time series are processed by sliding window to calculate the changing trend of each parameter in the continuous operating range.

[0014] The response delay time of the preset reaction process is used to align the trend of reaction rate change with the corresponding trend of dosage change in time. The dynamic response coefficient of reaction rate to dosage change is calculated. The dynamic response coefficient is then analyzed in conjunction with the trend of apparent viscosity change to calculate the modulation relationship of dynamic response coefficient with viscosity change.

[0015] Based on the apparent viscosity change trend, reaction rate change trend, and reaction response index after viscosity modulation, a multi-parameter time-series correlation vector is constructed within the same time window. The multi-parameter time-series correlation vector is sequentially arranged in a continuous time interval to form a trajectory curve reflecting the evolution of reaction control characteristics over time, thus obtaining the reaction control mechanism evolution trajectory curve.

[0016] As a further description of the above technical solution: the method for identifying the critical interval where the dosing behavior begins to fail includes:

[0017] The evolution trajectory curve of the reaction control mechanism is divided into multiple operating intervals by the length of the sliding window. The ratio between the change in reaction rate and the change in dosage in adjacent operating intervals is calculated to form a response coefficient sequence that characterizes the degree of change in reaction rate caused by a unit change in dosage.

[0018] The response coefficient sequence is mapped to the apparent viscosity change trend. When the response coefficient remains above the preset stable threshold in a continuous time interval, it is determined that the reaction process in the corresponding time interval is still in a state of effective response to the change in drug addition.

[0019] Mark the continuous time intervals that satisfy the state determination conditions for effective response on the evolution trajectory curve of the reaction control mechanism to form the effective response interval of the reaction rate to the change in drug dosage;

[0020] When the response coefficient shows a decreasing trend over three consecutive operating intervals on the evolution trajectory curve of the reaction control mechanism, and the apparent viscosity change trend within the corresponding operating interval shows a monotonically increasing characteristic over three consecutive operating intervals, the corresponding operating interval is marked as the critical interval of the effective response interval.

[0021] As a further description of the above technical solution: the method for quantitatively characterizing the reaction interface renewal capacity coefficient by combining the apparent viscosity change trend includes:

[0022] Within the critical range, the apparent viscosity data of the reaction unit are continuously collected. A sliding time window Δt was used to smooth the collected data, and the apparent viscosity change trend over time was calculated. ;

[0023] Based on the apparent viscosity change trend Define the function for the reaction interface update capability coefficient: ;in, Indicates time The response interface update capability coefficient, Update the capability coefficient calibration values ​​for the initial interface. It is a monotonically decreasing function.

[0024] As a further description of the above technical solution: the method for determining whether the reaction process has entered a restricted operating state dominated by mass transfer conditions includes: calculating the decay rate per unit time of the reaction interface renewal capacity coefficient; when the decay rate per unit time of the reaction interface renewal capacity coefficient is greater than the dynamic response coefficient, it is determined that the mass transfer process has become the dominant factor restricting the reaction; otherwise, the mass transfer process has not become the dominant factor restricting the reaction.

[0025] As a further description of the above technical solution: the method for generating the control constraint matrix includes:

[0026] Under continuous operation conditions, a set of characteristic parameters affecting the reaction interface update capability coefficient are synchronously collected within the critical interval;

[0027] Within the established critical range, the normalized perturbation analysis method is used to calculate the sensitivity coefficient of each characteristic parameter to the interface update capability. The sensitivity coefficient is used to characterize the contribution of a unit change in characteristic parameter to the change in interface update capability, thus forming a sensitivity mapping relationship between characteristic parameter and interface update capability.

[0028] A preset sensitivity threshold is set, and feature parameters with sensitivity coefficients higher than the sensitivity threshold are selected as the dominant control parameters for interface update capability, and the dominant control parameters are marked as an adjustable parameter set.

[0029] After determining the set of adjustable parameters, for each control parameter, a corresponding control constraint matrix is ​​established in combination with the equipment capacity boundary, operational safety constraints and process stability requirements.

[0030] As a further description of the above technical solution: the characteristic parameters include at least stirring speed, shear rate, wastewater discharge rate and pollutant influent rate, wherein each characteristic parameter is obtained in the form of a time series and is aligned with the reaction interface update capability coefficient on the same time axis.

[0031] As a further description of the above technical solution: the expression for the control constraint condition matrix is: ,in Indicates the first One control parameter, and These represent the lower and upper limits of the control parameter during the current operating phase, respectively. The lower and upper limits are used to restrict the range of change of the control parameter during the process of restoring the interface update capability.

[0032] As a further description of the above technical solution: the method for calculating the sensitivity coefficient of each feature parameter to the interface update capability is as follows:

[0033] Update the reaction interface capability coefficient As target response variables, the stirring speed, shear rate, wastewater discharge rate, and pollutant influent rate are denoted as... And perform time synchronization and scale normalization on each feature parameter;

[0034] While keeping the other characteristic parameters unchanged, a preset amplitude perturbation is applied to any characteristic parameter, and the change in the corresponding reaction interface update capability coefficient is recalculated within the perturbation range.

[0035] Based on the changes before and after the disturbance, the sensitivity coefficient of the feature parameters to the interface update capability is calculated.

[0036] A multi-parameter intelligent control system for industrial wastewater treatment, used to implement the aforementioned multi-parameter intelligent control method for industrial wastewater treatment, the system comprising:

[0037] The mechanism trajectory construction module simultaneously acquires the apparent viscosity change trend, reaction rate change trend, and dosage change trend within the reaction unit during continuous operation of wastewater treatment, and constructs the reaction control mechanism evolution trajectory curve based on the temporal correlation between the three.

[0038] The regulation boundary identification module extracts the effective response range of the reaction rate to changes in drug dosage based on the evolution trajectory curve of the reaction control mechanism, constructs dynamic response features that reflect the actual boundary of the regulation action, and identifies the critical interval where the drug dosage behavior begins to fail.

[0039] The mass transfer limitation identification module, after identifying the critical interval, combines the apparent viscosity change trend to quantitatively characterize the reaction interface update capacity coefficient, and determines whether the reaction process has entered a limited operating state dominated by mass transfer conditions.

[0040] The control strategy switching module, after confirming the restricted operating state, shifts the control focus from increasing the reaction rate to restoring the reaction interface update capability coefficient, and forms a matching control constraint condition matrix.

[0041] The model prediction and control module inputs the unit time decay rate of the reaction interface update capacity coefficient, the reaction interface update capacity coefficient, the control parameters, and the control constraint matrix into the pre-constructed wastewater treatment control and prediction model, outputs the control parameters and control values, and automatically controls the wastewater treatment unit based on the output control parameters and control values.

[0042] The beneficial effects of this invention are as follows:

[0043] This invention acquires key operating parameters such as apparent viscosity, reaction rate, and dosage simultaneously during continuous wastewater treatment operation. It constructs a reaction control mechanism evolution trajectory curve reflecting the inherent coupling relationship of the reaction process, enabling dynamic identification of the actual boundary of the control behavior. By analyzing the effective response range of the reaction rate to changes in dosage, it can accurately determine the critical range where the dosing behavior transitions from effective control to marginal failure, avoiding extensive control based solely on experience or a single indicator, thereby improving the objectivity and accuracy of operational status identification.

[0044] Building upon this foundation, this invention further introduces a reaction interface renewal capacity coefficient to quantitatively identify the transition of the reaction process from kinetic control to mass transfer-constrained control. When the system enters a constrained operating state, the control focus can be promptly shifted from simply increasing the reaction rate to restoring the reaction interface renewal capacity, forming a matching control constraint matrix. By inputting the interface renewal capacity decay characteristics, control parameters, and constraints into the prediction model, adaptive optimization of the control parameters is achieved, thereby improving the stability, responsiveness, and overall operating efficiency of the wastewater treatment system under complex operating conditions. Attached Figure Description

[0045] The present invention will be further explained below with reference to the accompanying drawings and embodiments:

[0046] Figure 1 This is a flowchart illustrating the multi-parameter intelligent control method for industrial wastewater treatment provided in Embodiment 1 of the present invention.

[0047] Figure 2 This is a flowchart of the method for constructing the evolution trajectory of the reaction control mechanism provided in Embodiment 1 of the present invention;

[0048] Figure 3 This is a flowchart of the method for identifying the critical interval where the dosing behavior begins to fail, as provided in Embodiment 1 of the present invention.

[0049] Figure 4 This is a module connection diagram of the multi-parameter intelligent control system for industrial wastewater treatment provided in Embodiment 2 of the present invention. Detailed Implementation

[0050] To make the technical means, creative features, objectives, and effects of this invention readily understandable, the invention is further described below with reference to specific illustrations. It should be noted that, unless otherwise specified, the embodiments and features described in these embodiments can be combined with each other.

[0051] Example 1:

[0052] Please see Figures 1-3 This invention provides a technical solution: a multi-parameter intelligent control method for industrial wastewater treatment, comprising:

[0053] During the continuous operation of wastewater treatment, the trends of apparent viscosity, reaction rate, and dosage within the reaction unit are acquired simultaneously. Based on the temporal correlation among these three factors, an evolution trajectory curve of the reaction control mechanism is constructed to characterize the dynamic changes of the dominant factors controlling the reaction process.

[0054] Methods for constructing the evolution trajectory curve of the reaction control mechanism include:

[0055] Apparent viscosity time series, reaction rate time series, and dosage time series are collected synchronously according to a unified sampling period, and the data are time-calibrated to ensure that each parameter has a one-to-one correspondence on the same time axis.

[0056] It should be noted that the method for acquiring the apparent viscosity time series includes: setting up a measurement module in the stirring zone of the reaction unit. The measurement module includes a shear structure with known geometric parameters, such as a narrow slit channel, capillary tube, or rotating shear element, and a matching pressure, rotation speed, or torque sensor. Under fixed shear rate or known shear conditions, pressure drop, rotation speed, or driving torque signals are acquired in real time, and the apparent viscosity value at the corresponding time is calculated according to a pre-calibrated rheological model. The rheological model is used to describe the functional relationship between shear stress, shear rate, and viscosity of the tested reaction medium under given shear conditions. It converts the measurable signal into apparent viscosity based on the geometric parameters of the known shear structure and the corresponding mechanical measurements. The rheological model can be an empirical or semi-empirical model applicable to non-Newtonian fluids, such as a power-law model, Bingham model, or its equivalent form. Its model parameters are obtained by offline calibration of standard samples or typical reactants and remain unchanged or are updated according to a calibration strategy during online measurement.

[0057] The method for acquiring the reaction rate time series includes: deploying an online water quality sensor within the reaction unit to acquire characterization parameters directly related to the target reaction in real time, including but not limited to COD, ammonia nitrogen, phosphorus concentration, dissolved organic matter concentration, conductivity, or characteristic absorbance; calculating the amount of reactant removed or converted per unit time based on the changes in characterization parameters at adjacent sampling times, combined with the effective volume of the reaction unit, thereby obtaining the equivalent reaction rate value.

[0058] With a preset sliding window length, the apparent viscosity time series, reaction rate time series, and dosage time series are processed by sliding window to calculate the changing trend of each parameter in the continuous operating range.

[0059] The response delay time of the preset reaction process is used to align the trend of reaction rate change with the corresponding trend of dosage change in time, and the dynamic response coefficient of reaction rate to dosage change is calculated.

[0060] Optionally, the formula for calculating the dynamic response coefficient is: In the formula, for The dynamic response coefficient at time t. This represents the change in reaction rate within the current sampling window. The change in dosage that caused this change, This is the response delay time of the reaction process.

[0061] The dynamic response coefficient and the apparent viscosity change trend are analyzed together to calculate the modulation relationship of the dynamic response coefficient with viscosity change, which is used to characterize the degree of influence of mass transfer condition changes on reaction response capability.

[0062] Specifically, the following related quantities can be constructed: In the formula, The response index after viscosity modulation. A monotonic mapping function characterizes the degree of inhibition of reaction response by viscosity changes. This monotonic mapping function maps the apparent viscosity change trend to the attenuation of effective shear strength and interface refresh frequency near the interface. Its specific form is not limited to an exponential function; it can also be a power function or a piecewise linear function. The exponential form is used to reflect the nonlinear inhibition effect of viscosity increase on interface renewal capability. The apparent viscosity change trend, In the formula, For at any time Apparent viscosity data of the wastewater.

[0063] Optionally, The function expression is: In the formula, This is the viscosity suppression sensitivity coefficient; the viscosity trend increases. Decays exponentially. It is an exponential function with the natural constant as its base. The reference viscosity value is set manually. This is the dimensionless value after normalization.

[0064] Viscosity suppression sensitivity coefficient The value is obtained by fitting historical operating data, and the optimal method is to determine the value by minimizing the error between the predicted value of the interface update capability and the actual operating effect.

[0065] Based on the apparent viscosity change trend, the reaction rate change trend, and the reaction response index after viscosity modulation, a multi-parameter time-series correlation vector is constructed within the same time window.

[0066] The multi-parameter time-series correlation vector It can be represented as: In the formula, The apparent viscosity change trend, This represents the trend of reaction rate change. This is the reaction response index after viscosity modulation; among which, ; This represents the change in reaction rate within the length of the sliding window. The length of the sliding window;

[0067] The multi-parameter time-series correlation vectors are sequentially arranged within a continuous time interval to form a trajectory curve reflecting the evolution of reaction control characteristics over time, thus obtaining the evolution trajectory curve of the reaction control mechanism.

[0068] In this embodiment, instead of relying solely on the reaction rate as a single indicator for dosing control, the evolution trajectory curve of the reaction control mechanism is constructed by simultaneously introducing the trends of apparent viscosity change, reaction rate change, and dosing amount change. This allows for the dynamic characterization of the dominant control factors of the reaction process from the perspective of the operational mechanism, enabling timely identification of key stages in the evolution of the reaction process from kinetic control to mass transfer control. This avoids the misjudgment problem caused by only observing the apparent removal rate in existing technologies, and improves the accuracy and foresight of operational status identification.

[0069] Based on the evolution trajectory curve of the reaction control mechanism, the effective response range of the reaction rate to the change in drug dosage is extracted, dynamic response characteristics reflecting the actual boundary of the regulatory action are constructed, and the critical interval at which the drug dosage behavior begins to fail is identified.

[0070] The method for identifying the critical interval at which the dosing action begins to fail includes:

[0071] The evolution trajectory curve of the reaction control mechanism is divided into multiple operating intervals by the sliding window length. The ratio between the change in reaction rate and the change in dosage within adjacent operating intervals is calculated to form a response coefficient sequence that characterizes the degree of reaction rate change caused by a unit change in dosage. This sequence is used to reflect the actual response intensity of the reaction process to dosage regulation.

[0072] The response coefficient sequence is mapped to the apparent viscosity change trend. When the response coefficient remains above a preset stability threshold in a continuous time interval, it is determined that the reaction process in the corresponding time interval is still in a state of effective response to the change in drug addition. The preset stability threshold is set by those skilled in the art based on the actual situation or obtained by simulation of a large amount of data.

[0073] Mark the continuous time intervals that satisfy the state determination conditions for effective response on the evolution trajectory curve of the reaction control mechanism to form the effective response interval of the reaction rate to the change in drug dosage;

[0074] When the response coefficient shows a decreasing trend over three consecutive operating intervals on the reaction control mechanism evolution trajectory curve, and the apparent viscosity change trend within the corresponding operating interval shows a monotonically increasing characteristic over three consecutive operating intervals, the corresponding operating interval is marked as the critical interval of the effective response interval; used to indicate the critical interval of the reaction control mechanism transitioning from kinetic control to mass transfer control.

[0075] In this embodiment, by extracting the effective response range of the reaction rate to changes in drug addition and identifying the critical range where the drug addition behavior begins to fail, a quantitative characterization of the "effective action boundary" of the control action is achieved. This enables the control system to distinguish between the operating states of "drug addition is still effective" and "drug addition has failed but has not yet caused obvious abnormalities." This fundamentally solves the problem of blindly adding drugs and continuously increasing costs in the prior art, which is difficult to detect in a timely manner, and effectively reduces drug waste and operating costs.

[0076] After identifying the critical interval, the reaction interface renewal capacity coefficient is quantitatively characterized by combining the apparent viscosity change trend to determine whether the reaction process has entered a restricted operating state dominated by mass transfer conditions.

[0077] The method for quantitatively characterizing the reaction interface renewal capacity coefficient by combining the apparent viscosity change trend includes:

[0078] Within the critical range, the apparent viscosity data of the reaction unit are continuously collected. A sliding time window Δt was used to smooth the collected data, and the apparent viscosity change trend over time was calculated. ;

[0079] Based on the apparent viscosity change trend Define the function for the reaction interface update capability coefficient: ;in, Indicates time The reaction interface renewal capacity coefficient is used to characterize the overall ability of the liquid phase near the reaction interface to be continuously replaced by fresh reactants during the reaction process. Update the capability coefficient calibration values ​​for the initial interface. A monotonic mapping function is used to characterize the degree of inhibition of reaction response by viscosity change. The monotonic mapping function is used to map the apparent viscosity change trend to the degree of decay of effective shear strength near the interface and interface refresh frequency.

[0080] The methods for determining whether a reaction process has entered a restricted operating state dominated by mass transfer conditions include:

[0081] Calculate the decay rate per unit time of the reaction interface renewal capacity coefficient. If the decay rate per unit time of the reaction interface renewal capacity coefficient is greater than the dynamic response coefficient, the mass transfer process is determined to be the dominant factor limiting the reaction. Otherwise, the mass transfer process is not the dominant factor limiting the reaction.

[0082] In this embodiment, after identifying the critical interval, a reaction interface renewal capacity coefficient is constructed based on the apparent viscosity change trend to quantitatively characterize the liquid phase renewal capacity near the reaction interface. This coefficient serves as the core basis for judging the mass transfer-limited operating state, so that the determination of the limiting mechanism of the reaction process no longer depends on empirical thresholds or post-event result analysis, but is based on calculable and traceable physical evolution characteristics, which significantly improves the scientificity and reliability of the limiting state determination.

[0083] After confirming the restricted operating state, the focus of regulation is shifted from increasing the reaction rate to restoring the reaction interface update capability coefficient, and a matching regulation constraint matrix is ​​formed.

[0084] The method for generating the regulation constraint matrix includes:

[0085] Under continuous operation conditions, a set of characteristic parameters affecting the reaction interface renewal capacity coefficient are synchronously collected within the critical interval. The characteristic parameters include at least stirring speed, shear rate, wastewater discharge rate and pollutant influent rate. Each characteristic parameter is obtained in the form of a time series and is aligned with the reaction interface renewal capacity coefficient on the same time axis.

[0086] Within the established critical range, the normalized perturbation analysis method is used to calculate the sensitivity coefficient of each characteristic parameter to the interface update capability. The sensitivity coefficient is used to characterize the contribution of a unit change in characteristic parameter to the change in interface update capability, thereby forming a sensitivity mapping relationship between characteristic parameter and interface update capability.

[0087] A preset sensitivity threshold is set, and feature parameters with sensitivity coefficients higher than the sensitivity threshold are selected as the dominant control parameters for interface update capability. The dominant control parameters are marked as an adjustable parameter set for subsequent control strategy execution.

[0088] After determining the set of adjustable parameters, for each control parameter, a corresponding control constraint matrix C is established based on the equipment capacity boundary, operational safety constraints, and process stability requirements. The control constraint matrix satisfies the following: ,in Indicates the first One control parameter, and These represent the lower and upper limits of the control parameters allowed during the current operating phase, respectively. These limits restrict the magnitude of change in the control parameters during the recovery of the interface update capability, preventing new flow instability or system overload. The lower and upper limits are set by those skilled in the art based on historical experience.

[0089] The sensitivity coefficients of each feature parameter to the interface update capability are calculated as follows:

[0090] Update the reaction interface capability coefficient As target response variables, characteristic parameters such as stirring speed, shear rate, wastewater discharge rate, and pollutant influent rate are denoted as follows: Furthermore, time synchronization and scale normalization are performed on each characteristic parameter to ensure comparability of characteristic parameters with different dimensions within the same analytical space.

[0091] While keeping the other characteristic parameters unchanged, a preset amplitude perturbation is applied to any characteristic parameter, and the corresponding change in the reaction interface update capability coefficient is recalculated within the perturbation range. The preset amplitude perturbation is taken from a preset proportion of the allowable change range of the characteristic parameter under the current operating state, which is used to reflect the influence of the characteristic parameter on the interface update capability within the actual adjustable range.

[0092] Based on the changes before and after the disturbance, the sensitivity coefficient of the feature parameter to the interface update capability is calculated, and its expression is as follows: In the formula, Indicates the first Each feature parameter in A normalized sensitivity coefficient for the interface update capability at any given time; the sensitivity coefficient is used to characterize the degree of relative change in the interface update capability caused by a unit relative parameter change; Indicates at time The first factor affecting the ability to update the interface The current value of each feature parameter. This indicates that within the critical interval, for the first... The controlled perturbation applied to each characteristic parameter Indicates the first The relative variation of each characteristic parameter is used to eliminate the influence of differences in the dimensions and absolute magnitudes of different parameters on the sensitivity coefficient calculation results. Indicates at time The reaction interface renewal capacity coefficient of the reaction unit. This indicates that only the first one is changed. The change in the reaction interface update capability coefficient, assuming one characteristic parameter and keeping the other characteristic parameters constant. This represents the relative magnitude of change in the interface update capability, used to characterize the degree to which the interface update capability responds to perturbations in the feature parameters.

[0093] The decay rate per unit time of the reaction interface renewal capacity coefficient, the reaction interface renewal capacity coefficient, the control parameters, and the control constraint matrix are input into the pre-constructed wastewater treatment control and prediction model, and the control parameters are output. The wastewater treatment unit is automatically controlled based on the output control parameters.

[0094] The training method for the wastewater treatment regulation and prediction model includes:

[0095] Obtain a historical control dataset for wastewater treatment regulation; train a pre-built wastewater treatment regulation prediction model based on the historical control dataset; the historical control dataset is recorded when the regulation meets the standards.

[0096] The historical regulation dataset includes the decay rate per unit time of the reaction interface update capability coefficients of P groups, the reaction interface update capability coefficients, regulation parameters and regulation constraint condition matrices, and the corresponding regulation values ​​of each regulation parameter, where P is a positive integer greater than 0.

[0097] Gradient boosting regression tree was selected as the wastewater treatment regulation and prediction model, and the training method is as follows:

[0098] Before training the model, complete the initial configuration and set the initial hyperparameters. The initial hyperparameters include: 100-150 decision trees, 4-6 maximum depth of a single tree, 8-12 minimum number of samples for node splitting, 3 maximum number of features to consider during splitting, 0.05-0.1 learning rate, and 0.1-0.2 regularization coefficient (L2).

[0099] During the model training phase, mean squared error is used as the loss function to measure the deviation between the predicted and actual control parameters. The decay rate of the reaction interface update capability coefficient, the reaction interface update capability coefficient, the control parameters, and the control constraint matrix are used as model input features, and the corresponding actual control parameters are used as prediction targets. Model training is carried out based on training data. The construction of each new tree is aimed at fitting the regression residual of the training set loss function. The optimal splitting feature is selected by the mean squared error criterion, such as prioritizing the decay rate of the reaction interface update capability coefficient as the splitting feature. The samples are divided into different child nodes until the stopping conditions are met, such as reaching the preset maximum depth, the number of child node samples being less than the minimum number of samples, or the splitting gain being lower than the threshold.

[0100] The gradient descent method is used to optimize the weights of the leaf nodes of each new tree. The contribution weight of the new tree to the prediction result of the final control parameter is adjusted by the learning rate, so as to avoid a single tree dominating the prediction process and ensure the stability of the model prediction.

[0101] In the hyperparameter optimization phase, a Bayesian optimization method is employed to search for the optimal combination of hyperparameters within a pre-defined optimization range. The optimization objective is to minimize the root mean square error of the validation set. Specifically, the hyperparameter optimization range is defined as follows: 80-180 decision trees, maximum depth of a single tree of 3-7, learning rate of 0.03-0.12, and regularization coefficient (L2) of 0.05-0.25.

[0102] An early stopping mechanism is introduced during training. The collected training data is divided into training set, validation set and test set in a ratio of 7:2:1. The root mean square error of the validation set is calculated every 20 trees. When the root mean square error of the validation set decreases by less than 0.001 after 3 consecutive iterations and a total of 60 trees, the model training is stopped to avoid the model overfitting the training data.

[0103] After training, save the model parameters with the lowest root mean square error in the validation set to ensure that the model has stable predictive ability for different decay rates, reaction interface update coefficients, and control parameters under different constraints.

[0104] During the model evaluation phase, the trained model is validated using a test set. The root mean square error, mean absolute error, and coefficient of determination (R²) are calculated. If the root mean square error of the test set is ≤ 5% of the target control parameter adjustment amount, the mean absolute error is ≤ 3% of the target control parameter adjustment amount, and the coefficient of determination (R²) is ≥ 0.9, then the model performance evaluation meets the standards and can be put into practical deployment for automatic control of wastewater treatment units.

[0105] In this embodiment, after confirming the mass transfer-limited operating state, the control focus can be automatically shifted from improving the reaction rate to restoring the interface update capability. Sensitivity analysis is used to screen the control parameters that have the greatest impact on the interface update capability. A control constraint matrix is ​​established by combining the equipment capability and the operational safety boundary, thereby avoiding over-adjustment of ineffective control parameters, reducing the risk of system fluctuations, and improving the pertinence and stability of the control strategy.

[0106] Furthermore, the reaction interface update capacity coefficient, its decay rate per unit time, control parameters, and control constraint matrix are all incorporated into the wastewater treatment control prediction model. This ensures that the control parameters output by the model not only conform to historical effective control experience but are also constrained by the current operating mechanism and safety constraints. This effectively avoids the problem of model prediction results being distorted or out of control under complex operating conditions. Finally, by introducing a gradient boosting regression tree model trained based on historical compliance operating conditions and combining hyperparameter optimization and early shutdown mechanisms, the wastewater treatment control prediction model constructed in this invention has high prediction accuracy and stability under different interface update capacity decay levels and operating constraints. This enables the wastewater treatment system to maintain strong adaptive control capabilities under fluctuating influent loads or sudden changes in operating conditions, thereby reducing the probability of system instability, shortening the recovery cycle, and significantly reducing the risk of exceeding emission standards.

[0107] Example 2:

[0108] Please see Figure 4This invention provides a technical solution: a multi-parameter intelligent control system for industrial wastewater treatment, which is used to implement the aforementioned multi-parameter intelligent control method for industrial wastewater treatment. The system includes:

[0109] The mechanism trajectory construction module simultaneously acquires the apparent viscosity change trend, reaction rate change trend, and dosage change trend within the reaction unit during continuous operation of wastewater treatment, and constructs the reaction control mechanism evolution trajectory curve based on the temporal correlation between the three.

[0110] The regulation boundary identification module extracts the effective response range of the reaction rate to changes in drug dosage based on the evolution trajectory curve of the reaction control mechanism, constructs dynamic response features that reflect the actual boundary of the regulation action, and identifies the critical interval where the drug dosage behavior begins to fail.

[0111] The mass transfer limitation identification module, after identifying the critical interval, combines the apparent viscosity change trend to quantitatively characterize the reaction interface update capacity coefficient, and determines whether the reaction process has entered a limited operating state dominated by mass transfer conditions.

[0112] The control strategy switching module, after confirming the restricted operating state, shifts the control focus from increasing the reaction rate to restoring the reaction interface update capability coefficient, and forms a matching control constraint condition matrix.

[0113] The model prediction and control module inputs the unit time decay rate of the reaction interface update capacity coefficient, the reaction interface update capacity coefficient, the control parameters, and the control constraint matrix into the pre-constructed wastewater treatment control and prediction model, outputs the control parameters and control values, and automatically controls the wastewater treatment unit based on the output control parameters and control values.

[0114] The foregoing has shown and described the basic principles, main features, and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The embodiments and descriptions in the specification are merely illustrative of the principles of the invention. Various changes and modifications can be made to the invention without departing from its spirit and scope, and all such changes and modifications fall within the scope of the present invention as claimed. The scope of protection of the present invention is defined by the appended claims and their equivalents.

Claims

1. A multi-parameter intelligent control method for industrial wastewater treatment, characterized in that, include: During the continuous operation of wastewater treatment, the trends of apparent viscosity, reaction rate, and dosage within the reaction unit are acquired simultaneously, and the evolution trajectory curve of the reaction control mechanism is constructed based on the temporal correlation between the three. Based on the evolution trajectory curve of the reaction control mechanism, the effective response range of the reaction rate to the change in drug dosage is extracted, dynamic response characteristics reflecting the actual boundary of the regulatory action are constructed, and the critical interval at which the drug dosage behavior begins to fail is identified. After identifying the critical interval, the reaction interface renewal capacity coefficient is quantitatively characterized by combining the apparent viscosity change trend to determine whether the reaction process has entered a restricted operating state dominated by mass transfer conditions. After confirming the restricted operating state, the focus of regulation is shifted from increasing the reaction rate to restoring the reaction interface update capability coefficient, and a matching regulation constraint matrix is ​​formed. The decay rate per unit time of the reaction interface renewal capacity coefficient, the reaction interface renewal capacity coefficient, the control parameters, and the control constraint matrix are input into the pre-constructed wastewater treatment control and prediction model, and the control parameters are output. The wastewater treatment unit is automatically controlled based on the output control parameters.

2. The multi-parameter intelligent control method for industrial wastewater treatment according to claim 1, characterized in that, The method for constructing the evolution trajectory curve of the reaction control mechanism includes: Apparent viscosity time series, reaction rate time series, and dosage time series are collected synchronously according to a unified sampling period and time calibration is performed to ensure that each parameter has a one-to-one correspondence on the same time axis. With a preset sliding window length, the apparent viscosity time series, reaction rate time series, and dosage time series are processed by sliding window to calculate the changing trend of each parameter in the continuous operating range. The response delay time of the preset reaction process is used to align the trend of reaction rate change with the corresponding trend of dosage change in time. The dynamic response coefficient of reaction rate to dosage change is calculated. The dynamic response coefficient is then analyzed in conjunction with the trend of apparent viscosity change to calculate the modulation relationship of dynamic response coefficient with viscosity change. Based on the apparent viscosity change trend, reaction rate change trend, and reaction response index after viscosity modulation, a multi-parameter time-series correlation vector is constructed within the same time window. The multi-parameter time-series correlation vector is sequentially arranged in a continuous time interval to form a trajectory curve reflecting the evolution of reaction control characteristics over time, thus obtaining the reaction control mechanism evolution trajectory curve.

3. The multi-parameter intelligent control method for industrial wastewater treatment according to claim 2, characterized in that, The method for identifying the critical interval at which the dosing action begins to fail includes: The evolution trajectory curve of the reaction control mechanism is divided into multiple operating intervals by the length of the sliding window. The ratio between the change in reaction rate and the change in dosage in adjacent operating intervals is calculated to form a response coefficient sequence that characterizes the degree of change in reaction rate caused by a unit change in dosage. The response coefficient sequence is mapped to the apparent viscosity change trend. When the response coefficient remains above the preset stable threshold in a continuous time interval, it is determined that the reaction process in the corresponding time interval is still in a state of effective response to the change in drug addition. Mark the continuous time intervals that satisfy the state determination conditions for effective response on the evolution trajectory curve of the reaction control mechanism to form the effective response interval of the reaction rate to the change in drug dosage; When the response coefficient shows a decreasing trend over three consecutive operating intervals on the evolution trajectory curve of the reaction control mechanism, and the apparent viscosity change trend within the corresponding operating interval shows a monotonically increasing characteristic over three consecutive operating intervals, the corresponding operating interval is marked as the critical interval of the effective response interval.

4. The multi-parameter intelligent control method for industrial wastewater treatment according to claim 1, characterized in that, The method for quantitatively characterizing the reaction interface renewal capacity coefficient by combining the apparent viscosity change trend includes: Within the critical range, the apparent viscosity data of the reaction unit are continuously collected. A sliding time window Δt was used to smooth the collected data, and the apparent viscosity change trend over time was calculated. ; Based on the apparent viscosity change trend Define the function for updating the reaction interface coefficient.

5. The multi-parameter intelligent control method for industrial wastewater treatment according to claim 2, characterized in that, The method for determining whether the reaction process has entered a restricted operating state dominated by mass transfer conditions includes: calculating the decay rate per unit time of the reaction interface renewal capacity coefficient; when the decay rate per unit time of the reaction interface renewal capacity coefficient is greater than the dynamic response coefficient, it is determined that the mass transfer process has become the dominant factor restricting the reaction; otherwise, the mass transfer process has not become the dominant factor restricting the reaction.

6. The multi-parameter intelligent control method for industrial wastewater treatment according to claim 1, characterized in that, The method for generating the regulation constraint matrix includes: Under continuous operation conditions, a set of characteristic parameters affecting the reaction interface update capability coefficient are synchronously collected within the critical interval; Within the established critical range, the normalized perturbation analysis method is used to calculate the sensitivity coefficient of each characteristic parameter to the interface update capability. The sensitivity coefficient is used to characterize the contribution of a unit change in characteristic parameter to the change in interface update capability, thus forming a sensitivity mapping relationship between characteristic parameter and interface update capability. A preset sensitivity threshold is set, and feature parameters with sensitivity coefficients higher than the sensitivity threshold are selected as the dominant control parameters for interface update capability, and the dominant control parameters are marked as an adjustable parameter set. After determining the set of adjustable parameters, for each control parameter, a corresponding control constraint matrix is ​​established in combination with the equipment capacity boundary, operational safety constraints and process stability requirements.

7. The multi-parameter intelligent control method for industrial wastewater treatment according to claim 6, characterized in that, The characteristic parameters include at least stirring speed, shear rate, wastewater discharge rate, and pollutant influent rate. Each characteristic parameter is obtained in time series form and aligned with the reaction interface update capacity coefficient on the same time axis.

8. The multi-parameter intelligent control method for industrial wastewater treatment according to claim 6, characterized in that, The expression for the control constraint matrix is: ,in Indicates the first One control parameter, and These represent the lower and upper limits of the control parameter during the current operating phase, respectively. The lower and upper limits are used to restrict the range of change of the control parameter during the process of restoring the interface update capability.

9. The multi-parameter intelligent control method for industrial wastewater treatment according to claim 6, characterized in that, The method for calculating the sensitivity coefficients of each feature parameter to the interface update capability is as follows: Update the reaction interface capability coefficient As target response variables, the stirring speed, shear rate, wastewater discharge rate, and pollutant influent rate are denoted as... And perform time synchronization and scale normalization on each feature parameter; While keeping the other characteristic parameters unchanged, a preset amplitude perturbation is applied to any characteristic parameter, and the change in the corresponding reaction interface update capability coefficient is recalculated within the perturbation range. Based on the changes before and after the disturbance, the sensitivity coefficient of the feature parameters to the interface update capability is calculated.

10. A multi-parameter intelligent control system for industrial wastewater treatment, used to implement the multi-parameter intelligent control method for industrial wastewater treatment as described in any one of claims 1-9, characterized in that, The system includes: The mechanism trajectory construction module simultaneously acquires the apparent viscosity change trend, reaction rate change trend, and dosage change trend within the reaction unit during continuous operation of wastewater treatment, and constructs the reaction control mechanism evolution trajectory curve based on the temporal correlation between the three. The regulation boundary identification module extracts the effective response range of the reaction rate to changes in drug dosage based on the evolution trajectory curve of the reaction control mechanism, constructs dynamic response features that reflect the actual boundary of the regulation action, and identifies the critical interval where the drug dosage behavior begins to fail. The mass transfer limitation identification module, after identifying the critical interval, combines the apparent viscosity change trend to quantitatively characterize the reaction interface update capacity coefficient, and determines whether the reaction process has entered a limited operating state dominated by mass transfer conditions. The control strategy switching module, after confirming the restricted operating state, shifts the control focus from increasing the reaction rate to restoring the reaction interface update capability coefficient, and forms a matching control constraint condition matrix. The model prediction and control module inputs the unit time decay rate of the reaction interface update capacity coefficient, the reaction interface update capacity coefficient, the control parameters, and the control constraint matrix into the pre-constructed wastewater treatment control and prediction model, outputs the control parameters and control values, and automatically controls the wastewater treatment unit based on the output control parameters and control values.

Images