Parameter tracking control method for sewage treatment equipment

Abstract

The invention relates to the technical field of sewage treatment automation control, in particular to a sewage treatment equipment parameter tracking control method. The method comprises the steps of data acquisition, load prediction, state evaluation, prediction control and feedback correction. The system determines a biochemical operation state by acquiring water inlet parameters and real-time data of a reaction area; the core of the method is to construct a virtual water mass tracking model to generate a future load impact curve, evaluate biological reaction inertia based on an oxygen uptake rate, and output actual oxygen demand and metabolic response delay time; after the impact arrival time is aligned with the metabolic delay, generating an elastic tracking envelope by using a model predictive control algorithm, adaptively adjusting the aeration equipment, and performing closed-loop correction based on the effluent quality; according to the method, conversion from passive hysteresis control to accurate feedforward prediction is realized, inflow disturbance and metabolic inertia can be coordinated in advance, and insufficient oxygen supply and excessive aeration are effectively avoided.

Description

Technical Field

[0001] This invention relates to the field of automated control technology for wastewater treatment, specifically a method for tracking and controlling parameters of wastewater treatment equipment. Background Technology

[0002] In the current wastewater treatment operating environment, biological treatment systems often face the impact of sudden increases in influent load; the treatment system needs to continuously collect water quality and flow data on the influent side and status parameters on the reaction zone side, and rely on the aeration equipment at the bottom to supply oxygen to the biological reaction zone in order to maintain the metabolic degradation capacity of microorganisms. To control the oxygen supply process, existing solutions generally adopt a rigid control strategy based on a fixed dissolved oxygen setpoint. Although this solution can maintain basic operation under stable conditions, it cannot predict the spatiotemporal propagation trajectory of the load and does not take into account the metabolic response inertia of the biological system and the optimal efficiency range of the equipment. This passive adjustment mode carries the risk of insufficient oxygen supply before the load arrives. Complex operating disturbances cause the system response to be sluggish, leading to an increase in the reactive power of the blower and the risk of unstable effluent water quality. Therefore, how to coordinate influent disturbance with biochemical metabolic characteristics to achieve feedforward prediction and flexible tracking of control parameters for wastewater treatment equipment has become an urgent technical problem to be solved. Summary of the Invention

[0003] The purpose of this invention is to provide a parameter tracking and control method for wastewater treatment equipment, addressing the following technical problems: Existing wastewater treatment oxygen supply control technologies suffer from a lack of feedforward prediction capability in predicting the spatiotemporal propagation trajectory of load when facing sudden surges in influent load, and in comprehensively considering the metabolic response inertia of biological systems and the optimal efficiency range of equipment. Therefore, there is an urgent need for a wastewater treatment equipment parameter tracking and control method that can coordinate influent disturbances with biochemical metabolic characteristics to achieve feedforward prediction, and combine this with equipment energy efficiency for flexible tracking and feedback correction. The objective of this invention can be achieved through the following technical solutions: A method for tracking and controlling parameters of wastewater treatment equipment, comprising: The system acquires influent flow rate data, influent water quality data, and preset effluent water quality tolerance boundaries. The system includes a reaction zone and execution equipment configured in the reaction zone, including aeration equipment. Acquire real-time dissolved oxygen data, redox potential data, and sludge concentration data in the reaction zone; Based on influent flow rate data, influent water quality data, and spatial parameters of the reaction zone, a virtual water mass tracking model is constructed, and a future load impact curve with timestamps and spatial coordinates is generated through the virtual water mass tracking model. Based on real-time dissolved oxygen data, redox potential data, and sludge concentration data, the oxygen uptake rate is calculated, the biological reaction inertia state is assessed based on the oxygen uptake rate, and the actual oxygen demand and metabolic response delay time are output. Align the arrival time of the future load impact curve with the metabolic response delay time on the time axis, combine the actual oxygen demand, and use the model predictive control algorithm to generate an elastic tracking envelope of the equipment control parameters that changes over time. The flexible tracking envelope of the equipment control parameters is converted into equipment execution instructions. Based on the equipment execution instructions and the preset optimal efficiency range of the equipment, the operating frequency path is determined within the flexible tracking envelope of the equipment control parameters. The operating frequency path is used to control the oxygen supply of the aeration equipment to match the actual oxygen demand. Based on the operating frequency path adjustment execution equipment, and collecting the adjusted effluent water quality data, the effluent water quality data is fed back to the model predictive control algorithm for parameter correction, forming a parameter tracking control closed loop.

[0004] Optionally, based on influent flow rate data, influent water quality data, and spatial parameters of the reaction zone, a virtual water mass tracking model is constructed. This virtual water mass tracking model generates future load impact curves with timestamps and spatial coordinates, including: Extract the instantaneous change rate and peak characteristics of influent flow rate data and influent water quality data within a preset time window; The instantaneous rate of change and peak characteristics are mapped to a preset spatiotemporal three-dimensional grid to construct a virtual water mass tracking model; The flow trajectory of water masses in physical space is calculated using a virtual water mass tracking model; Based on the flow trajectory, the arrival time and load intensity of the water mass to the preset target area in the reaction zone are predicted, and a future load impact curve with timestamps and spatial coordinates is generated.

[0005] Optionally, based on real-time dissolved oxygen data, redox potential data, and sludge concentration data, the oxygen uptake rate is calculated. The biological reaction inertia state is assessed based on the oxygen uptake rate, and the actual oxygen demand and metabolic response delay time are output, including: Input real-time dissolved oxygen data, oxidation-reduction potential data, and sludge concentration data into the preset activated sludge model; The oxygen uptake rate at the current moment is calculated by solving the kinetic equations in the activated sludge model. The oxygen uptake rate is compared with a preset activity threshold range to determine the biological reaction inertia state. The actual oxygen demand and metabolic response delay time corresponding to the biological reaction inertia state are queried and extracted from the preset state mapping table.

[0006] Optionally, the oxygen uptake rate is compared with a preset activity threshold range to determine the biological reaction inertia state, including: If the oxygen uptake rate is less than the preset lower limit activity threshold, the biological reaction inertia state is determined to be a starvation state. If the oxygen uptake rate is not less than the lower limit activity threshold and not greater than the preset upper limit activity threshold, then the biological reaction inertia state is determined to be a suitable state. If the oxygen uptake rate is greater than the upper limit of the activity threshold, the biological reaction inertia state is determined to be an overload state.

[0007] Optionally, the arrival time of the future load shock curve is aligned with the metabolic response delay time on the time axis, and combined with the actual oxygen demand, a model predictive control algorithm is used to generate a time-varying elastic tracking envelope of the equipment control parameters, including: Input the future load impact curve, actual oxygen demand, and metabolic response delay time into the model predictive control algorithm; The model predictive control algorithm is used to extrapolate the control trajectory within a future time period in a preset virtual environment; The predicted water quality value of the control trajectory is subtracted from the preset effluent water quality tolerance boundary to obtain the deviation of the effluent index; the pre-built water quality and dissolved oxygen sensitivity mapping table is called to extract the dissolved oxygen fluctuation amplitude corresponding to the deviation of the effluent index, and the safe elastic range of allowable dissolved oxygen fluctuation is calculated based on the fluctuation amplitude. Based on the safety elasticity range, an elastic tracking envelope of device control parameters that varies over time is generated.

[0008] Optionally, the control trajectory for a future time period can be extrapolated in a preset virtual environment using a model predictive control algorithm, including: Identify the load change trend in the future load impact curve and extract the load change rate at adjacent time points; Combine the biological response inertia state with the corresponding control strategy; If the load change rate is less than zero and the absolute value is greater than the preset fluctuation threshold, the load change trend is determined to be decreasing. When the load change trend is decreasing and the biological reaction inertia state is suitable, a sliding trajectory that allows dissolved oxygen to decrease is generated. If the load change rate is greater than zero and the absolute value is greater than the preset fluctuation threshold, the load change trend is determined to be upward. When the load change trend is upward or the biological reaction inertia state is overloaded, a compensation trajectory to maintain or increase dissolved oxygen is generated. If the above conditions are not met, the current control trajectory will be maintained.

[0009] Optionally, the flexible tracking envelope of the equipment control parameters is converted into equipment execution instructions. Based on the equipment execution instructions and a preset optimal efficiency range for the equipment, the operating frequency path is determined within the flexible tracking envelope of the equipment control parameters, including: Extract boundary parameters from the elastic tracking envelope of the equipment control parameters; Obtain the current operating frequency of the execution device and the pre-stored energy consumption curve, and match the boundary parameters with the current operating frequency and energy consumption curve; Within the range defined by the elastic tracking envelope of the equipment control parameters, search for the frequency node that has the highest degree of overlap with the equipment's optimal efficiency range; Connect frequency nodes to generate operating frequency paths.

[0010] Optionally, the execution device adjusted based on the operating frequency path includes: Convert the operating frequency path into an electrical signal; The electrical signal is sent to the programmable logic controller that is communicatively connected to the execution device; The operating frequency of the execution device can be smoothly adjusted by a programmable logic controller.

[0011] Optionally, collecting adjusted effluent water quality data and feeding this data back to the model predictive control algorithm for parameter correction includes: The absolute difference between the effluent water quality data and the preset effluent water quality tolerance boundary is calculated as the deviation value; The model predictive control algorithm has a built-in objective function that includes target penalty weights. If the deviation value is greater than the preset deviation threshold, the target penalty weights of the model predictive control algorithm will be updated. If the deviation value is not greater than the preset deviation threshold, the current penalty weight of the model predictive control algorithm is maintained.

[0012] Compared with the prior art, the present invention has the following beneficial effects: 1. This method generates future load impact curves with timestamps and spatial coordinates by constructing a virtual water mass tracking model. Simultaneously, it calculates oxygen uptake based on real-time dissolved oxygen, oxidation-reduction potential, and sludge concentration data, assesses the biological reaction inertia state, and outputs the actual oxygen demand and metabolic response delay time. The arrival time of the future load impact curve is aligned with the metabolic response delay time on the time axis. This overcomes the lag and passivity of traditional rigid control based on fixed dissolved oxygen setpoints. This method can coordinate the spatiotemporal propagation trajectory of influent disturbances with the metabolic response inertia of the biochemical system, planning and initiating the oxygen supply path in advance before the actual impact of high loads, thus avoiding insufficient oxygen supply before the load arrives. Furthermore, it promptly adjusts back after the load passes, avoiding over-aeration and achieving precise feedforward predictive control. 2. This method utilizes a model predictive control algorithm combined with actual oxygen demand to generate an elastic tracking envelope of equipment control parameters that varies over time. This envelope is then converted into equipment execution commands. Within the range defined by the envelope and combined with a preset optimal efficiency range for the equipment, a smooth operating frequency path for adjusting the execution equipment is searched and determined. This overcomes the technical deficiency of maintaining a single fixed target value in existing technologies, transforming complex operating risks into a safe elastic range that allows for dissolved oxygen fluctuations. By prioritizing the selection of frequency nodes with the highest overlap with the optimal efficiency range within the envelope that satisfies the effluent boundary constraints for smooth adjustment, the method effectively avoids irregular fluctuations in equipment frequency between the upper and lower boundaries, reduces reactive power losses of equipment such as blowers, and balances the needs of wastewater biochemical treatment with the energy efficiency optimization of underlying equipment. 3. This method collects effluent water quality data based on the adjusted operating frequency path, calculates the deviation between this data and the preset effluent water quality tolerance boundary, and if the deviation exceeds the preset threshold, feeds the data back to the model predictive control algorithm to trigger a weight update mechanism for parameter correction. This solves the problem of model prediction error accumulation caused by mass transfer efficiency decay or operating condition drift in biochemical treatment systems under long-term disturbances under complex operating conditions. By using real effluent results to perform closed-loop correction of the feedforward prediction and execution strategy, the control model can adapt to changes in actual operating conditions, establishing the reliability of the system's control benchmark over long periods and reducing the risk of effluent water quality instability. Attached Figure Description

[0013] The present invention will be further explained below with reference to the accompanying drawings and embodiments: Figure 1 This is a flowchart of the method of the present invention. Detailed Implementation

[0014] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to specific embodiments.

[0015] like Figure 1 As shown, a method for tracking and controlling parameters of wastewater treatment equipment includes: The system acquires influent flow rate data, influent water quality data, and preset effluent water quality tolerance boundaries. The system includes a reaction zone and execution equipment configured in the reaction zone, including aeration equipment. Acquire real-time dissolved oxygen data, redox potential data, and sludge concentration data in the reaction zone; Based on influent flow rate data, influent water quality data, and spatial parameters of the reaction zone, a virtual water mass tracking model is constructed, and a future load impact curve with timestamps and spatial coordinates is generated through the virtual water mass tracking model. Based on real-time dissolved oxygen data, redox potential data, and sludge concentration data, the oxygen uptake rate is calculated, the biological reaction inertia state is assessed based on the oxygen uptake rate, and the actual oxygen demand and metabolic response delay time are output. Align the arrival time of the future load impact curve with the metabolic response delay time on the time axis, combine the actual oxygen demand, and use the model predictive control algorithm to generate an elastic tracking envelope of the equipment control parameters that changes over time. The flexible tracking envelope of the equipment control parameters is converted into equipment execution instructions. Based on the equipment execution instructions and the preset optimal efficiency range of the equipment, the operating frequency path is determined within the flexible tracking envelope of the equipment control parameters. The operating frequency path is used to control the oxygen supply of the aeration equipment to match the actual oxygen demand. Based on the operating frequency path adjustment execution equipment, and collecting the adjusted effluent water quality data, the effluent water quality data is fed back to the model predictive control algorithm for parameter correction, forming a parameter tracking control closed loop.

[0016] This embodiment provides a parameter tracking and control mechanism for wastewater treatment equipment. Specifically, the following description uses the biological treatment section of a municipal wastewater treatment plant with a treatment capacity of 100,000 tons / day as the main scenario. This treatment section includes an inlet channel, a selection zone, an aeration reaction zone, a secondary sedimentation tank, and a blower frequency conversion control system connected to the aeration reaction zone. When industrial wastewater flows in at night, the plant experiences a sudden increase in influent load. Traditional control methods based on fixed dissolved oxygen setpoints are prone to problems such as insufficient oxygen supply before the high load and over-aeration after the high load has passed. Therefore, this embodiment is adopted for continuous control. Specifically, the control system collects two types of data: data from the influent side and data from the reaction zone side. The influent side data includes influent flow rate data and influent water quality data, which may include at least one of chemical oxygen demand (COD), ammonia nitrogen, and total nitrogen. Simultaneously, it pre-stores effluent water quality tolerance boundaries, for example, setting the upper limit for effluent ammonia nitrogen to 5 mg / L and the upper limit for effluent COD to 50 mg / L. The reaction zone side data includes real-time dissolved oxygen data, oxidation-reduction potential (ORP) data, and sludge concentration data. For ease of explanation, let's assume that at a certain moment, the influent flow rate is 5200 m³ / h, the influent COD is 310 mg / L, and the influent ammonia nitrogen is 38 mg / L; the dissolved oxygen in the front, middle, and rear sections of the reaction zone are 0.9 mg / L, 1.4 mg / L, and 1.8 mg / L, respectively; the ORPs are -85 mV, -30 mV, and 40 mV, respectively; and the sludge concentration is 3500 mg / L. After acquiring the above data, the system constructs a virtual water mass tracking model. Here, a water mass can be understood as a specific volume of inlet unit with similar load characteristics within a certain time window or an inlet unit within a preset time window. To establish a simulation test scenario, the incoming water for 20 consecutive minutes can be divided into 4 virtual water masses, denoted as W1, W2, W3, and W4, respectively. Among them, W3 corresponds to the industrial wastewater superposition section, which has the highest flow rate increase and the highest ammonia nitrogen concentration. The system adds timestamps and spatial coordinates to each water mass. For example, W3 enters the inlet channel at 21:10, is expected to enter the front section of the aeration reaction zone at 21:32, and reach the middle section at 21:56. This forms a future load impact curve, which at least contains information representing a specific time, a specific spatial area, and the corresponding load intensity in the future. Subsequently, the system calculates the oxygen uptake rate based on real-time dissolved oxygen data, redox potential data, and sludge concentration data, and assesses the inertia state of the biological reaction accordingly; for ease of explanation, the current oxygen uptake rate output by the activated sludge model can be denoted as... The system then uses the pre-stored mapping relationship to determine the actual oxygen demand at that moment, for example, 430 kg O2 / h. It also determines that the microorganisms need a response time to adjust from the current state to the corresponding metabolic level, for example, a metabolic response delay time of 18 minutes. This delay time is used to describe the internal time lag characteristics that exist in the process of the microbial population building a matching stable oxygen uptake capacity after the external gas supply is forcibly increased. The system aligns the arrival time in the future load impact curve with the metabolic response delay time on the time axis. Using the aforementioned data as an example, if W3 will arrive at the front of the reaction zone at 21:32, and the metabolic response delay time is 18 minutes, the control system will not passively increase aeration or oxygen supply after 21:32, but will instead begin planning the oxygen supply path around 21:14. The control core employs a model predictive control algorithm, no longer simply maintaining a fixed dissolved oxygen target value, but rather searching for a control trajectory within a predicted time domain that ensures the effluent boundary is not breached while maintaining stable output power of the oxygen supply system. The algorithm outputs not a single control point, but an elastic tracking envelope of the equipment control parameters that changes over time. For example, during the period from 21:14 to 21:40, the initial dissolved oxygen is allowed to operate between 1.2 mg / L and 1.8 mg / L, and the middle stage is allowed to operate between 1.5 mg / L and 2.1 mg / L. This envelope provides a safe allowable variation range, rather than a statically set single target point. At the execution layer, the system translates the elastic tracking envelope into equipment execution instructions and determines the operating frequency path based on the blower's optimal efficiency range. For example, the high-efficiency operating range of a certain blower is 38Hz to 44Hz, and the current operating frequency is 36Hz. When the envelope requires a gradual increase in oxygen supply over the next 20 minutes, the system does not directly jump to the maximum frequency or abruptly change to the maximum frequency. Instead, it searches within the envelope range for frequency nodes that meet the oxygen supply demand and fall within the 38Hz to 44Hz range as much as possible, such as selecting 39Hz, 41Hz, and 42Hz sequentially, thus forming a smoothly increasing operating frequency path. The programmable logic controller adjusts the frequency converter according to this path to drive the blower to output the corresponding oxygen supply. After control is executed, the system continues to collect the adjusted effluent water quality data and feeds it back to the model predictive control algorithm for parameter correction, forming a closed loop. For example, if the effluent ammonia nitrogen is detected to be 4.3 mg / L and chemical oxygen demand is 43 mg / L at 22:30, both within the tolerance boundary, the current model parameters are maintained. If the effluent ammonia nitrogen rises to 5.6 mg / L, it indicates that the predictive model's estimate of the actual hydraulic delay or biological activity is lower than the actual value. The system will increase the penalty weight of the corresponding state parameters or recalibrate the delay amount so that the next round of control can intervene earlier. As an anomaly handling mechanism or as an alternative, if the influent sensor experiences a momentary communication interruption or signal loss, the sliding average or reliable redundant sensor data within the most recent preset time window can be used as a substitute. If any of the three parameters—dissolved oxygen, oxidation-reduction potential, and sludge concentration—abnormally deviates from the physically reasonable range, for example, if dissolved oxygen momentarily jumps to 15 mg / L while adjacent times and adjacent measuring points are all around 2 mg / L, it is determined to be a sampling anomaly. In this round, model reconstruction is not triggered; only the frequency path of the previous control cycle is retained and the envelope width is reduced to prevent control instability. If the future load impact curve cannot form a clear peak, the system enters a narrow-band envelope tracking mode to maintain stable oxygen supply with a narrower elastic tracking range. For example, during the nighttime operation of this municipal wastewater treatment plant, the industrial influent gradually increases between 21:00 and 21:20. The virtual water mass tracking module identifies that a high ammonia nitrogen water mass will enter the front section of the reaction zone at 21:32. At the same time, the biological status assessment results indicate that the sludge is currently highly active but has an 18-minute metabolic response delay. Based on this, the system starts to increase the blower frequency in advance at 21:14 and controls the dissolved oxygen in the front section between 1.4 mg / L and 1.7 mg / L, instead of passively increasing it to a fixed 2.0 mg / L after 21:32. After the high-load water mass passes through, the system allows the dissolved oxygen to slowly fall back along the safety envelope according to the trajectory of the subsequent low-load water mass. In this way, the entire aeration process is always coordinated around the arrival time of the future load and the microbial response delay. The purpose of this step is to integrate influent disturbance, microbial metabolic inertia and equipment power limitation characteristics into the same control closed loop, thereby achieving feedforward prediction, flexible tracking and feedback correction of wastewater treatment equipment parameters, and reducing the risk of blower power waste and effluent instability caused by excessive aeration and hysteresis adjustment. In this embodiment, a virtual water mass tracking model is constructed based on influent flow rate data, influent water quality data, and spatial parameters of the reaction zone. The future load impact curve obtained through the virtual water mass tracking model includes: Extract the instantaneous change rate and peak characteristics of influent flow rate data and influent water quality data within a preset time window; The instantaneous rate of change and peak characteristics are mapped to a preset spatiotemporal three-dimensional grid to construct a virtual water mass tracking model; The flow trajectory of water masses in physical space is calculated using a virtual water mass tracking model; Based on the flow trajectory, the arrival time and load intensity of the water mass to the preset target area in the reaction zone are predicted, and a future load impact curve with timestamps and spatial coordinates is generated.

[0017] This embodiment provides a virtual water mass tracking mechanism. Specifically, in the same operating scenario of the aforementioned municipal wastewater treatment plant, relying solely on the instantaneous influent value at a certain moment can easily miss the situation where the load is rapidly increasing but has not yet fully entered the reaction zone. Therefore, this embodiment further performs spatiotemporal mapping on the changes in influent load, so that subsequent control no longer relies solely on the current influent load, but further determines the spatiotemporal distribution trend of a specific load in the reaction zone. Specifically, the system first extracts the instantaneous change rate and peak characteristics of influent flow rate and influent water quality data within a preset time window. For example, the most recent 20 minutes can be taken as the time window, and a sampling segment is formed every 5 minutes, resulting in 4 segments P1, P2, P3, and P4. Let the flow rates corresponding to P1 to P4 be 4800, 4950, 5300, and 5250 m³ / h, respectively, and the corresponding ammonia nitrogen be 28, 29, 41, and 39 mg / L, respectively. Then, the flow rate change rate can be represented by the difference between adjacent segments. For example, P2 increases by 150 m³ / h relative to P1, and P3 increases by 350 m³ / h relative to P2. The ammonia nitrogen peak appears in P3. Based on this, the system identifies P3 as a sudden increase segment, which can be regarded as the central segment of a high-load water mass. The instantaneous rate of change and peak characteristics are mapped onto a preset three-dimensional spatial grid. This three-dimensional spatial grid is not limited to pure geometric coordinates; it can also be a discrete grid composed of flow direction position, pool lateral position, and time-progression layers. For ease of deduction, the reaction zone can be divided into three longitudinal segments Z1, Z2, and Z3, each segment further divided into three lateral units: left, middle, and right, thus forming a 3×3 planar grid. Then, four future prediction time layers T1, T2, T3, and T4 are superimposed to form a three-dimensional grid. If the high-load water mass corresponding to P3 is located in the Z1-middle unit near the inlet at time T1, the system can project it sequentially to the Z1-right, Z2-right, and Z3-middle units at times T2, T3, and T4, respectively, based on the hydraulic residence time and the guiding structure. This completes the mapping from temporal variation characteristics to spatial propagation trajectory. When calculating the flow trajectory of the water mass, the flow velocity in the channel, the effective volume of the pool, the recirculation ratio, and historical calibration parameters can be combined. The specific flow calculation process is as follows: The system extracts the effective volume of the pool in the target reaction section, combines the influent flow rate data at the same moment with the internal recirculation ratio flow rate data of the recirculation pump station, and calculates the instantaneous hydraulic residence time of the section; based on the instantaneous hydraulic residence time, the hydraulic transfer step of the water mass between longitudinal sections is obtained; the historical calibration parameters obtained in advance through historical tracer tests or computational fluid dynamics simulations are fused to perform delay correction on the dispersion effect caused by the guide structure and aeration disturbance, thereby obtaining the average transfer time with spatiotemporal accuracy. For simplification, if the average transfer time from the inlet to the center of Z1 is calculated to be 8 minutes, from Z1 to Z2 20 minutes, and from Z2 to Z3 18 minutes, then after the high-load water mass enters the system at 21:10, it can be predicted to reach Z1 at 21:18, Z2 at 21:38, and Z3 at 21:56. Simultaneously, the system adds load intensity values ​​to each arrival point; for example, the ammonia nitrogen load intensity is 1.0 at Z1, decreases to 0.82 at Z2 due to dilution and mixing, and decreases to 0.67 at Z3. The resulting curve includes both a time axis and spatial location and load decay characteristics. When generating the future load impact curve, it can be represented by several discrete points; for example, the curve node Q1 can be recorded as 21:18, Z1-middle, intensity 1.0; Q2 is 21:38, Z2-right, intensity 0.82; Q3 is 21:56, Z3-middle, intensity 0.67; the subsequent model predictive control algorithm can then know that the future impact will not act synchronously on the entire reaction pool, but will arrive at different regions at different times. As an anomaly handling mechanism or as an alternative, if no obvious peak is detected within the time window, but only fluctuations with an amplitude less than the preset intensity threshold are observed, a low-intensity, gentle water mass can still be formed to prevent the model from losing its feedforward capability due to the absence of peaks. If adjacent segment data contradict each other, for example, a sudden increase in flow but a simultaneous and significant decrease in chemical oxygen demand and ammonia nitrogen, which is inconsistent with the upstream conditions, the credibility of the segment can be reduced through historical consistency verification, and it can participate in the mapping in a weighted manner. If a certain cell of the 3D mesh cannot obtain calibration information due to sensor maintenance, the average propagation parameters of adjacent cells can be temporarily used to supplement it to avoid trajectory breakage. For example, during the nighttime operation of this wastewater treatment plant, in four segments after 21:00, the ammonia nitrogen in P3 jumped from 29 mg / L to 41 mg / L, and the flow rate simultaneously increased to 5300 m³ / h. The system identified this as a high-load center segment. After mapping, the algorithm determined that this high-load water mass would affect the front aeration zone at 21:18, the middle zone at 21:38, and the rear zone at 21:56. Therefore, subsequent control no longer uniformly increased the overall aeration volume of the pool, but instead prioritized the allocation of oxygen resources to the areas that would be impacted. The purpose of this step is to transform the load changes that were originally only observed at the water inlet into spatiotemporal propagation information that can be used for control decisions, thereby enabling feedforward identification of future load impacts and more refined regional regulation. In this embodiment, based on real-time dissolved oxygen data, redox potential data, and sludge concentration data, the oxygen uptake rate is calculated. Based on the oxygen uptake rate, the biological reaction inertia state is assessed, and the actual oxygen demand and metabolic response delay time are output, including: Input real-time dissolved oxygen data, oxidation-reduction potential data, and sludge concentration data into the preset activated sludge model; The oxygen uptake rate at the current moment is calculated by solving the kinetic equations in the activated sludge model. The oxygen uptake rate is compared with a preset activity threshold range to determine the biological reaction inertia state. The actual oxygen demand and metabolic response delay time corresponding to the biological reaction inertia state are queried and extracted from the preset state mapping table.

[0018] This embodiment provides a biological reaction inertia assessment mechanism. Specifically, the previous embodiment was able to predict when a high-load water mass would enter the reaction zone, but knowing only that the load was about to cause an impact was not enough to accurately determine when and how much to increase the oxygen supply. The reason is that when the same incoming water impact acts on sludge groups in different active states, their metabolic response speeds are not the same. Therefore, this embodiment characterizes the current microbial activity and inertia through oxygen uptake rate. Specifically, the system inputs real-time dissolved oxygen data, oxidation-reduction potential data, and sludge concentration data into a preset activated sludge model. The activated sludge model includes Activated Sludge Model 1, Activated Sludge Model 2, or Activated Sludge Model 2D based on the International Water Association (IWA) standards, or a simplified kinetic model based on the standard that includes substrate degradation and microbial growth and decay processes. A simplified kinetic model can be used here; it does not require a complete and complex reaction network at the control level, but at least a calculable relationship between biodegradable substrate, microbial activity, and oxygen demand must be established. For ease of understanding, assume that at a certain moment, the dissolved oxygen in the front section of the reaction zone is 1.0 mg / L, the oxidation-reduction potential is -70 mV, and the sludge concentration is 3600 mg / L; the dissolved oxygen in the middle section is 1.5 mg / L, the oxidation-reduction potential is -20 mV, and the sludge concentration is the same. The model calculates the current oxygen uptake rate based on these inputs. In the specific kinetic calculation process, the execution of this equation follows the following data flow logic: The current biomass tank is characterized by sludge concentration data, and a basic oxygen uptake potential is generated by combining this with the system's preset maximum substrate degradation rate; real-time dissolved oxygen data and redox potential data are used as evaluation inputs for electron acceptor conditions to construct a dimensionless acceptor limiting factor. For example, when the dissolved oxygen concentration is below the preset critical concentration and the redox potential tends towards the reduced state, this acceptor limiting factor decreases proportionally; the basic oxygen uptake potential is multiplied by the acceptor limiting factor, and the system-calibrated endogenous respiration oxygen uptake number is added. The endogenous respiration oxygen uptake number is obtained through offline sludge decay experiments or fitting historical low-load period energy consumption baselines, thereby outputting the current oxygen uptake rate; based on this decomposition process, the calculation result can be simplified to the initial oxygen uptake rate. The oxygen uptake rate in the middle section is ; After obtaining the oxygen uptake rate, the system compares it with a preset activity threshold range to determine the inertial state of the biological response. For simulation calculations or analysis, a three-segment mapping table can be preset: if the oxygen uptake rate is below 25, it corresponds to low activity; if the oxygen uptake rate is between 25 and 45, it corresponds to suitable activity; if the oxygen uptake rate is above 45, it corresponds to high load and high activity. Based on the above example, the first segment of 48 falls into the high load and high activity range, and the middle segment of 35 falls into the suitable activity range. Thus, it can be seen that microorganisms in different regions are not in the same state at the same time. The system queries and extracts the actual oxygen demand and metabolic response delay time corresponding to the biological reaction inertia state from the state mapping table. Continuing the example above, the initial stage is in a high-load, high-activity zone, which can be mapped to an actual oxygen demand of 470 kg O2 / h and a metabolic response delay time of 12 minutes. The middle stage is in a suitable activity zone, which can be mapped to an actual oxygen demand of 350 kg O2 / h and a metabolic response delay time of 18 minutes. The delay time here is not simply the device response time, but rather describes the time required for the biological system as a whole to transition from the current state to a new equilibrium processing capacity. The reason for introducing this mapping is that simply judging oxygen demand based on dissolved oxygen levels can easily lead to misjudgments. For example, if the dissolved oxygen level is only 1.0 mg / L at the beginning, it may seem that the oxygen supply needs to be increased immediately. However, if the oxygen uptake rate is already high at this time, it means that the microorganisms have been fully activated. Over-adjusting the oxygen supply or increasing the oxygen supply beyond the actual demand threshold may only raise the oxygen concentration in the pool without simultaneously increasing the actual degradation rate. Conversely, if the dissolved oxygen level is not low but the oxygen uptake rate is lower than the actual value, it may mean that the microorganisms have not yet reached the normal metabolic level. In this case, feedforward pre-aeration or pre-increasing the oxygen supply is necessary, rather than compensating after the index drops. As an anomaly handling mechanism or as an alternative, if missing input data causes the model to be unable to solve stably, the oxygen uptake rate of the previous valid moment can be used with an additional decay factor as a temporary result; if the obtained oxygen uptake rate is negative or exceeds the reasonable upper limit of the equipment and process, it is judged as an invalid solution, and resampling or median correction is initiated; if the difference in calculation results between different regions exceeds the preset consistency threshold, for example, the oxygen uptake rate of the front section is 50 while that of the adjacent middle section is only 5, and it does not match the water mass trajectory, then a consistency check can be triggered to rule out anomalies caused by local probe contamination or short-term sludge deposition. For example, at 21:15 in this wastewater treatment plant, the high-load water mass had not yet fully entered the front reaction zone, but the oxygen uptake rate in the front zone had already increased to 48%, indicating that the microbial community was already at a high respiration level. The oxygen uptake rate in the middle zone was still 35%, while that in the rear zone was only 22%. Based on this, the control system judged that the current metabolic response delay in the front zone was relatively short, and it could handle the additional load at a higher response rate. The rear zone was closer to the edge of low activity, and it was necessary to reserve an appropriate oxygen supply margin before the subsequent high load spread backward. The purpose of this step is to quantify the hidden biological state of whether microorganisms are ready to handle future loads, thereby determining the actual oxygen demand and metabolic response delay time, and providing a basis for subsequent timeline alignment and flexible control. In this embodiment, comparing the oxygen uptake rate with a preset activity threshold range to determine the biological reaction inertia state includes: If the oxygen uptake rate is less than the preset lower limit activity threshold, the biological reaction inertia state is determined to be a starvation state. If the oxygen uptake rate is not less than the lower limit activity threshold and not greater than the preset upper limit activity threshold, then the biological reaction inertia state is determined to be a suitable state. If the oxygen uptake rate exceeds the upper limit activity threshold, the biological reaction inertia state is determined to be an overload state. This embodiment provides a biological state discrimination mechanism based on threshold segmentation. Specifically, although the above implementation has provided a scheme for obtaining oxygen demand and delay time through a mapping table, if the state classification boundaries are not clear, different operating shifts or different control cycles may have inconsistent interpretations, affecting the continuity of control. Therefore, this embodiment further adopts clear upper and lower limit activity thresholds to classify the biological reaction inertia state into starvation state, adequate state, and overload state. Specifically, the system pre-sets a lower activity threshold and an upper activity threshold; for example, the lower activity threshold can be set to... The upper limit of the activity threshold is set to When the real-time calculated oxygen uptake rate is less than 20, it is determined to be a starvation state; when the oxygen uptake rate is between 20 and 45, it is determined to be a suitable state; when the oxygen uptake rate is greater than 45, it is determined to be an overload state. For ease of explanation, simulation calculations or analyses can be performed on the three regions. Assuming the oxygen uptake rate is 52% in the front section, 31% in the middle section, and 16% in the rear section, the front section is marked as overloaded, the middle section as adequate, and the rear section as starved. At this time, the system will not adopt the same oxygen supply strategy for the three regions, but will treat them differently according to their state classification. Overloaded means that the microorganisms are in a high-intensity respiration stage. If there is a load in the future, oxygen supply needs to be maintained earlier and more stably. Adequate means that the system is in a relatively balanced zone and can be flexibly adjusted within the envelope. Starved means that there is insufficient substrate or the activity has not been activated. If there is no control strategy for continuous high-intensity aeration, the power output of the equipment will exceed the actual oxygen demand matching value, which may also inhibit processes such as denitrification. Compared to simply outputting continuous values, using discrete state labels is also beneficial for the rapid matching of subsequent control strategies; for example, the controller can map starvation, suitability, and overload to three different strategy templates, instead of having to re-evaluate all biochemical behaviors from scratch in each cycle. As an anomaly handling mechanism or as an alternative, if the oxygen uptake rate is exactly equal to the threshold boundary, such as 20 or 45, it is included in the middle or upper interval according to the interval inclusion relationship to avoid classification blind spots or undefined state intervals; if short-term noise causes the oxygen uptake rate to fluctuate frequently near the threshold, a hysteresis interval can be set, such as maintaining the previous state unchanged between 19 and 21 until it exceeds the boundary for two consecutive sampling cycles before switching, to prevent frequent state switching; if multiple areas are overloaded at the same time, but the water inlet tracking shows that the future load is decreasing, the system can regard the state as a residual overload and gradually release oxygen supply in the control trajectory instead of continuing to increase it sharply. For example, at 22:00 in the same wastewater treatment plant, a high-load water mass has passed through the front section and entered the middle section. The oxygen uptake rate in the front section drops to 33%, in the middle section it rises to 47%, and in the back section it is 18%. At this time, the system will correct the front section from overload to suitable, judge the middle section as overloaded, and the back section as starved. The control focus will also shift from the front section to the middle section, instead of the whole tank acting uniformly. The purpose of this step is to transform continuously changing biological activity into a finite set of states that are easy for the controller to invoke by using clear, stable and repeatable threshold segmentation rules, thereby improving the consistency and interpretability of control decisions.

[0019] In this embodiment, the arrival time of the future load shock curve is aligned with the metabolic response delay time on the time axis. Combined with the actual oxygen demand, a model predictive control algorithm is used to generate a time-varying elastic tracking envelope of the equipment control parameters, including: Input the future load impact curve, actual oxygen demand, and metabolic response delay time into the model predictive control algorithm; The model predictive control algorithm is used to extrapolate the control trajectory within a future time period in a preset virtual environment; The predicted water quality value of the control trajectory is subtracted from the preset effluent water quality tolerance boundary to obtain the deviation of the effluent index; the pre-built water quality and dissolved oxygen sensitivity mapping table is called to extract the dissolved oxygen fluctuation amplitude corresponding to the deviation of the effluent index, and the safe elastic range of allowable dissolved oxygen fluctuation is calculated based on the fluctuation amplitude. Based on the safety elasticity range, an elastic tracking envelope of device control parameters that varies over time is generated.

[0020] This embodiment provides a flexible tracking envelope generation mechanism. Specifically, the aforementioned scheme can identify the future load location and the current microbial state, but if a single target value is still output, such as always requiring the dissolved oxygen at the front end to be equal to 2.0 mg / L, it will still return to the traditional rigid control path, making it difficult to reflect the advantage of allowing the index to fluctuate flexibly within a safe range. Therefore, this embodiment generates a safety envelope that changes over time through a model predictive control algorithm, rather than a fixed setpoint. Specifically, the system inputs the future load impact curve, actual oxygen demand, and metabolic response delay time into the model predictive control algorithm. For ease of explanation, the prediction time domain can be set to the next 60 minutes, and the control step size to 5 minutes. Assume that a high-load water mass will affect the initial stage at 21:30 and the middle stage at 21:50; the actual oxygen demand in the initial stage is 460 kg O2 / h, with a metabolic response delay time of 12 minutes; the actual oxygen demand in the middle stage is 360 kg O2 / h, with a metabolic response delay time of 18 minutes. Therefore, when solving the problem, the system does not start satisfying the initial stage demand from 21:30, but instead shifts the control actions for the initial stage forward by 12 minutes and the control actions for the middle stage forward by 18 minutes. When extrapolating future control trajectories in a pre-defined virtual environment, factors such as pool volume, oxygen transfer efficiency, blower output capacity, and effluent boundary constraints can be considered simultaneously. The process of extrapolating and predicting water quality values ​​follows structured data flow rules to avoid untraceable control decision-making processes: For each calculation step within the prediction time domain, the algorithm extracts the input substrate load at the corresponding time from the future load impact curve, combines this with the oxygen supply of the currently traversed candidate control trajectories at that time node, and subtracts the corresponding actual oxygen demand to quantitatively assess the surplus or deficit state of the oxygen supply environment. Simultaneously, the effective microbial uptake capacity at this time domain stage is lagging and reduced using the metabolic response delay time, thereby calculating the actual substrate removal under the dual constraints of oxygen supply and biological conditions. The remaining unremoved substrate is spatiotemporally mixed and delayed based on the pool volume, and the predicted water quality value at the end of the candidate control trajectory is output. To simplify parameter extrapolation, the algorithm can first provide three candidate trajectories: Trajectory A prioritizes equipment output limits, with dissolved oxygen gradually increasing from 1.2 to 1.6 in the initial stage; Trajectory B prioritizes constraint balancing, with dissolved oxygen increasing from 1.4 to 1.9 in the initial stage; Trajectory C prioritizes high redundancy, maintaining dissolved oxygen around 2.1 in the initial stage. The system simulates the predicted values ​​of future effluent ammonia nitrogen under each trajectory. For example, the predicted peak value for trajectory A is 5.3 mg / L, for trajectory B it is 4.6 mg / L, and for trajectory C it is 4.4 mg / L. When the upper limit of the effluent ammonia nitrogen tolerance boundary is 5.0 mg / L, trajectory A has exceeded the preset boundary, while trajectories B and C are both safe. After further comparing the differences in the underlying dynamic power request corresponding to each trajectory, the system selects trajectory B as the central trajectory. The difference between the predicted water quality value of the control trajectory and the limit value of the effluent water quality tolerance boundary is used to calculate the safe elasticity range for dissolved oxygen fluctuations. Continuing with the example above, trajectory B corresponds to a predicted peak of 4.6 mg / L, and the difference between it and the boundary of 5.0 mg / L is 0.4 mg / L, indicating that there is still a certain safety margin. The difference of trajectory C is 0.6 mg / L, which has a larger safety margin but results in a higher level of equipment load. When calculating the safe elasticity range, the system does not subjectively set the range or set it without physical basis. Instead, it calls a preset water quality and dissolved oxygen sensitivity mapping table. This mapping table is pre-constructed from historical steady-state operating data through multiple regression analysis. It is used to characterize the dissolved oxygen adjustment tolerance corresponding to the deviation of a unit effluent index. The system extracts the dissolved oxygen fluctuation amplitude corresponding to the above difference of 0.4 mg / L under the current biological state. For example, the calculated safe fluctuation amplitude is 0.15 mg / L. Then, the system can superimpose this fluctuation amplitude around the central baseline of trajectory B to generate a dissolved oxygen fluctuation range that allows fluctuation. For example, from 21:20 to 21:35, if the initial dissolved oxygen control trajectory line is 1.60 mg / L, the superimposed range is 1.45 to 1.75 mg / L. From 21:35 to 21:50, if the initial dissolved oxygen control trajectory line rises to 1.75 mg / L, the superimposed range is 1.60 to 1.90 mg / L. This range is the elastic tracking envelope of the equipment control parameters. The envelope here can be a hyperbola formed by the upper and lower boundaries, or it can be an array of upper and lower limits at discrete time nodes. For example, at a 5-minute resolution, the first segment can form the following nodes: 21:20 is [1.45, 1.75], 21:25 is [1.50, 1.80], and 21:30 is [1.55, 1.85]. As long as the execution layer ensures that the actual control quantity falls within this interval, it is considered to meet the current control requirements. As an anomaly handling mechanism or as an alternative, if the prediction results show that the effluent boundary will be breached regardless of the candidate trajectory used, the system enters an enhanced mode, temporarily narrowing the envelope and raising the entire range, while simultaneously issuing a process risk warning to the host computer; if the prediction shows that the load will be stable for a long period of time and the safety margin is greater than the preset safety tolerance threshold, the system can appropriately widen the envelope width to improve the tolerance of load reduction adjustment under smooth operating conditions; if at a certain moment the effluent prediction value cannot be reliably estimated due to abnormal sensor data, the envelope of the previous cycle will be used and the adjustment slope will be reduced in this control cycle to avoid the controller making large adjustments based on unreliable data; For example, before the nighttime surge at the wastewater treatment plant, the system, based on the arrival time of the high load at 21:30 and a 12-minute metabolic response delay, starts increasing the upstream oxygen supply from 21:18 in advance, and calculates that the upstream dissolved oxygen can safely operate at 1.45 to 1.85 mg / L for the next 30 minutes without being forcibly increased to a fixed 2.0 mg / L; after the high load passes, the system gradually lowers the envelope based on the subsequent load decline trend, releasing the downward adjustment power buffer threshold for the equipment; The purpose of this step is to transform the complex uncertainties of future operating conditions into dynamic operating ranges that are easy for the execution layer to access, thereby achieving safe tracking control based on the water outlet boundary as a constraint and the equipment input balance as a guide.

[0021] In this embodiment, the model predictive control algorithm extrapolates the control trajectory over a future time period in a preset virtual environment, including: Identify the load change trend in the future load impact curve and extract the load change rate at adjacent time points; Combine the biological response inertia state with the corresponding control strategy; If the load change rate is less than zero and the absolute value is greater than the preset fluctuation threshold, the load change trend is determined to be decreasing. When the load change trend is decreasing and the biological reaction inertia state is suitable, a sliding trajectory that allows dissolved oxygen to decrease is generated. If the load change rate is greater than zero and the absolute value is greater than the preset fluctuation threshold, the load change trend is determined to be upward. When the load change trend is upward or the biological reaction inertia state is overloaded, a compensation trajectory to maintain or increase dissolved oxygen is generated. If the above conditions are not met, the current control trajectory will be maintained.

[0022] This embodiment provides a control trajectory matching mechanism. Specifically, although the previous embodiment can generate an elastic tracking envelope, if all time periods are solved by uniform optimization, it is easy to increase the computational burden when the working conditions do not change significantly, and it is not conducive to reflecting process experience under different trends. Therefore, this embodiment further introduces a dual-condition judgment of load change trend + biological reaction inertia state for rapid matching of sliding trajectory, compensation trajectory or maintenance trajectory. Specifically, the system identifies the load change trend in the future load impact curve. The rate of change can be represented by the difference in load intensity between adjacent time points. For simplicity, if the load intensities of the four future time points are 1.00, 0.92, 0.80, and 0.70 respectively, the rates of change are -0.08, -0.12, and -0.10. If the preset fluctuation threshold is 0.05, the absolute values ​​of these rates of change are all greater than the threshold and are negative, thus indicating a downward trend. Conversely, if the future time points are 0.75, 0.88, 0.98, and 1.06, the rates of change are positive and their absolute values ​​exceed the threshold, indicating an upward trend. After identifying the load change trend, the system combines the biological reaction inertia state matching control strategy; if the trend is downward and the microorganisms are in a suitable state, a gliding trajectory that allows dissolved oxygen to decrease is generated; this gliding is not a sudden drop, but a smooth descent constrained by the envelope; for example, if the current dissolved oxygen is 1.8 mg / L, the system can provide trajectory nodes for gradually decreasing to 1.5 mg / L over the next 15 minutes, instead of dropping to 1.2 mg / L in one step; this allows the existing activity of the microorganisms to treat the residual load, while reducing ineffective aeration; If the trend is upward, or if the trend is not obvious but the biological reaction inertia is already in an overload state, a compensation trajectory to maintain or increase dissolved oxygen will be generated. For example, if the current dissolved oxygen at the front end is 1.4 mg / L, and the load intensity increases from 0.8 to 1.1 in the next 20 minutes, and the oxygen uptake rate at the front end is already in the overload range, the system will generate a compensation trajectory of 1.4→1.6→1.75 mg / L. The purpose is to provide sufficient oxygen transfer conditions for microorganisms before the load truly arrives in full force. If neither a significant decrease nor a significant increase is met, and the biological state has not reached the level requiring compensation, then the current control trajectory is maintained; for example, if the future load intensity is 0.82, 0.84, 0.83, 0.81, and the rate of change is less than 0.05, then it is considered a stable operating condition, and the controller does not replan the large-scale trajectory, but only maintains the existing envelope centerline. The reason for this trend matching is that relying solely on the results of a single optimization may overlook the differences in operating conditions, such as when energy consumption should be actively released when the load is continuously decreasing or when the microorganisms are already on the verge of overload even though the load has not yet increased significantly. This implementation method is equivalent to superimposing a strategy constraint that is more in line with the mechanism of biochemical processes on top of the optimization layer. As an anomaly handling mechanism or as an alternative, if the load change rate fluctuates frequently around the positive and negative thresholds, the weighted average change rate of the three most recent consecutive nodes can be used as the final judgment to prevent frequent switching of trajectory types; if the load trend is decreasing but the biological state is in a state of starvation, the gliding trajectory is not directly applied, but the current trajectory is maintained or only slightly reduced to avoid slow system recovery due to excessive oxygen reduction; if the load trend is increasing but the equipment capacity is close to the limit, the compensation trajectory prioritizes oxygen supply to critical areas and restricts non-critical areas to operate near the lower boundary of the envelope. For example, after 22:10 in this wastewater treatment plant, the intensity of the upstream load mass decreases from 0.95 and 0.83 to 0.72, and the microorganisms are in a suitable state. The system then generates a sliding trajectory for the upstream that allows for a slow decrease in dissolved oxygen. At the same time, the future load intensity in the midstream increases from 0.80 to 1.02, and the midstream is in an overload state. Therefore, the midstream is assigned to the compensation trajectory. In this way, the entire tank will not miss the pre-compensation for the midstream shock due to a global reduction in aeration, nor will it cause over-aeration in the upstream due to a global increase in aeration. The purpose of this step is to use clearer trend rules to strategically constrain the predictive control results, thereby achieving trajectory extrapolation that is more consistent with the dynamic characteristics of the wastewater biochemical process. In this embodiment, the flexible tracking envelope of device control parameters is converted into device execution instructions. Based on the device execution instructions and a preset optimal efficiency range for the device, the operating frequency path is determined within the flexible tracking envelope of the device control parameters, including: Extract boundary parameters from the elastic tracking envelope of the equipment control parameters; Obtain the current operating frequency of the execution device and the pre-stored energy consumption curve, and match the boundary parameters with the current operating frequency and energy consumption curve; Within the range defined by the elastic tracking envelope of the equipment control parameters, search for the frequency node that has the highest degree of overlap with the equipment's optimal efficiency range; Connect frequency nodes to generate operating frequency paths.

[0023] This embodiment provides a mechanism for determining the operating frequency path. Specifically, the aforementioned scheme has provided an elastic tracking envelope for the equipment control parameters, but the envelope is essentially still an allowable range of the process layer and cannot directly drive the variable frequency blower. If the efficiency characteristics of the equipment itself are ignored and execution is only required if the parameters fall within the envelope, it may result in a situation where the water quality is safe, but the motor operates in a low-efficiency frequency band for a long time. Therefore, this embodiment further searches for the most suitable frequency path for efficient operation of the equipment within the envelope. Specifically, the system first extracts the boundary parameters from the elastic tracking envelope; for example, within a future 30-minute window, the lower and upper boundaries of the frequency corresponding to the oxygen supply required in the front end are [37,41]Hz, [38,43]Hz, [39,44]Hz, and [38,42]Hz, respectively; these boundary parameters can be obtained by converting the dissolved oxygen range into the oxygen supply range from the process model, and then further converting it into the blower frequency range; The system acquires the current operating frequency and pre-stored energy consumption curves. The energy consumption curves and the optimal efficiency range of the equipment are determined by performance bench test data or on-site power consumption statistics provided by the equipment manufacturer. For example, the blower is currently operating at 36Hz. Its unit oxygen supply energy consumption curve shows that 38Hz to 42Hz is the optimal efficiency range, efficiency gradually decreases above 43Hz, and it is not suitable for long-term operation below 35Hz due to surge risk and insufficient air volume. The system matches the boundary parameters at each of the aforementioned moments with the current frequency and energy consumption curves, and eliminates frequency points that are within the envelope but have poor efficiency or are not suitable for stable operation. Within the envelope constraint, the system searches for the frequency node with the highest overlap with the optimal efficiency interval. To ensure the final uniqueness of multiple candidate nodes and eliminate uncertain allocations in the algorithm, the system introduces a frequency comprehensive cost evaluation formula for quantitative assessment. Among them, the cost assessment value The candidate frequencies for the current traversal The value and the peak frequency of energy efficiency in the optimal efficiency range The absolute value of the difference between the values, and the alternative frequency. The numerical value and the frequency node of the previous adjacent time in the path The absolute value of the difference between the values ​​is the same as the smoothing constraint weight coefficient. The product of the two, and the sum of the two; For ease of explanation, if the regulations are as follows It is 41Hz. The value is 0.6, and the initial production scheduling setting is 36Hz. When searching for the next four control nodes, for the first node, the allowable range overlapping with the optimal efficiency is 38Hz to 41Hz. Substituting these values ​​into the evaluation formula, if the alternative is 39Hz, then... If 40Hz is selected as an alternative, then... It should be noted that in calculating the above cost assessment value... At this point, the dimensionless pure numerical value is obtained by extracting frequency values ​​and calculating them, which is used for the optimization comparison of the evaluation function; at this point, the node with relatively smoother and more efficient can be selected based on the minimum cost evaluation value; as the time series progresses, the result of the previous node is updated. Similarly, the allowable overlap ranges for the second node (e.g., 38Hz to 42Hz), the third node (e.g., 39Hz to 44Hz), and the fourth node (e.g., 38Hz to 42Hz) are solved one by one. Thus, by comparing the cost evaluation values, the fuzzy decision that any one of them can be chosen can be eliminated, and after quantization and connection, a smooth and efficient operating frequency path such as 39Hz, 40Hz, 41Hz, and 40Hz is finally formed. These frequency nodes are connected to form an operating frequency path; this path not only meets the process envelope requirements, but also tries to keep the equipment operating in the high-efficiency range; if multiple blowers are connected in parallel at the same time, the system can also distribute the total demand to multiple devices, giving priority to one of them to operate in the optimal efficiency band, while the other undertakes the fine-tuning task, thereby further reducing the total operating power of the system. Compared with directly calculating the single-point frequency based on the control quantity, this path-based generation method can avoid the frequency jumping irregularly between the upper and lower boundaries of the envelope. For example, if a node allows both 39Hz and 43Hz, both of which meet the oxygen supply requirements, but if 43Hz exceeds the high-efficiency zone while 39Hz can still meet the requirements, then 39Hz will be selected first to reduce the power output requirements of the equipment. As an anomaly handling mechanism or as an alternative, if the envelope range does not overlap with the optimal efficiency range at all, for example, if extreme operating conditions necessitate operation above 45Hz, the system will select the feasible frequency within the envelope that is closest to the upper boundary of the optimal efficiency, based on the principle of prioritizing process safety and then considering efficiency. If the span between the current frequency and the target frequency exceeds the preset frequency adjustment range, for example, if it is necessary to rapidly increase from 35Hz to 44Hz, the system can add intermediate transition nodes to avoid mechanical shock. If the equipment energy consumption curve is missing or maintenance and replacement have occurred, the system can temporarily use the equipment nameplate recommended range and historical statistical curves as substitutes.

[0024] For example, during the pre-compensation phase before the high load arrives at the wastewater treatment plant from 21:20 to 21:35, the allowable frequency ranges for the blowers calculated from the process envelope are 37Hz to 41Hz, 38Hz to 43Hz, and 39Hz to 44Hz, respectively. The system query finds that the optimal efficiency range for the blowers is 38Hz to 42Hz. Therefore, the cost evaluation function is used to converge 39Hz, 40Hz, and 41Hz as continuous nodes, instead of directly jumping to 44Hz or fluctuating randomly. This satisfies the oxygen supply requirements of the upstream section while avoiding unnecessary high-frequency or oscillating operation.

[0025] The purpose of this step is to further map the safety envelope of the process layer into an efficient operating path of the equipment layer, thereby achieving a balance between meeting processing requirements and ensuring equipment energy efficiency.

[0026] In this embodiment, the execution device for adjusting the operating frequency path includes: Convert the operating frequency path into an electrical signal; The electrical signal is sent to the programmable logic controller that is communicatively connected to the execution device; The operating frequency of the execution device can be smoothly adjusted by a programmable logic controller.

[0027] This embodiment provides a frequency path execution mechanism. Specifically, the previous embodiment has determined a frequency path suitable for efficient equipment operation. However, if discrete frequency nodes are directly hard-switched to the equipment during implementation, it may still cause fan surge, sudden changes in pipeline pressure, or motor overcurrent. Therefore, this embodiment further explains how to convert the operating frequency path into an actually executable electrical signal and smoothly adjust it by a programmable logic controller. Specifically, the system first converts the operating frequency path into an electrical signal; this electrical signal can be an analog control signal or a digital control frame encapsulated by an industrial communication protocol; for ease of explanation, if the frequency path nodes are 39Hz, 40Hz, 41Hz, and 40Hz, and the control cycle is one node every 5 minutes, then the system can further insert several transition points between adjacent nodes. For example, when transitioning from 39Hz to 40Hz, 39.2Hz, 39.4Hz, 39.6Hz, 39.8Hz, and 40Hz are generated in 30-second steps, corresponding to a set of continuous control quantities output; Afterwards, the system sends these control quantities to the programmable logic controller (PLC) that is connected to the blower. After receiving the path, the PLC can first perform local safety verification, such as confirming that the target frequency does not exceed the inverter's allowable upper limit and is not lower than the minimum stable operating frequency, and checking the current motor status, fault relay status and damper interlock status. If the verification passes, the PLC sends instructions to the inverter according to the predetermined gradient. During the smooth adjustment process, the programmable logic controller (PLC) does not directly execute node jumps, but instead adjusts the operating frequency of the device smoothly according to a preset slope limit. For example, if the maximum allowable frequency ramp slope is 0.05 Hz / s, it will take at least 20 seconds to ramp from 39 Hz to 40 Hz. If the control system wants to ramp from 40 Hz to 43 Hz within 1 minute, but the slope limit only allows 2 Hz / min, the PLC will adjust the target to 42 Hz and continue ramping in the next control cycle. This smoothing mechanism can reduce mechanical shock and airflow oscillations. If there is more than one execution device in the system, such as two blowers supplying air in parallel at the front end, the programmable logic controller can also decompose the total frequency path into a main machine path and an auxiliary machine path; for example, the main machine is kept running in the high-efficiency range of around 40Hz, and the auxiliary machine is slightly compensated between 36Hz and 38Hz, so as to balance air volume stability and energy consumption control. As an anomaly handling mechanism or as an alternative, if the communication between the programmable logic controller (PLC) and the upper control system is interrupted, the PLC can enter a local backup mode and continue to execute the end steady-state value of the most recent complete path, or return to the preset safe frequency. If the actual frequency fed back by the frequency converter deviates too much from the target frequency, for example, the target is 41Hz but the actual frequency stays at 38Hz for a long time, it can be determined that the equipment response is abnormal, the system will suspend the continued increase and issue a maintenance alarm. If the pipeline pressure exceeds the safety threshold during the smooth adjustment process, the pressure protection will be triggered first, and the frequency increase action will be temporarily frozen. For example, when the wastewater treatment plant started pre-compensation at 21:18, the upper control system issued the blower path for the next 15 minutes: 39Hz→40Hz→41Hz. The programmable logic controller (PLC) refined this path into several control points at the 30-second level, and after confirming that the blower bearing temperature and the air valve opening were normal, it gradually increased the frequency according to the limit slope. In this way, the oxygen supply was steadily increased before the high load, rather than causing fluctuations in the blower system by a one-time sudden increase in frequency. The purpose of this step is to transform the frequency path generated by the optimization layer into a control process that the device can execute stably, thereby enabling the reliable implementation of control commands from the algorithm layer to the field hardware layer.

[0028] In this embodiment, collecting adjusted effluent water quality data and feeding the effluent water quality data back to the model predictive control algorithm for parameter correction includes: The absolute difference between the effluent water quality data and the preset effluent water quality tolerance boundary is calculated as the deviation value;

[0029] The model predictive control algorithm has a built-in objective function that includes target penalty weights. If the deviation value is greater than the preset deviation threshold, the target penalty weights of the model predictive control algorithm will be updated. If the deviation value is not greater than the preset deviation threshold, the current penalty weight of the model predictive control algorithm is maintained.

[0030] This embodiment provides a feedback correction mechanism. Specifically, the aforementioned scheme can achieve feedforward control based on load tracking and biological state prediction. However, the wastewater treatment process still has long-term drift factors such as seasonal changes, sludge age changes, and mass transfer efficiency decay. If the model parameters are fixed for a long time, the prediction error will gradually accumulate. Therefore, this embodiment uses the adjusted effluent water quality data to dynamically correct the model prediction control algorithm. Specifically, after executing the frequency path, the system collects effluent water quality data and calculates the absolute difference between this data and the preset effluent water quality tolerance boundary as the deviation value. For ease of explanation, if the upper limit of the effluent ammonia nitrogen tolerance boundary is 5.0 mg / L, and the current detection value is 4.2 mg / L, then the deviation value is... If the detected value is 5.4 mg / L, then the deviation value is... Although the absolute difference between the two is different, their technological significance is not exactly the same. Therefore, in practical applications, a status flag indicating whether the boundary has been crossed can be added to participate in the judgment. If the deviation value exceeds the preset deviation threshold, the weight update mechanism of the model predictive control algorithm is triggered. For specific data calculation and deduction, the deviation threshold can be set to 0.3 mg / L. If the measured ammonia nitrogen in the effluent is 5.4 mg / L, the deviation value of 0.4 mg / L exceeds the threshold, and the system triggers an update. The weight update mechanism here is based on the execution of quantitative closed-loop feedback logic: the system calculates the excess difference of the deviation value exceeding the deviation threshold, and uses this difference as a driving factor to positively map the incremental adjustment step size of the penalty weight. For example, for every 0.1 mg / L of deviation, the target tracking risk weight of the predictive control model is increased by a preset quantization unit. The preset quantization unit is a dimensionless fixed step size or proportional coefficient pre-configured according to the sensitivity of the control system. The system uses time-series data queues to backtrack and locate the upstream water mass batch corresponding to the out-of-bounds deviation, and extracts its operating condition attribute classification when entering the reaction zone. The incremental step size calculated above is added to the objective function of the model to correspond to the penalty term of the effluent boundary constraint for the corresponding attribute condition. With this mechanism, updates can be reflected at multiple levels and implemented precisely. For example, the weight of the effluent boundary constraint in the objective function can be increased, the risk penalty for future high-load nodes can be increased, the allowable decline in dissolved oxygen can be shortened, or the hydraulic delay parameter and oxygen transfer efficiency parameter can be corrected. For example, if the system finds that the last three predictions underestimated the peak effluent ammonia nitrogen, the risk penalty correlation parameter for the high-load response node can be increased from 1.0 to 1.3, making the next round of solution more inclined to increase the compensation frequency in advance. If the deviation value is not greater than the preset deviation threshold, the current weights of the model predictive control algorithm are maintained. This avoids frequent modifications to the model when the system is running stably and maintains control continuity. For example, if the effluent ammonia nitrogen is stable between 4.7 mg / L and 4.9 mg / L for several consecutive cycles, although it is close to the boundary, it is still within the acceptable prediction error range. The system will not adjust the weights and will continue to operate according to the current envelope strategy. Compared to simply correcting based on instantaneous deviation, this implementation can also incorporate the principle of continuity; for example, a significant weight update is only performed when the deviation exceeds the threshold twice consecutively, or exceeds the threshold once and simultaneously crosses the boundary; otherwise, only a slight correction is performed to prevent model oscillation caused by a single accidental measurement error. As an anomaly handling mechanism or as an alternative, if the effluent detection device has a long analysis cycle, such as an online ammonia nitrogen meter updating every 15 minutes while the control cycle is 5 minutes, the most recent valid effluent data can be combined with the soft measurement results at intermediate moments. If a effluent sensor malfunction causes a data jump, such as from 4.5 mg / L to 9.8 mg / L instantaneously, and manual verification and other indicators do not support this result, the weight update will not be triggered temporarily and will be processed after secondary confirmation. If the deviation remains small for a long time but the average load increment of the supporting execution equipment shows a divergent trend, it indicates that the constraint judgment setting tends to the conservative limit. In this case, it is allowed to downgrade and adjust the weight parameters of some risk-related items, thereby widening the operational margin for frequency reduction control to explore the lower limit. For example, during three consecutive nightly operation cycles at the wastewater treatment plant, the system predicted that the peak ammonia nitrogen level in the effluent after the high load should be 4.7 mg / L, but the actual measured values ​​were 5.2 mg / L, 5.3 mg / L, and 5.1 mg / L, respectively, all exceeding the deviation threshold of 0.3 mg / L. Therefore, the controller automatically increased the penalty weight associated with the high load node and corrected the upstream metabolic response delay to a more conservative value. After the adjustment, the peak value for the next night dropped to 4.8 mg / L, and the model weights remained stable and no longer increased. The purpose of this step is to perform closed-loop correction of the feedforward prediction and execution strategy using the actual results from the effluent side, thereby achieving adaptive updates of model parameters as operating conditions change and establishing the reliability of the control benchmark over a long period of time.

[0031] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. A method for tracking and controlling parameters of wastewater treatment equipment, characterized in that, include: The system acquires influent flow rate data, influent water quality data, and preset effluent water quality tolerance boundaries. The system includes a reaction zone and execution equipment configured in the reaction zone, including aeration equipment. Acquire real-time dissolved oxygen data, redox potential data, and sludge concentration data in the reaction zone; Based on influent flow rate data, influent water quality data, and spatial parameters of the reaction zone, a virtual water mass tracking model is constructed, and a future load impact curve with timestamps and spatial coordinates is generated through the virtual water mass tracking model. Based on real-time dissolved oxygen data, redox potential data, and sludge concentration data, the oxygen uptake rate is calculated, the biological reaction inertia state is assessed based on the oxygen uptake rate, and the actual oxygen demand and metabolic response delay time are output. Align the arrival time of the future load impact curve with the metabolic response delay time on the time axis, combine the actual oxygen demand, and use the model predictive control algorithm to generate an elastic tracking envelope of the equipment control parameters that changes over time. The flexible tracking envelope of the equipment control parameters is converted into equipment execution instructions. Based on the equipment execution instructions and the preset optimal efficiency range of the equipment, the operating frequency path is determined within the flexible tracking envelope of the equipment control parameters. The operating frequency path is used to control the oxygen supply of the aeration equipment to match the actual oxygen demand. Based on the operating frequency path adjustment execution equipment, and collecting the adjusted effluent water quality data, the effluent water quality data is fed back to the model predictive control algorithm for parameter correction, forming a parameter tracking control closed loop.

2. The wastewater treatment equipment parameter tracking and control method as described in claim 1, characterized in that, Based on influent flow rate data, influent water quality data, and spatial parameters of the reaction zone, a virtual water mass tracking model is constructed. This model generates future load impact curves with timestamps and spatial coordinates, including: Extract the instantaneous change rate and peak characteristics of influent flow rate data and influent water quality data within a preset time window; The instantaneous rate of change and peak characteristics are mapped to a preset spatiotemporal three-dimensional grid to construct a virtual water mass tracking model; The flow trajectory of water masses in physical space is calculated using a virtual water mass tracking model; Based on the flow trajectory, the arrival time and load intensity of the water mass to the preset target area in the reaction zone are predicted, and a future load impact curve with timestamps and spatial coordinates is generated.

3. The wastewater treatment equipment parameter tracking and control method as described in claim 2, characterized in that, Based on real-time dissolved oxygen data, redox potential data, and sludge concentration data, oxygen uptake rate is calculated. The biological reaction inertia state is assessed based on the oxygen uptake rate, and the actual oxygen demand and metabolic response delay time are output, including: Input real-time dissolved oxygen data, oxidation-reduction potential data, and sludge concentration data into the preset activated sludge model; The oxygen uptake rate at the current moment is calculated by solving the kinetic equations in the activated sludge model. The oxygen uptake rate is compared with a preset activity threshold range to determine the biological reaction inertia state. The actual oxygen demand and metabolic response delay time corresponding to the biological reaction inertia state are queried and extracted from the preset state mapping table.

4. The wastewater treatment equipment parameter tracking and control method as described in claim 3, characterized in that, The oxygen uptake rate is compared with a preset activity threshold range to determine the biological reaction inertia state, including: If the oxygen uptake rate is less than the preset lower limit activity threshold, the biological reaction inertia state is determined to be a starvation state. If the oxygen uptake rate is not less than the lower limit activity threshold and not greater than the preset upper limit activity threshold, then the biological reaction inertia state is determined to be a suitable state. If the oxygen uptake rate is greater than the upper limit of the activity threshold, the biological reaction inertia state is determined to be an overload state.

5. The wastewater treatment equipment parameter tracking and control method as described in claim 4, characterized in that, Aligning the arrival time of the future load shock curve with the metabolic response delay time on the time axis, and combining this with the actual oxygen demand, a model predictive control algorithm is used to generate a time-varying elastic tracking envelope for the equipment control parameters, including: Input the future load impact curve, actual oxygen demand, and metabolic response delay time into the model predictive control algorithm; The model predictive control algorithm is used to extrapolate the control trajectory within a future time period in a preset virtual environment; The predicted water quality value of the control trajectory is subtracted from the preset effluent water quality tolerance boundary to obtain the deviation of the effluent index; the pre-built water quality and dissolved oxygen sensitivity mapping table is called to extract the dissolved oxygen fluctuation amplitude corresponding to the deviation of the effluent index, and the safe elastic range of allowable dissolved oxygen fluctuation is calculated based on the fluctuation amplitude. Based on the safety elasticity range, an elastic tracking envelope of device control parameters that varies over time is generated.

6. The wastewater treatment equipment parameter tracking and control method as described in claim 5, characterized in that, The model predictive control algorithm extrapolates the control trajectory over a future time period in a preset virtual environment, including: Identify the load change trend in the future load impact curve and extract the load change rate at adjacent time points; Combine the biological response inertia state with the corresponding control strategy; If the load change rate is less than zero and the absolute value is greater than the preset fluctuation threshold, the load change trend is determined to be decreasing. When the load change trend is decreasing and the biological reaction inertia state is suitable, a sliding trajectory that allows dissolved oxygen to decrease is generated. If the load change rate is greater than zero and the absolute value is greater than the preset fluctuation threshold, the load change trend is determined to be upward. When the load change trend is upward or the biological reaction inertia state is overloaded, a compensation trajectory to maintain or increase dissolved oxygen is generated. If the above conditions are not met, the current control trajectory will be maintained.

7. The wastewater treatment equipment parameter tracking and control method as described in claim 6, characterized in that, The flexible tracking envelope of equipment control parameters is converted into equipment execution commands. Based on the equipment execution commands and the preset optimal efficiency range of the equipment, the operating frequency path is determined within the flexible tracking envelope of the equipment control parameters, including: Extract boundary parameters from the elastic tracking envelope of the equipment control parameters; Obtain the current operating frequency of the execution device and the pre-stored energy consumption curve, and match the boundary parameters with the current operating frequency and energy consumption curve; Within the range defined by the elastic tracking envelope of the equipment control parameters, search for the frequency node that has the highest degree of overlap with the equipment's optimal efficiency range; Connect frequency nodes to generate operating frequency paths.

8. The wastewater treatment equipment parameter tracking and control method as described in claim 7, characterized in that, The execution devices based on operating frequency path adjustment include: Convert the operating frequency path into an electrical signal; The electrical signal is sent to the programmable logic controller that is communicatively connected to the execution device; The operating frequency of the execution device can be smoothly adjusted by a programmable logic controller.

9. The wastewater treatment equipment parameter tracking and control method as described in claim 8, characterized in that, Collect effluent water quality data after adjustment, and feed this data back to the model predictive control algorithm for parameter correction, including: The absolute difference between the effluent water quality data and the preset effluent water quality tolerance boundary is calculated as the deviation value; The model predictive control algorithm has a built-in objective function that includes target penalty weights. If the deviation value is greater than the preset deviation threshold, the target penalty weights of the model predictive control algorithm will be updated. If the deviation value is not greater than the preset deviation threshold, the current penalty weight of the model predictive control algorithm is maintained.

Images