Short-cut denitrification desulfurization wastewater treatment method and system

Abstract

The invention belongs to the field of wastewater treatment, and particularly relates to a short-cut denitrification desulfurization wastewater treatment method and system. Determining a main sliding mode surface parameter of the sliding mode controller and a weight coefficient of a process parameter according to the inlet water quality index and the short-cut denitrification desulfurization target parameter, and adjusting the main sliding mode surface parameter and the weight coefficient if a track deviates from an effective operation region corresponding to the target parameter; performing weighted combination on the process parameters by using the weight coefficient to generate a state variable, and calculating according to the state variable and the main sliding mode surface parameters to obtain a sliding mode function value of the current state; if the absolute value of the sliding mode function value is greater than a switching threshold value, generating a control instruction by adopting a first approaching control mode, otherwise, generating the control instruction by adopting a second approaching control mode; and sending the control instruction to an execution unit in the wastewater treatment system. And the accuracy of short-cut denitrification desulfurization wastewater treatment control is effectively improved.

Description

Technical Field

[0001] This application belongs to the field of wastewater treatment, and particularly relates to a method and system for treating short-cut denitrification desulfurization wastewater. Background Art

[0002] The rapid development of industrialization and urbanization has led to the discharge of a large amount of nitrogen- and sulfur-containing wastewater, causing increasingly severe pollution to the water environment, such as water eutrophication, black and odorous phenomena, and threats to aquatic ecosystems and human health. Traditional biological nitrogen removal processes, such as complete nitrification-denitrification, although widely used, have disadvantages such as high energy consumption, large carbon source requirements, high sludge production, and long reaction cycles. Short-cut nitrification-denitrification can significantly reduce the aeration energy consumption by about 25%, reduce the organic carbon source requirement by about 40%, and reduce sludge production by controlling the oxidation of ammonia nitrogen at the nitrite nitrogen stage and then performing denitrification. However, the stable operation of short-cut nitrification requires strict operating conditions, such as dissolved oxygen, pH, temperature, free ammonia / nitrous acid concentration. If not properly controlled, it is prone to collapse or excessive production of by-products. To address these complexities and uncertainties, traditional control methods such as proportional-integral-derivative (PID) control or simple open-loop / closed-loop logic control are actually difficult to achieve good control effects due to their linear characteristics and over-simplification of the system model, especially when dealing with biological wastewater treatment systems with strong nonlinearity, large time delays, and multi-variable coupling characteristics. Sliding mode control (SMC) has advantages such as strong robustness to system parameter uncertainties and external disturbances, and fast response, and has been widely studied and applied in the field of nonlinear system control. However, SMC cannot guarantee in the reaching phase, the control output is prone to chattering phenomenon, and it is difficult for fixed sliding mode surface parameters to adapt to the dynamic changes of the system working conditions, especially in systems such as short-cut denitrification desulfurization that require high precision of operating conditions and have complex internal biochemical reaction paths. How to dynamically adjust the control strategy and suppress chattering is the key to realizing the efficient and stable operation of short-cut denitrification desulfurization wastewater treatment. Summary of the Invention

[0003] Aiming at the problems in the existing short-cut denitrification desulfurization wastewater treatment method, such as simple control strategy, inability to effectively predict the influence of disturbances, and inherent chattering of sliding mode control, resulting in unstable operation and difficulty in maintaining ideal short-cut denitrification and high-efficiency desulfurization effects for a long time. This application provides a method for treating short-cut denitrification desulfurization wastewater, including: Collect the influent water quality indexes of the wastewater in real time, and synchronously obtain the process parameters during the wastewater treatment process; predict the trajectory of the key biochemical indexes based on the influent water quality indexes and process parameters, determine the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the influent water quality indexes and the short-cut denitrification and desulfurization target parameters. If the trajectory deviates from the effective operation area corresponding to the key biochemical indexes, adjust the main sliding mode surface parameters and the weight coefficients. Use the weight coefficients to perform weighted combination on the process parameters to generate state variables, and calculate the sliding mode function value of the current state according to the state variables and the main sliding mode surface parameters; if the absolute value of the sliding mode function value is greater than the switching threshold, generate a control instruction by using the first approaching control method, otherwise generate a control instruction by using the second approaching control method. Send the control instruction to the execution unit in the wastewater treatment system.

[0004] Optionally, the predicting the trajectory of the key biochemical indexes based on the influent water quality indexes and process parameters includes: Obtain the influent water quality index data and process parameter data within a preset time length before the current moment. For each influent water quality index, obtain an index sequence, and for each process parameter, obtain a parameter sequence. Input the index sequence and the parameter sequence into a pre-trained recurrent neural network to obtain the trajectory of the key biochemical indexes; the key biochemical indexes include the effluent nitrite nitrogen accumulation rate and the effluent sulfide concentration.

[0005] Optionally, the determining the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the influent water quality indexes and the short-cut denitrification and desulfurization target parameters includes: Obtain the preset short-cut denitrification and desulfurization target parameters, and the target parameters include the target effluent nitrite nitrogen concentration range, the lower threshold of the target nitrite nitrogen accumulation rate, and the upper threshold of the target effluent sulfide concentration. Based on the current working condition level reflected by the influent water quality indexes and the preset target parameters, determine the main sliding mode surface parameters through a preset multi-objective optimization solving program. The main sliding mode surface parameters include the coefficients of each state variable error term in the sliding mode switching function; and synchronously determine the weight coefficients of the process parameters, and the weight coefficients are used to perform weighting on the process parameters to form the state variables of the sliding mode control.

[0006] Optionally, the adjusting the main sliding mode surface parameters and the weight coefficients if the trajectory deviates from the effective operation area corresponding to the key biochemical indexes includes: Obtain the effective operating region of key biochemical indicators, compare the predicted trajectory of key biochemical indicators with the effective operating region, and if the time when the predicted trajectory exceeds the boundary of the effective operating region within a preset determination duration in the future exceeds a preset ratio, it is considered a deviation. When a deviation occurs, adjust the weight coefficients of the main sliding mode surface parameters and process parameters currently in use according to the direction and degree of the deviation through a preset adjustment rule.

[0007] Optionally, calculating the sliding mode function value of the current state based on the state variable and the main sliding mode surface parameters includes: Substitute the state variable and the main sliding mode surface parameters into to calculate the sliding mode function value of the current state , where is the main sliding mode surface parameter, is the state variable error term.

[0008] Optionally, generating the control instruction by adopting the first reaching control method includes: The change rate of the sliding mode function The calculation formula is: , where s is the sliding mode function value, , are preset control parameters, is the power exponent parameter and , is a positive constant.

[0009] Based on the model of the actuator to be controlled and the change rate of the sliding mode function obtain the control instruction.

[0010] Optionally, generating the control instruction by adopting the second reaching control method includes: The desired change rate of the sliding mode function The calculation formula is: , where s is the sliding mode function value, is the saturation function, C is the value corresponding to the working condition level, , are the control gain parameters adjusted with C, is the boundary layer thickness parameter adjusted with C; Based on the model of the actuator to be controlled and the change rate of the sliding mode function obtain the control instruction.

[0011] The present invention also provides a short-range denitrifying and desulfurizing wastewater treatment system, including: A sliding mode determination unit, configured to collect the influent water quality indexes of the wastewater in real time, and synchronously obtain the process parameters during the wastewater treatment process; predict the trajectory of key biochemical indexes based on the influent water quality indexes and process parameters, determine the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the influent water quality indexes and the short-cut denitrification and desulfurization target parameters, and if the trajectory deviates from the effective operation area corresponding to the key biochemical indexes, adjust the main sliding mode surface parameters and the weight coefficients; A control instruction generation unit, configured to generate a state variable by performing weighted combination on the process parameters by using the weight coefficients, and calculate the sliding mode function value of the current state according to the state variable and the main sliding mode surface parameters; if the absolute value of the sliding mode function value is greater than the switching threshold, generate a control instruction by using a first approaching control method, otherwise generate a control instruction by using a second approaching control method; A control unit, configured to send the control instruction to an execution unit in the wastewater treatment system.

[0012] Optionally, the predicting the trajectory of the key biochemical indexes based on the influent water quality indexes and the process parameters includes: Obtaining the influent water quality index data and the process parameter data within a preset time length before the current moment, obtaining an index sequence for each influent water quality index, and obtaining a parameter sequence for each process parameter; Inputting the index sequence and the parameter sequence into a pre-trained recurrent neural network to obtain the trajectory of the key biochemical indexes; the key biochemical indexes include the effluent nitrite nitrogen accumulation rate and the effluent sulfide concentration.

[0013] Optionally, the determining the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the influent water quality indexes and the short-cut denitrification and desulfurization target parameters includes: Obtaining the preset short-cut denitrification and desulfurization target parameters, where the target parameters include the target effluent nitrite nitrogen concentration range, the lower threshold of the target nitrite nitrogen accumulation rate, and the upper threshold of the target effluent sulfide concentration; Based on the current working condition level reflected by the influent water quality indexes and the preset target parameters, determining the main sliding mode surface parameters through a preset multi-objective optimization solving program, where the main sliding mode surface parameters include the coefficients of each state variable error term in the sliding mode switching function; and synchronously determining the weight coefficients of the process parameters, where the weight coefficients are used to perform weighted combination on the process parameters to form the state variables of the sliding mode control.

[0014] Optionally, the adjusting the main sliding mode surface parameters and the weight coefficients if the trajectory deviates from the effective operation area corresponding to the key biochemical indexes includes: Obtain the effective operating region of key biochemical indicators, compare the predicted trajectory of the key biochemical indicators with the effective operating region, and if the time when the predicted trajectory exceeds the boundary of the effective operating region within a preset determination duration in the future exceeds a preset ratio, it is considered a deviation. When a deviation occurs, adjust the weight coefficients of the main sliding mode surface parameters and process parameters currently in use according to the direction and degree of the deviation through a preset adjustment rule.

[0015] Optionally, calculating the sliding mode function value of the current state based on the state variable and the main sliding mode surface parameters includes: Substitute the state variable and the main sliding mode surface parameters into to calculate the sliding mode function value of the current state , where is the main sliding mode surface parameter, is the state variable error term.

[0016] Optionally, generating a control instruction by adopting the first reaching control method includes: The change rate of the sliding mode function The calculation formula is: , where s is the sliding mode function value, , are preset control parameters, is the power exponent parameter and , is a positive constant.

[0017] Based on the model of the actuator to be controlled and the change rate of the sliding mode function obtain the control instruction.

[0018] Optionally, generating a control instruction by adopting the second reaching control method includes: The change rate of the sliding mode function The calculation formula is: , where s is the sliding mode function value, is the saturation function, C is the value corresponding to the working condition level, , are control gain parameters adjusted with C, is the boundary layer thickness parameter adjusted with C; Based on the model of the actuator to be controlled and the change rate of the sliding mode function obtain the control instruction.

[0019] By collecting the influent water quality and system process parameters in real time, this application predicts the trends of key biochemical indicators, and dynamically adjusts the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the prediction results and target parameters, enabling the control system to anticipatively respond to changes in influent load and internal state fluctuations, and optimize the control strategy in advance, thereby enhancing the adaptability and robustness to disturbances. Moreover, by using the sliding mode surface parameters and weight coefficients determined based on multi-objective optimization, it is possible to better coordinate multiple control objectives in the short-cut denitrification and desulfurization processes and achieve the optimization of the overall treatment efficiency. In addition, the strategy of switching different reaching control laws according to the magnitude of the sliding mode function value proposed in this invention effectively takes into account both the rapid reaching of the state and the chattering suppression near the sliding mode surface, ensuring both the rapidity and accuracy of the control and improving the smoothness of the control signal. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 is a flowchart of an embodiment of this application; Figure 2 is a structural diagram of the trajectory for predicting key biochemical indicators; Figure 3 is a schematic diagram of the effective operating region; Figure 4 is a schematic diagram of the exponential reaching law, the saturation function reaching law, the improved exponential reaching law, and the improved saturation function reaching rate. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] Next, the technical solutions in the embodiments of this application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of this application. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all the embodiments. Based on the embodiments in this application, all other embodiments obtained by those skilled in the art without making creative efforts belong to the scope of protection of this application.

[0022] The terms "first", "second" and their corresponding term numbers in the specification and claims of this application and the above-mentioned drawings are used to distinguish similar objects, and do not necessarily need to be used to describe a specific order or sequence. It should be understood that such terms can be interchanged under appropriate circumstances, which is only a way of distinguishing objects with the same attributes when describing the embodiments of this application. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusion, so that a process, method, system, product or device including a series of units does not necessarily have to be limited to those units, but may include other units not clearly listed or inherent to these processes, methods, products or devices.

[0023] In addition, in the description of the present application, unless otherwise specified, the meaning of "a plurality of" is two or more. The term "and / or" or the character " / " in the present application is merely an association relationship describing associated objects, indicating that three relationships may exist. For example, A and / or B, or A / B, may represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0024] Specific embodiments, such as Figure 1 shown, provides a short-range denitrifying desulfurization wastewater treatment method, including: Step 1, collect the influent water quality indicators of the wastewater in real time, and synchronously obtain the process parameters during the wastewater treatment process; predict the trajectory of the key biochemical indicators based on the influent water quality indicators and process parameters, determine the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the influent water quality indicators and the short-range denitrifying desulfurization target parameters, and if the trajectory deviates from the effective operation area corresponding to the key biochemical indicators, adjust the main sliding mode surface parameters and the weight coefficients; Obtain the influent water quality indicators entering the wastewater treatment system through sensors or monitoring instruments. The influent water quality indicators include, but are not limited to, chemical oxygen demand, biochemical oxygen demand, ammonia nitrogen concentration, total nitrogen concentration, total phosphorus concentration, sulfate concentration, pH value, temperature, conductivity, turbidity, and influent flow rate, etc. While collecting the influent water quality, the operating state parameters of each key process unit or reaction area inside the wastewater treatment system are also monitored and recorded. The process parameters include, but are not limited to, dissolved oxygen concentration, oxidation-reduction potential, pH value, temperature, sludge concentration, sludge age, nitrite nitrogen concentration, nitrate nitrogen concentration, sulfide concentration, volatile fatty acid concentration, aeration volume, stirrer speed, internal and external reflux flow rates, chemical agent dosage, etc.

[0025] Predict the trend of key biochemical reactions in the future through a model. Among them, the key biochemical indicators refer to parameters that can directly or indirectly characterize the efficiency and stability of the core reactions of short-range denitrification and desulfurization. In one embodiment, the key biochemical indicators include, but are not limited to, the effluent nitrite nitrogen accumulation rate, the ratio of effluent nitrate nitrogen to nitrite nitrogen, the effluent sulfide removal rate, the activity indicators of specific microbial populations, etc. The trajectory of the key biochemical indicators can be generated in various ways. For example, a biochemical reaction kinetic model is constructed based on the activated sludge model and the coupled sulfur cycle model, or support vector regression and artificial neural networks are used for prediction. Figure 2 Shows a generation method based on RNN. The short-range denitrifying desulfurization target parameters are the desired operating effects, such as the target effluent water quality standards, such as the desired range of nitrite accumulation rate, the desired degree of sulfide oxidation, etc. For example, based on the data at time t1, the model predicts that after 2 hours, the effluent nitrite nitrogen accumulation rate will drop to 50%.

[0026] The methods for determining the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the influent water quality indexes and the short-cut denitrifying desulfurization target parameters include, but are not limited to, determining based on experience, determining based on model optimization, etc. Among them, determining based on model optimization specifically means calculating the optimal parameter combination through optimization algorithms such as genetic algorithms and particle swarm optimization to meet specific performance indexes, such as minimizing the tracking error, the fastest response speed, the best anti-disturbance ability, etc. The optimization objective can combine the current influent water quality and the set treatment target. More specifically, the weight coefficient of the process parameter of dissolved oxygen is 0.6, and the weight coefficient of the carbon source dosage is 0.4.

[0027] The effective operation region is composed of multi-dimensional target intervals of key biochemical indexes. For example, the nitrite nitrogen concentration is between A and B mg / L, and at the same time, the sulfide concentration is lower than C mg / L. In one embodiment, the determination of the trajectory deviation is based on the integral value or the maximum deviation amplitude of the distance between the predicted trajectory and the effective operation region. When it is determined that a deviation occurs, the main sliding mode surface parameters and the weight coefficients are adjusted through an optimization algorithm based on gradient descent, aiming to minimize the future deviation between the predicted trajectory and the target region, and the backpropagation error is used to update these parameters; or a case-based reasoning method is adopted to learn from historical similar deviation events and their successful corrective measures and adjust the parameters.

[0028] Step 2, use the weight coefficients to perform weighted combination on the process parameters to generate state variables, and calculate the sliding mode function value of the current state according to the state variables and the main sliding mode surface parameters; if the absolute value of the sliding mode function value is greater than the switching threshold, generate a control instruction by using the first reaching control method, otherwise generate a control instruction by using the second reaching control method; The wastewater treatment process involves many interacting parameters, such as temperature, pH, flow rate, concentrations of various substances, etc. Observing these parameters alone is not sufficient to comprehensively understand the system state and may even be misleading. For the continuously monitored process parameters, their respective weight coefficients are used for weighted combination to generate one or more state variables. For example, the dissolved oxygen is 0.5 mg / L, the carbon source dosage is 0.6 kg / h, and the state variable is 0.54. In one embodiment, the process parameters are all converted to 0-1. Based on the new state variables and the main sliding mode surface parameters, the sliding mode function value of the current state is obtained. Assuming that the expected state variable X is stable at the target value of 0.6, and the state variable calculated above is 0.54, if the sliding mode function s is defined as , the sliding mode function value is -0.06. The sliding mode function quantifies the deviation between the current state and the target state defined by the main sliding mode surface. Comparing the absolute value of the sliding mode function value with a preset switching threshold, if the absolute value of the sliding mode function value exceeds this switching threshold, it indicates that the system significantly deviates from the expected operating trajectory. In this case, the first reaching control method will be adopted to generate a control instruction. The first control strategy is usually more controllable and can quickly guide the system state back to the sliding mode surface. On the contrary, if the absolute value of the sliding mode function value is less than or equal to the switching threshold, it means that the system operating state is closer to the ideal state or already on the sliding mode surface. The second reaching control method will be adopted to generate a control instruction. The second control strategy is usually smoother and aims to maintain the system on the sliding mode surface while minimizing chattering or oscillation to ensure stable and precise control. The switching between the two control modes enables the system to quickly respond to large disturbances and perform fine adjustment when approaching the target state.

[0029] In an alternative embodiment, the step of generating a control instruction by adopting the first reaching control method includes: Adopting an exponential reaching law Calculating the desired rate of change of the sliding mode function , where and are preset first reaching control parameters, is the sign function; Based on the model of the actuator to be controlled , by solving the control instruction u is obtained, where f(x) and g(x) are functions representing the system characteristics, and x is the current state.

[0030] By the exponential reaching law formula Calculate the desired rate of change of the sliding mode function . Where s is the sliding mode function value calculated at the current moment, and are preset first reaching control parameters, and they are both positive numbers; affects the smoothness and response speed of the reaching process, then ensures the coerciveness and finite-time reachability of the reaching process, is the sign function, which takes the value of 1 when its parameter is greater than 0, -1 when less than 0, and 0 when equal to 0. represents at what rate the desired sliding mode function s changes towards 0 in the current state. For example, if the current s = 2, , , then , indicating that it is expected that s decreases towards 0 at a rate of -1.1. Based on the desired rate of change of the sliding mode function , combined with the mathematical model of the actuator to be controlled to solve the actual control instruction u, and this model is expressed as the derivative of the sliding mode function s The differential equation describing the relationship between the system state x and the control input u , where f(x) represents the part of the system's own dynamic characteristics and is not directly affected by the control input u, and g(x) represents the gain of the control input u on the system dynamics; x represents the current overall state of the system, which can include multiple state variables. To make the actual change rate of the sliding mode function equal to the desired change rate , that is, to achieve , directly solve the above equation to obtain the control instruction u. The control instruction is obtained through calculation, and the calculated u is the specific operation amount finally applied to the corresponding execution units in the wastewater treatment system, such as pumps, valves, aerators, etc.

[0031] In an alternative embodiment, the generation of the control instruction using the second reaching control method includes: Using the saturation function reaching law Calculate the desired change rate of the sliding mode function , where and are preset second reaching control parameters, is the preset boundary layer thickness parameter, is the saturation function; Based on the model of the actuator to be controlled , solve to obtain the control instruction u, where f(x) and g(x) are functions characterizing the system characteristics, and x is the current state.

[0032] Calculate the desired change rate of the sliding mode function through the reaching law formula based on the saturation function , where s is the value of the sliding mode function calculated at the current moment, and and are preset second reaching control parameters, both of which are positive numbers and are used to adjust the reaching speed and the system response characteristics, is a preset boundary layer thickness parameter, , which defines a thin layer area around the sliding mode surface s = 0. When , , this term in the reaching law becomes , which is a linear term about s, making the control action smooth. When , the state is outside the boundary layer but still satisfies switching condition, , this term becomes , providing a strong action. It represents that in the current state, it is expected that the sliding mode function s changes towards 0 in a smoother manner. Similar to the first reaching control method, based on this expected change rate of the sliding mode function , the u calculated by combining the mathematical model of the actuator to be controlled is the control operation amount applied to the corresponding execution unit in the wastewater treatment system, which will not be elaborated here.

[0033] Step 3: Send the control instruction to the execution unit in the wastewater treatment system.

[0034] Send the control instruction to the corresponding execution unit in the wastewater treatment system. For example, if the control instruction is to increase the carbon source feeding rate by 10%, then this instruction is sent to the carbon source feeding pump to increase its output frequency; if the control instruction is to reduce the dissolved oxygen in the aerobic tank by 0.5 mg / L, then this instruction is sent to the actuator that controls the valve opening of the aeration blower to reduce the valve opening.

[0035] In the complex biochemical treatment process of short-cut denitrifying desulfurization, the change of state often has the characteristics of time lag and non-linearity. Relying solely on instantaneous data for control adjustment is difficult to cope with the impact brought by the fluctuation of influent water quality, and it is also impossible to predict potential risks in advance. In an alternative embodiment, the trajectory of the key biochemical indicators predicted based on the influent water quality indicators and process parameters includes: Obtain the influent water quality indicator data and process parameter data within a preset time length before the current moment. For each influent water quality indicator, obtain an indicator sequence, and for each process parameter, obtain a parameter sequence; Input the indicator sequence and the parameter sequence into a pre-trained recurrent neural network to obtain the trajectory of the key biochemical indicators; the key biochemical indicators include the effluent nitrite nitrogen accumulation rate and the effluent sulfide concentration.

[0036] Specifically, continuously monitor and record the influent water quality indicators entering the wastewater treatment system and the process parameters at each key point inside the system. For example, data is collected at fixed time intervals. For each monitored influent water quality indicator, such as influent COD, influent sulfide concentration, influent nitrate concentration, etc., a preset time length before the current time point is intercepted. For example, all the data points collected within the past 24 hours or the past 48 hours are used to form an indicator sequence for this indicator. Similarly, for each monitored process parameter, such as the pH value of reaction tank A, the ORP value of reaction tank B, the aeration rate, the dosing rate of external carbon source, etc., all the data points within the same preset time length are intercepted to form the corresponding parameter sequence. These processed indicator sequences and parameter sequences are used as inputs and fed into a pre-trained recurrent neural network model. The recurrent neural network model has learned the complex non-linear mapping relationship and time dynamics between the influent conditions, operating parameters and effluent biochemical indicators in a large amount of historical operation data during the training phase. When the new indicator sequences and parameter sequences are input, the RNN model will deduce step by step according to the patterns it has learned and output a sequence of predicted values of the key biochemical indicators for a future period of time. This sequence of predicted values constitutes the future change direction of the key biochemical indicators. Those skilled in the art should know that when there are multiple water quality indicators, the water quality indicator sequence input into the recurrent neural network model should be a matrix, and the number of rows of the matrix is equal to the number of water quality indicators. The same applies to the parameter sequence, which will not be elaborated here.

[0037] The operating state and external disturbances are dynamically changing. If the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters are set to be fixed, then when facing different working condition loads, the optimal state may not be achieved, and even the control may become unstable. In an optional embodiment, the determining of the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the influent water quality indicators and the short-cut denitrifying desulfurization target parameters includes: Obtain the preset short-cut denitrifying desulfurization target parameters, where the target parameters include the target effluent nitrite nitrogen concentration range, the lower threshold of the target nitrite nitrogen accumulation rate, and the upper threshold of the target effluent sulfide concentration; Based on the current working condition level reflected by the influent water quality indicators and the preset target parameters, through a preset multi-objective optimization solving program, determine the main sliding mode surface parameters, where the main sliding mode surface parameters include the coefficients of each state variable error term in the sliding mode switching function; and synchronously determine the weight coefficients of the process parameters, where the weight coefficients are used to weight the process parameters to form the state variables of the sliding mode control.

[0038] Specifically, obtain and load the target parameters of the preset short-cut denitrifying desulfurization process. The target parameters include, but are not limited to, the target effluent nitrite nitrogen concentration range, the lower threshold of the target nitrite nitrogen accumulation rate, and the upper threshold of the target effluent sulfide concentration. Continuously analyze the real-time collected influent water quality indicators, and classify or quantify the operating load represented by the current influent water quality through a preset model, such as determining that the current belongs to low load, medium load, or high shock load, etc. Then, use the evaluated current operating condition level information and the preset short-cut denitrifying desulfurization target parameters as input conditions to a preset multi-objective optimization solving program, such as a genetic algorithm or a particle swarm algorithm. The core goal of this optimization program is to find a set of optimal sliding mode controller parameters on the premise of meeting multiple treatment objectives. The output includes the main sliding mode surface parameters, which define the coefficients of the error terms of each state variable in the sliding mode switching function, that is, the difference between the actual state and the desired state, thus determining the characteristics of the system state converging to the sliding mode surface. At the same time, the optimization program will also synchronously determine the weight coefficients of each process parameter such as aeration volume, carbon source dosage, internal reflux ratio, etc. These weight coefficients are used to perform weighted combination on the original process parameters to form more comprehensive and sensitive state variables in the sliding mode control, ensuring that these state variables can effectively reflect the system dynamics that have the greatest impact on the control objectives.

[0039] In an optional embodiment, the adjusting the main sliding mode surface parameters and the weight coefficients if the trajectory deviates from the effective operation region corresponding to the key biochemical indicators includes: Obtain the effective operation region of the key biochemical indicators, compare the predicted trajectory of the key biochemical indicators with the effective operation region. If the time when the predicted trajectory exceeds the boundary of the effective operation region within a preset determination time length in the future exceeds a preset ratio, it is considered a deviation. When a deviation occurs, adjust the currently used main sliding mode surface parameters and the weight coefficients of the process parameters according to the direction and degree of the deviation through a preset adjustment rule.

[0040] Obtain the boundary of the effective operation region of the key biochemical indicators and determine whether the predicted trajectory deviates. The effective operation region is jointly defined by various short-cut denitrifying desulfurization target parameters. For example, the target effluent nitrite nitrogen concentration range, the lower limit of the target nitrite nitrogen accumulation rate, and the upper limit of the target effluent sulfide concentration, etc. Compare the predicted trajectory with the boundary of the effective operation region. If the predicted trajectory shows that within a preset determination time length in the future, the total time length during which one or some key biochemical indicators will exceed the boundary of their effective operation region exceeds a preset ratio. For example, if the time in a state exceeding the standard accounts for more than 20% of the preset determination time length, a deviation has occurred, as Figure 3As shown. When a deviation occurs in the system, analyze the specific situation of the deviation, including the direction of the deviation, such as whether the predicted value of the nitrite nitrogen accumulation rate is too low or the predicted value of the sulfide concentration is too high, and the degree of the deviation, such as the magnitude and duration of the predicted value exceeding the boundary. According to the characteristics of these deviations, call the preset adjustment rule library. These rules are usually obtained based on expert experience and historical data analysis. For example, if it is predicted that the sulfide concentration in the effluent will continue to exceed the standard, it is indicated to increase the weight coefficient of the process parameters related to carbon source addition and correspondingly adjust a certain coefficient of the main sliding mode surface to strengthen the desulfurization effect. The adjusted parameters will be immediately applied to the subsequent control calculations to pull the future operating trajectory of the system back into the effective operating area.

[0041] In an optional embodiment, calculating the sliding mode function value of the current state according to the state variable and the main sliding mode surface parameter includes: Substitute the state variable and the main sliding mode surface parameter into to calculate the sliding mode function value of the current state , where is the main sliding mode surface parameter, and is the state variable error term.

[0042] Specifically, the state variable Xi is a synthesis of several key process parameters in the wastewater treatment process, and the main sliding mode surface parameter is determined according to the influent water quality, target parameters, and multi-objective optimization. They are the coefficients corresponding to each state variable in the sliding mode switching function. Substitute the obtained state variables and the corresponding main sliding mode surface parameters into the expression of the predefined sliding mode function or switching function for calculation. In an optional embodiment, the sliding mode function is in the form of a linear combination . For each state variable , it will be multiplied by the corresponding coefficient to obtain the contribution term of this state variable to the total value of the sliding mode function. Finally, perform an algebraic sum of all these contribution terms to obtain the sliding mode function value corresponding to the system state at the current moment.

[0043] In order for the system state to approach at a higher rate when it is far from the sliding mode surface and smoothly transition when it is close to the sliding mode surface, effectively reducing the chattering problem caused by the discontinuity of the sign function in the traditional reaching law. In an optional embodiment, adopting the first reaching control method to generate a control instruction includes: The change rate of the sliding mode function The calculation formula is: , where s is the sliding mode function value, , are preset control parameters, is the power exponent parameter and , is a positive constant; Based on the model of the actuator to be controlled and the change rate of the sliding mode function the control command is obtained.

[0044] The key power exponent such that when the sliding mode variable 0 < < 1, is larger relative to the value of, thus providing a faster convergence rate than the traditional linear reaching term, which helps to achieve the system state reaching the sliding mode surface in a finite time. When the system state is far from the sliding mode surface, i.e., is much larger than the positive constant , , approximates the traditional constant-speed reaching part , overcoming the system uncertainties and external disturbances. When the system state is very close to the sliding mode surface, i.e., becomes very small and approaches 0, in this case, the second part approaches , near the sliding mode surface, the discontinuous sign function provides a switching control action that is replaced. The smooth transition can suppress the control chattering caused by the high-frequency switching of the sign function. This embodiment focuses on fast reaching when far from the sliding mode surface to ensure rapidity; when approaching the sliding mode surface, it focuses on smooth transition to ensure the steady-state performance of the system and reduce chattering, as Figure 4 shown. Among them, is a small positive number, in a more specific embodiment, .

[0045] Under different working conditions, the content of the control command should also be different. In an alternative embodiment, the control command is generated by adopting the second reaching control method, including: The change rate of the sliding mode function The calculation formula is: , where s is the value of the sliding mode function, is the saturation function, C is the value corresponding to the working condition level, , are the control gain parameters adjusted with C, is the boundary layer thickness parameter adjusted with C; Based on the model of the actuator to be controlled and the change rate of the sliding mode function the control command is obtained.

[0046] When the working condition level C indicates that the future water quality may deteriorate and the load will increase significantly, reduce the boundary layer thickness , making the sliding mode control more accurate near the sliding mode surface, and at the same time increasing the control gains and To provide stronger approaching power and convergence speed, so as to quickly respond to and suppress the expected disturbances. On the contrary, when the prediction shows that the system is operating smoothly, the water quality is good, and the key indicators are stable within the target range, the system will increase the boundary layer thickness , making the control action softer to maximize the suppression of chattering, and can reduce and to reduce the control energy consumption, as Figure 4 shown.

[0047] More specifically, the larger the operating condition level C, the worse the operating condition. For example, 0 represents low load, 0.5 represents medium load, and 1 represents high impact load. In an alternative embodiment, , , , those skilled in the art should know that these formulas are merely exemplary, and the present application is not limited to this calculation method.

[0048] In several embodiments provided by the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division, and there may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some characteristic data can be ignored or not executed. Another point, the displayed or discussed coupling or direct coupling or communication connection between each other can be through some interfaces, and the indirect coupling or communication connection of the device or unit can be in electrical, mechanical, or other forms.

[0049] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place, or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0050] In addition, the functional units in each embodiment of the present application can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit. The above integrated units can be implemented in the form of hardware or in the form of software functional units.

[0051] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: various media that can store program codes, such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs.

Claims

1. A short-cut denitrification and desulfurization wastewater treatment method, characterized in that, Including: Real-time collect the influent water quality indicators of the wastewater and synchronously obtain the process parameters during the wastewater treatment process; Predict the trajectory of key biochemical indicators based on the influent water quality indicators and process parameters, determine the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the influent water quality indicators and the short-cut denitrification and desulfurization target parameters. If the trajectory deviates from the effective operation area corresponding to the key biochemical indicators, adjust the main sliding mode surface parameters and the weight coefficients; Use the weight coefficients to perform weighted combination on the process parameters to generate state variables, and calculate the sliding mode function value of the current state according to the state variables and the main sliding mode surface parameters; if the absolute value of the sliding mode function value is greater than the switching threshold, generate a control instruction by using the first reaching control method, otherwise generate a control instruction by using the second reaching control method; Send the control instruction to the execution unit in the wastewater treatment system.

2. The method according to claim 1, characterized in that, The predicting the trajectory of key biochemical indicators based on the influent water quality indicators and process parameters includes: Obtain the influent water quality indicator data and process parameter data within a preset time length before the current moment. For each influent water quality indicator, obtain an indicator sequence, and for each process parameter, obtain a parameter sequence; Input the indicator sequence and the parameter sequence into a pre-trained recurrent neural network to obtain the trajectory of key biochemical indicators; the key biochemical indicators include the effluent nitrite nitrogen accumulation rate and the effluent sulfide concentration.

3. The method according to claim 1, characterized in that, The determining the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the influent water quality indicators and the short-cut denitrification and desulfurization target parameters includes: Obtain the preset short-cut denitrification and desulfurization target parameters, and the target parameters include the target effluent nitrite nitrogen concentration range, the lower threshold of the target nitrite nitrogen accumulation rate, and the upper threshold of the target effluent sulfide concentration; Based on the current working condition level reflected by the influent water quality indicators and the preset target parameters, determine the main sliding mode surface parameters through a preset multi-objective optimization solution program. The main sliding mode surface parameters include the coefficients of each state variable error term in the sliding mode switching function; and synchronously determine the weight coefficients of the process parameters. The weight coefficients are used to perform weighted combination on the process parameters to form the state variables of the sliding mode control.

4. The method according to claim 1, characterized in that, The if the trajectory deviates from the effective operation area corresponding to the key biochemical indicators, adjust the main sliding mode surface parameters and the weight coefficients includes: Obtain the effective operation area of the key biochemical indicators, compare the predicted trajectory of the key biochemical indicators with the effective operation area. If the time when the predicted trajectory exceeds the boundary of the effective operation area within a preset determination time length in the future exceeds the preset ratio, it is considered a deviation; When a deviation occurs, adjust the currently used main sliding mode surface parameters and the weight coefficients of the process parameters according to the direction and degree of the deviation through a preset adjustment rule.

5. The method according to claim 1, characterized in that, The calculating the sliding mode function value of the current state according to the state variables and the main sliding mode surface parameters includes: Substitute the state variable and the main sliding mode surface parameter into to calculate the sliding mode function value of the current state , where is the main sliding mode surface parameter is the state variable error term 6. The method according to claim 1, characterized in that, The generating a control instruction by using the first reaching control method includes: Sliding mode function change rate The calculation formula is as follows: , where s is the sliding mode function value, and are preset control parameters, is the power exponent parameter and , is a positive constant; Based on the model of the actuator to be controlled and the change rate of the sliding mode function The control command is obtained.

7. The method according to claim 1, characterized in that, The generating a control instruction by using the second reaching control method includes: Sliding mode function change rate The calculation formula is as follows: , where s is the sliding mode function value, is the saturation function, C is the value corresponding to the operating condition level, , are the control gain parameters adjusted with C, is the boundary layer thickness parameter adjusted with C; Based on the model of the actuator to be controlled and the change rate of the sliding mode function The control command is obtained.

8. A short-cut denitrification and desulfurization wastewater treatment system, characterized in that, Including: A sliding mode determination unit for real-time collecting the influent water quality indexes of the wastewater and synchronously obtaining the process parameters during the wastewater treatment process; Predicting the trajectory of key biochemical indexes based on the influent water quality indexes and process parameters, determining the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the influent water quality indexes and the short-cut denitrification and desulfurization target parameters, and adjusting the main sliding mode surface parameters and the weight coefficients if the trajectory deviates from the effective operation area corresponding to the key biochemical indexes; A control instruction generation unit for generating a state variable by weighted combination of the process parameters using the weight coefficients, calculating the sliding mode function value of the current state according to the state variable and the main sliding mode surface parameters; if the absolute value of the sliding mode function value is greater than the switching threshold, generating a control instruction by using a first reaching control method, otherwise generating a control instruction by using a second reaching control method; A control unit for sending the control instruction to an execution unit in the wastewater treatment system.

9. The system according to claim 8, characterized in that, The predicting the trajectory of key biochemical indexes based on the influent water quality indexes and process parameters includes: Obtaining the influent water quality index data and process parameter data within a preset time length before the current moment, obtaining an index sequence for each influent water quality index, and obtaining a parameter sequence for each process parameter; Inputting the index sequence and the parameter sequence into a pre-trained recurrent neural network to obtain the trajectory of key biochemical indexes; the key biochemical indexes include the effluent nitrite nitrogen accumulation rate and the effluent sulfide concentration.

10. The system according to claim 8, characterized in that, The determining the main sliding mode surface parameters of the sliding mode controller and the weight coefficients of the process parameters according to the influent water quality indexes and the short-cut denitrification and desulfurization target parameters includes: Obtaining the preset short-cut denitrification and desulfurization target parameters, where the target parameters include the target effluent nitrite nitrogen concentration range, the lower threshold of the target nitrite nitrogen accumulation rate, and the upper threshold of the target effluent sulfide concentration; Based on the current working condition level reflected by the influent water quality indexes and the preset target parameters, determining the main sliding mode surface parameters through a preset multi-objective optimization solution program, where the main sliding mode surface parameters include the coefficients of each state variable error term in the sliding mode switching function; and synchronously determining the weight coefficients of the process parameters, where the weight coefficients are used to weighted-combine the process parameters to form the state variable of the sliding mode control.

Images