Wastewater treatment plant operation assistance system based on inflow water quantity dynamics and water quality real-time monitoring

Abstract

The present application relates to sewage treatment automatic control technical field, disclose a sewage plant operation auxiliary system based on water inflow dynamic and water quality real-time monitoring, including perception execution layer and control layer. The control layer divides the biochemical reactor volume into fluid microelement to establish volume displacement matrix, and the position of field device is mapped as matrix constant index. The cumulative volume is obtained by integrating the inflow, and the matrix node is triggered to delay and assign the inflow characteristic data when the volume reaches the set value. According to the backflow and air supply flow, the data of the node is re-annotated and smoothed to correct the flow state. When the data node is displaced to the corresponding index, the material demand is calculated combined with the stoichiometry constant matrix, and the feedforward control instruction is generated. The space-time difference between the predicted concentration and the actual concentration of the effluent is calculated, and the constant matrix is dynamically updated. The present application replaces the fixed time step with volume integration driving, adapts to the inflow fluctuation, eliminates the space-time lag of dosing and aeration instruction, and improves the feedforward control precision and system stability.

Description

Technical Field

[0001] This invention relates to the field of automatic control technology for wastewater treatment, specifically to an auxiliary system for wastewater treatment plant operation based on dynamic monitoring of influent flow and real-time water quality. Background Technology

[0002] In the daily operation of wastewater treatment plants, the stable operation of the biological reactor depends on the precise control of aeration rate and chemical dosage. Existing wastewater treatment plant operation control systems typically use a fixed time step combined with an influent proportion model to estimate the migration process of pollutants within the biological reactor. However, the actual influent flow rate of a wastewater treatment plant fluctuates continuously. With a fixed time step, because the impact of flow rate changes on hydraulic retention time is not considered, the calculation model cannot accurately reflect the physical displacement process of water within the reactor, resulting in the control system's inability to accurately track the spatial distribution and dynamic changes in concentration of pollutants within the biological reactor.

[0003] Meanwhile, the hydraulic flow pattern inside the bioreactor is not an ideal plug flow state; most existing control models simplify the internal hydraulic conditions, neglecting the material extraction and re-injection effects caused by sludge recirculation in different reaction sections, and failing to effectively quantify the back-mixing disturbance effect of aeration equipment on local water bodies during operation. The lack of real-time correction for these physical mixing phenomena causes the theoretical pollutant concentration calculated within the system to gradually deviate from the actual water quality conditions.

[0004] Because the system cannot accurately grasp the spatial migration patterns of pollutants and lacks flow regime correction, the generated dosing and aeration feedforward control commands are out of sync with the actual biochemical reaction material requirements, exhibiting significant spatial and temporal lags. Furthermore, when using effluent monitoring data for feedback regulation, existing technologies often directly compare the current effluent measurement value with the theoretical calculation value. This approach does not consider the response delay of online water quality analyzers and the physical transmission time of the biological treatment tank itself, resulting in the data used for feedback comparison not being strictly aligned in time and space. Due to the bias in the comparison benchmark, the control system struggles to calculate accurate correction parameters to update the underlying chemical kinetic model constants, leading to low feedback regulation accuracy and an inability to adapt to long-term, slow changes in microbial population activity, ultimately affecting the stability of the effluent quality. Summary of the Invention

[0005] The technical problem solved by this invention is that existing auxiliary operation schemes for wastewater treatment plants are difficult to accurately reflect the actual migration location of pollutants in the biochemical reactor when there are fluctuations in influent flow rate, water quality, return flow status, and aeration intensity. This can easily cause deviations in the timing of chemical dosing and aeration adjustment, affecting operational stability and control accuracy.

[0006] To address the above problems, the present invention provides the following technical solution:

[0007] The first aspect of this invention provides a wastewater treatment plant operation support system based on dynamic monitoring of influent flow and real-time water quality. This system includes a sensing and execution layer and a control layer. The sensing and execution layer is located on-site and includes monitoring and execution units for collecting water quality and quantity data and executing actions. The control layer is communicatively connected to the sensing and execution layer and is used to receive on-site monitoring data and execution feedback data, and to perform operational auxiliary calculations for carbon source addition and aeration adjustment.

[0008] The control layer includes a data acquisition module, a matrix construction module, a volume-driven module, a flow regime correction module, a feedforward execution module, and a feedback regulation module. The data acquisition module receives instantaneous influent flow rate, influent characteristic data, instantaneous return flow rate, actual air supply flow rate, and actual effluent concentration data, and performs engineering quantity conversion, timestamp unification, and baseline value extraction on the collected data. The matrix construction module divides the effective physical volume of the bioreactor into a set number of equal-volume fluid micro-elements, establishes a volume displacement matrix, and maps the physical spatial positions of the field actuators and return pipelines to constant indices of the volume displacement matrix. The volume-driven module integrates the instantaneous influent flow rate to obtain the cumulative influent volume. When the cumulative influent volume reaches the set volume value of a single equal-volume fluid micro-element, a baseline shift command is triggered, causing the nodes within the volume displacement matrix to extend sequentially, and influent characteristic data is assigned to the nodes released at the head end. The flow correction module performs data re-injection between constant indices in the corresponding return pipeline based on the instantaneous return flow rate, calculates the dynamic backmixing coefficient based on the actual air supply flow rate, and performs lateral smoothing calculations on continuous nodes in the corresponding aerobic zone. The feedforward execution module calculates material requirements using the stoichiometric constant matrix when a data node shifts to its corresponding constant index, and generates feedforward control commands to be sent to the execution unit. The feedback adjustment module spatiotemporally aligns the actual effluent concentration data with the predicted concentration at the end node of the volume displacement matrix and generates a residual sequence to output correction parameters to update the stoichiometric constant matrix.

[0009] The first aspect of this invention uses a volumetric displacement matrix as a unified data carrier and the cumulative influent volume as the driving force for spatial propulsion, transforming the migration process of pollutants within the bioreactor from a fixed-time estimation to a dynamic tracking based on actual volume propulsion. The matrix construction module establishes a logical computation domain through spatial discretization, establishing a correspondence between on-site physical locations and digital indices. The volume-driven module updates the matrix using the actual influent volume, ensuring that pollutant location judgments are consistent with dynamic changes in the influent. The flow regime correction module introduces recirculation and backmixing smoothing to correct the effects of recirculation and aeration disturbances, enabling the logical model to reflect the actual process flow regime. The feedforward execution module performs material demand calculations based on the characteristic concentration of pollutants at specific index locations, ensuring that control actions correspond to the spatial arrival location of pollutants. The feedback adjustment module does not directly compensate for the underlying execution quantity but instead uses the effluent residual to correct the stoichiometric constant matrix, maintaining the material balance within the feedforward calculation framework and improving parameter adaptability during long-term operation.

[0010] In one implementation, the volume drive module determines the number of shifts to be executed in the current cycle based on the integer part of the ratio of the current cumulative influent volume to the set volume value, and continuously executes the corresponding number of reference shift instructions. The data of each node in the volume displacement matrix is ​​then sequentially extended by the corresponding step size in the index-incrementing direction. Simultaneously, the average influent characteristic data within the corresponding integration interval is sequentially assigned to the nodes released at the head end, and the margin of the cumulative influent volume relative to the set volume value is retained as the initial integration value for the next control cycle. This process avoids the volume omission problem that occurs when a single cycle spans multiple micro-volumes under high flow conditions.

[0011] In one implementation, the flow correction module can perform time integration on the instantaneous recirculation flow rate to obtain the cumulative recirculation volume, and construct a dynamic recirculation ratio constant based on the ratio of the cumulative recirculation volume to a set volume value. On this basis, the data vector at the recirculation extraction index is extracted, and the data vectors extending to the recirculation access index are weighted and summed to cover them. In this way, the cross-regional migration and remixing process of matter caused by recirculation can be characterized within the volume displacement matrix.

[0012] In one implementation, the flow regime correction module can use a preset nonlinear mapping function to convert the actual air supply flow rate of each air supply branch in the aerobic zone into a dynamic backmixing coefficient. Then, for the target node within the aerobic zone span, it calls an adjacent three-point weighted algorithm to perform lateral smoothing on the basic characteristic concentration vectors of the target node, upstream node, and downstream node. In this way, the local axial mixing effect caused by aeration disturbance can be introduced into the spatial discretization model.

[0013] In one implementation, the feedforward execution module can extract the matrix data vector at the constant index corresponding to the dosing point in the anoxic zone, combine it with the theoretical carbon-nitrogen consumption ratio constant and volume setpoint retrieved from the stoichiometric constant matrix, calculate the absolute carbon source mass, and convert it into the corresponding pulse frequency or execution frequency to be sent to the dosing execution unit. It can also extract the characteristic concentration vector of all continuous fluid micro-elements within the continuous index interval of the aerobic zone, perform interval integration calculation on all nodes within the continuous index interval using the theoretical oxygen consumption equivalent coefficient, obtain the theoretical total oxygen demand, and combine it with the standard oxygen transfer efficiency of the aeration equipment and the current water temperature parameters to convert it into an actual air target flow rate signal to be sent to the aeration execution unit. In this way, spatial fixed-point calculation and interval integration calculation for dosing control and aeration control can be achieved respectively.

[0014] In one implementation, the feedback adjustment module can allocate a data buffer queue in memory to temporarily store the predicted concentration output from the end node of the volume displacement matrix with a delay time determined based on the sampling and measurement cycle of the on-site online analyzer. Then, it generates a residual sequence by subtracting the real-time acquired actual effluent concentration data from the delayed theoretical predicted concentration. When the actual effluent concentration data exceeds a set range, the residual value for the current cycle is forced to zero. Further, the feedback adjustment module can input the residual sequence into a discrete proportional-integral-derivative (DI-D) controller, combine it with anti-integral saturation logic to calculate the dynamic compensation amount of the model constants, and superimpose this compensation amount onto the basic stoichiometric constants called in the previous calculation cycle. After physical threshold clamping and limiting, the current execution constant is generated for the feedforward execution module to call. In this way, closed-loop correction of model parameters can be implemented without directly interfering with the output of the underlying actuator.

[0015] A second aspect of this invention provides a wastewater treatment plant operation support method based on dynamic influent flow and real-time water quality monitoring, comprising the following processes:

[0016] The effective physical volume of the biochemical reactor is divided into multiple equal-volume fluid micro-elements, a volume displacement matrix is ​​established, and the mapping between the physical position of the equipment and the constant index is completed.

[0017] The system continuously acquires the instantaneous influent flow rate, influent pollutant concentration vector, instantaneous return flow rate, actual effluent pollutant concentration data, and operational status data fed back by the execution unit at a set period.

[0018] The instantaneous influent flow rate is integrated to obtain the current cumulative influent volume. When the cumulative volume reaches the set volume value, a reference shift event is triggered and the volume displacement matrix is ​​updated.

[0019] Re-injection is performed based on the return flow rate, and back-mixing smoothing is performed based on the actual gas supply flow rate.

[0020] For data nodes that reach a specific operation index and data nodes that are within a continuous operation interval, feedforward calculations are performed to generate dosing and aeration control commands.

[0021] The predicted concentration output from the end node of the volume displacement matrix is ​​spatiotemporally aligned with the actual effluent concentration and the difference is calculated to generate an error sequence. The stoichiometric constant matrix is ​​then dynamically corrected using the error sequence.

[0022] The second aspect of this invention establishes a methodological process corresponding to the first aspect. This method employs a continuous process of spatial discretization, volume-driven updating, reflow re-injection, back-mixing correction, feedforward execution, and feedback adjustment, uniformly mapping on-site water quantity, water quality, and execution status data into the same logical computation framework. Compared to operation assistance methods based on fixed time constants or average residence times, this method maintains consistency between pollutant location identification, material demand calculation, and execution timing under dynamic changes in influent flow. Furthermore, by correcting model parameters through feedback, the system maintains good adaptability during continuous operation.

[0023] This invention provides an auxiliary system for wastewater treatment plant operation based on dynamic monitoring of influent flow and real-time water quality. It offers the following advantages:

[0024] 1. This invention adopts a computational architecture that uses volume integral to drive spatial matrix displacement, and uses the cumulative influent volume as the model stepping trigger condition. This overcomes the shortcomings of traditional fixed time step models that are difficult to adapt to frequent fluctuations in influent flow, and achieves strict matching between matrix flow calculation and actual inflow volume.

[0025] 2. By introducing a dynamic reflux ratio constant and a dynamic backmixing coefficient, this invention performs data re-injection and lateral smoothing operations on the data nodes within the matrix, effectively quantifying and compensating for the physical deviations caused by reflux injection and aeration disturbances on the actual hydraulic flow state of the biochemical reactor, thereby improving the reliability of predicting the concentration distribution of pollutants inside the reactor.

[0026] 3. This invention maps the physical locations of field pipelines and actuators to matrix constant indices, and directly calculates absolute material requirements using feedforward data from displacement to specific spatial nodes, eliminating the spatial and temporal lag of traditional control commands and improving the accuracy of dosing and aeration control.

[0027] 4. This invention establishes a data buffer queue to temporarily store the predicted concentration, thereby eliminating the measurement hysteresis error of online water quality analysis instruments. It also uses the spatiotemporally aligned residual sequence to dynamically update the chemometric constant matrix, enabling the control model to adapt to long-term changes in microbial community activity and ensuring the long-term stable operation of the system. Attached Figure Description

[0028] Figure 1 This is a structural diagram of a wastewater treatment plant operation support system based on dynamic monitoring of influent flow and real-time water quality, according to an embodiment of the present invention.

[0029] Figure 2 This is a flowchart of a wastewater treatment plant operation support method based on dynamic influent flow and real-time water quality monitoring according to an embodiment of the present invention.

[0030] Figure 3 This is a schematic diagram of the hardware layout and data topology connection of the perception execution layer according to an embodiment of the present invention;

[0031] Figure 4 This is a schematic diagram of spatial discretization and matrix mapping of a biochemical reactor according to an embodiment of the present invention;

[0032] Figure 5 This is a schematic diagram of the operation logic and data shifting of a volume driving module according to an embodiment of the present invention;

[0033] Figure 6 This is a schematic diagram illustrating the process of data re-injection performed by the flow correction module according to an embodiment of the present invention;

[0034] Figure 7 This is a schematic diagram of the calculation process for lateral data smoothing performed by the flow correction module according to an embodiment of the present invention;

[0035] Figure 8 This is a schematic diagram of the logic control of a feedforward execution module according to an embodiment of the present invention;

[0036] Figure 9 This is a schematic diagram of the control logic of the feedback adjustment module and the correction of model parameters according to an embodiment of the present invention;

[0037] Figure 10 This is a schematic diagram showing the relationship between the dynamic change of influent flow rate and the number of shifts of the volume driving module according to an embodiment of the present invention;

[0038] Figure 11 This is a schematic diagram showing the distribution of typical pollutant concentrations along the index direction in a volume displacement matrix according to an embodiment of the present invention.

[0039] Figure 12 This is a comparison chart of the changes in ammonia nitrogen and total nitrogen in the effluent of the test group and the control group over time according to an embodiment of the present invention; wherein, (a) is a comparison chart of the changes in ammonia nitrogen in the effluent of the test group and the control group over time; and (b) is a comparison chart of the changes in total nitrogen in the effluent of the test group and the control group over time.

[0040] Figure 13 This is a comparison chart of the operational performance indicators of the experimental group and the control group according to an embodiment of the present invention;

[0041] Figure 14 This is a comparison chart of the fitting relationship between the predicted effluent concentration and the measured effluent concentration according to an embodiment of the present invention; wherein, (a) is the fitting curve of the predicted and measured effluent concentration of the control group, and (b) is the fitting curve of the predicted and measured effluent concentration of the experimental group.

[0042] The system comprises the following components: 100, Wastewater Treatment Plant Operation Auxiliary System; 110, Sensing and Execution Layer; 111, Influent Monitoring Unit; 112, Effluent Monitoring Unit; 113, Backflow Monitoring Unit; 114, Chemical Dosing Unit; 115, Aeration Unit; 120, Control Layer; 121, Data Acquisition Module; 122, Matrix Construction Module; 123, Volume Driving Module; 124, Flow Pattern Correction Module; 125, Feedforward Execution Module; and 126, Feedback Adjustment Module. Detailed Implementation

[0043] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0044] See attached document Figure 1 This invention provides a wastewater treatment plant operation support system 100 based on dynamic monitoring of influent flow and real-time water quality. The wastewater treatment plant operation support system 100 may include a sensing and execution layer 110 and a control layer 120.

[0045] The sensing and execution layer 110 is installed in the biochemical reactor and its supporting piping network to collect physical parameters of the biochemical process and execute control commands issued by the system. The sensing and execution layer 110 includes an influent monitoring unit 111, an effluent monitoring unit 112, a reflux monitoring unit 113, a dosing execution unit 114, and an aeration execution unit 115.

[0046] The influent monitoring unit 111 is located at the influent end of the biochemical reactor and includes an electromagnetic flow meter and an online water quality analyzer, used to collect instantaneous influent flow rate and influent pollutant concentration.

[0047] The effluent monitoring unit 112 is located at the effluent end of the biochemical reactor and includes an ammonia nitrogen sensor, a nitrate nitrogen sensor and / or a total nitrogen analyzer to collect the actual effluent pollutant concentration. As a preferred embodiment, the effluent monitoring unit 112 may also include a dissolved oxygen sensor to assist in monitoring the operating status of the aerobic zone.

[0048] The backflow monitoring unit 113 is installed in the internal backflow pipeline and the external backflow pipeline, and includes a flow meter device for collecting instantaneous backflow flow.

[0049] The dosing execution unit 114 is located in the anoxic zone of the biochemical reactor and includes a variable frequency dosing pump. It is used to add carbon source to the anoxic zone according to control commands and to feed back the current operating frequency or stroke execution status of the dosing pump to the control layer 120.

[0050] The aeration execution unit 115 is located in the aerobic zone of the biochemical reactor and includes a variable frequency blower, an air regulating valve, and an air flow detection device installed on the main air supply pipeline and / or each air supply branch. It is used to adjust the air supply flow of each zone in the aerobic zone according to the control command and to feed back the actual air supply flow, valve opening degree and / or blower operating frequency to the control layer 120.

[0051] The control layer 120 is communicatively connected to the sensing and execution layer 110, receiving sensor data and actuator feedback data collected by the sensing and execution layer 110, and sending device drive signals to the sensing and execution layer 110. The control layer 120 includes a data acquisition module 121, a matrix construction module 122, a volume drive module 123, a flow correction module 124, a feedforward execution module 125, and a feedback adjustment module 126.

[0052] The data acquisition module 121 receives real-time data from the influent monitoring unit 111, the effluent monitoring unit 112, and the return flow monitoring unit 113, as well as the operating status data fed back by the dosing execution unit 114 and the aeration execution unit 115, according to a preset sampling cycle, and sends the data to the other calculation modules.

[0053] The matrix construction module 122 divides the effective physical volume of the biochemical reactor into a set number of equal-volume fluid micro-elements, establishes a volume displacement matrix of corresponding length in the memory space of the control layer 120, and maps the physical spatial positions of the dosing execution unit 114, the aeration execution unit 115 and the return pipeline to the constant index of the volume displacement matrix.

[0054] The volume drive module 123 performs time integration on the instantaneous influent flow rate collected by the influent monitoring unit 111 to obtain the cumulative influent volume. When the cumulative influent volume reaches the volume set value of one or more equal-volume fluid micro-elements, the volume drive module 123 triggers a corresponding number of global reference shift commands.

[0055] Upon receiving the global reference shift command, the flow correction module 124 calculates the mass balance equation based on the instantaneous backflow rate obtained by the backflow monitoring unit 113, and performs a data re-injection operation between the backflow access index and the backflow extraction index corresponding to the volume displacement matrix. Furthermore, the flow correction module 124 calculates the dynamic backmixing coefficient based on the actual air supply flow rate fed back by the aeration execution unit 115, or the actual branch air supply flow rate estimated based on the valve opening, blower frequency, main pipe pressure, and main pipe flow rate, and performs lateral weighted smoothing calculations on specified nodes in the volume displacement matrix.

[0056] The feedforward execution module 125 scans the data distribution of the volume displacement matrix in real time, and performs calculations on data nodes that have reached specific operation indices and data nodes that are within continuous operation intervals according to a preset execution order after each global reference shift event. When a data node with specific pollutant concentration characteristics shifts to the corresponding constant index, the feedforward execution module 125 calculates the absolute value of material demand and generates feedforward control commands, which are then sent to the dosing execution unit 114 and the aeration execution unit 115.

[0057] The feedback adjustment module 126 acquires the actual effluent pollutant concentration data collected by the effluent monitoring unit 112, calculates the difference between the actual effluent and the corresponding predicted concentration stored at the end node of the volume displacement matrix after spatiotemporal alignment, generates a residual sequence, processes the residual sequence using a PID algorithm, and outputs correction parameters to adjust the corresponding chemical reaction constants called in the feedforward execution module 125.

[0058] See attached document Figure 2 This invention provides a wastewater treatment plant operation support method based on dynamic influent flow and real-time water quality monitoring, applied to the aforementioned wastewater treatment plant operation support system 100 based on dynamic influent flow and real-time water quality monitoring. The method may include the following steps:

[0059] In step S110, the matrix construction module 122 divides the effective physical volume of the biochemical reactor into N incompressible, equal-volume fluid micro-elements, and establishes a volume displacement matrix of length N. Based on the spatial topology of the biochemical reactor, the matrix construction module 122 establishes a mapping relationship between the physical location of each device and a specific index number in the volume displacement matrix.

[0060] In step S120, the data acquisition module 121 continuously acquires the instantaneous influent flow rate, influent pollutant concentration vector, instantaneous return flow rate, and actual effluent target pollutant concentration data at a set period, and acquires the operating status data fed back by the dosing execution unit 114 and the aeration execution unit 115; wherein, the data fed back by the aeration execution unit 115 includes at least the actual air supply flow rate, or includes the valve opening degree, blower frequency, and main pipe air supply parameters used to estimate the actual air supply flow rate.

[0061] In step S130, the volume drive module 123 performs an integral calculation on the instantaneous influent flow rate to obtain the current cumulative influent volume. It then determines whether the current cumulative influent volume reaches the preset volume of a single fluid element of equal volume. When this condition is met, the volume drive module 123 calculates the number of shifts n to be performed in the current cycle, where n is the integer part of the ratio of the current cumulative influent volume to the volume of a single fluid element of equal volume. Subsequently, the volume drive module 123 triggers n global reference shift events, extending the data of each node in the volume displacement matrix by n steps in the index-incrementing direction, and calculating the average influent water quality characteristic value within the corresponding integration interval, assigning it sequentially to the n nodes released at the head of the volume displacement matrix. After the shift is completed, the margin of the current cumulative influent volume relative to the volume of a single fluid element of equal volume is retained as the initial integral value for the next control cycle.

[0062] In step S140, the flow correction module 124 acquires the instantaneous return flow rate at the trigger shift and calculates the return ratio for that cycle. Based on the law of conservation of mass, the flow correction module 124 extracts the data corresponding to the return extraction point index in the volume displacement matrix according to the return ratio, and sums and overwrites it with the upstream continuation data of the corresponding return access point index to complete the nonlinear data re-injection.

[0063] In step S150, the flow correction module 124 reads the actual air supply flow rate of each air supply branch in the aerobic zone and converts it into a dynamic backmixing coefficient using a preset mapping function. When some branches are not directly equipped with air flow detection devices, the flow correction module 124 reads the valve opening degree, blower frequency, main pipe air supply pressure and / or main pipe flow rate of the corresponding branch, and calculates the actual air supply flow rate of the branch in combination with the preset valve flow characteristic relationship, and then converts it into a dynamic backmixing coefficient using the preset mapping function. For continuous nodes in the volume displacement matrix corresponding to the aerobic zone, the flow correction module 124 performs a weighted smoothing operation of adjacent three points in combination with the dynamic backmixing coefficient to reconstruct the concentration distribution between nodes.

[0064] In step S160, the feedforward execution module 125 continuously extracts data from the volume displacement matrix that reaches a specific operation index. The feedforward execution module 125 calculates the absolute carbon source mass and theoretical oxygen demand required to eliminate the pollutant at that node by combining the stoichiometric constant matrix, and converts them into control electrical signals that are sent to the dosing execution unit 114 and the aeration execution unit 115.

[0065] In step S170, the feedback adjustment module 126 extracts the predicted water quality data output from the end node of the volume displacement matrix. It then aligns the predicted pollutant concentration components corresponding to the actual monitored items in the effluent monitoring unit 112 with the actual effluent pollutant concentration data in both time and space, and calculates the difference to obtain error sequences. The feedback adjustment module 126 processes these error sequences using one or more PID controllers and uses the controller outputs to fine-tune the stoichiometric constant matrix values ​​used by the feedforward execution module 125 in step S160. This ensures that the feedback errors of different pollutants act on their respective reaction constants, achieving adaptive convergence of parameters during long-term operation.

[0066] The stoichiometric constant matrix described in this article is a parameter set representing one or more reaction constants, oxygen consumption equivalent coefficients, and carbon-nitrogen consumption ratio constants called by the feedforward execution module 125; for a specific controlled object, the corresponding parameters in the parameter set can be called and modified as individual constants.

[0067] The specific implementation of the system of the present invention will be described in detail below with reference to the accompanying drawings.

[0068] See attached document Figure 3 In this embodiment, the sensing and execution layer 110 serves as the physical data source and action execution terminal of the system, directly affecting the consistency of basic data and the accuracy of subsequent control. The internal process environment of the biochemical reactor is complex, and the placement of various sensors and actuators needs to conform to the actual fluid dynamic characteristics.

[0069] Specifically, the influent monitoring unit 111 is installed at the main influent channel of the biochemical reactor to ensure stable water flow and reduce turbulence interference caused by water droplets. The electromagnetic flowmeter and online water quality analyzer integrated within the influent monitoring unit 111 are used to continuously acquire parameters of the untreated source wastewater. The effluent monitoring unit 112 is preferably installed at the effluent outlet of the biochemical reactor, using an ammonia nitrogen sensor, a nitrate nitrogen sensor, and / or a total nitrogen analyzer to monitor the actual water quality after the reaction. Alternatively, the effluent monitoring unit 112 may also include a dissolved oxygen sensor to assist in monitoring the operating status of the aerobic zone. For process fluid distribution, the reflux monitoring unit 113 is installed in the internal and external reflux pipelines connecting the various structures, and is equipped with a pipe-section electromagnetic flowmeter to obtain the actual flow rate and reflux ratio of the mixed liquor.

[0070] On the action execution side, the dosing unit 114 is located at the inlet confluence of the anoxic zone. It uses a variable frequency diaphragm metering pump to control the instantaneous dosing mass of the liquid carbon source by adjusting the operating frequency of the drive motor. Simultaneously, the dosing unit 114 also feeds back the current operating frequency or stroke execution status to the control layer 120. Furthermore, the aeration unit 115 is installed in the aerobic zone, relying on the linkage between the variable frequency centrifugal blower on the main air supply pipeline and the electric regulating valves on the branch pipelines to achieve dynamic distribution of airflow to each aerobic grid.

[0071] To obtain the accurate air supply data required for flow regime correction, the aeration execution unit 115 preferably installs air flow detection devices on the main air supply pipeline and / or each branch pipeline. In embodiments where no separate flow detection device is configured for each branch, the control layer 120 can estimate the actual air supply flow of each branch based on the main pipeline flow, main pipeline pressure, valve opening, blower frequency, and preset valve flow characteristic curves. The selection specifications, flange standards, and electrical wiring of the aforementioned sensors, frequency converters, and regulating valves can be conventionally configured by those skilled in the art based on the actual wastewater treatment plant's treatment scale and pipe diameter requirements; these are well-known technologies in the field and will not be elaborated upon here.

[0072] After acquiring the physical state of the site, the fundamental task in constructing a spatially discretized mapping model is to transform the heterogeneous analog signals at the lower levels into digital features that the control system can recognize and synchronize with time. To this end, the data acquisition module 121 in the control layer 120 establishes a data link with the sensing and execution layer 110 via a fieldbus. As a preferred approach, the fieldbus can employ industrial communication protocols such as Profinet or Modbus RTU. The data acquisition module 121 scans the bus network at a set high-frequency sampling period, reading the raw register values ​​of each lower-level device. These raw register values ​​include standard 4-20mA current or 0-10V voltage signals output by the inlet monitoring unit 111, outlet monitoring unit 112, and return monitoring unit 113, as well as digital status quantities, frequency feedback quantities, valve position opening quantities, and engineering quantity feedback values ​​output by the dosing execution unit 114, aeration execution unit 115, and air flow detection device.

[0073] Since the raw electrical signals cannot be directly used in process calculations, the data acquisition module 121 needs to perform a linear engineering quantity conversion on them. This conversion accurately maps the analog range of the electrical instruments to the actual fluid dynamics or chemical concentration range. The conversion formula is as follows:

[0074] ;

[0075] In the formula, The converted physical engineering quantity values; This refers to the raw electrical signal quantity read in real time. and These are the minimum and maximum electrical signal quantities corresponding to the sensor's range, and these two parameters are usually determined at the instrument's factory or during the field calibration stage. and This corresponds to the upper and lower limits of the measurement of the physical quantity.

[0076] In actual industrial settings, due to the complex chemical reaction measurement cycle of online water quality analyzers and the extremely short instantaneous response cycle of electromagnetic flowmeters, significant physical time differences often exist in the arrival time of different physical parameters at the control layer. Directly merging these parameters for calculation would lead to spatiotemporal misalignment in subsequent volume shift logic. Therefore, the data acquisition module 121 uses the clock of the central processing unit of the control layer 120 as a reference to assign a unified system timestamp to each set of converted engineering quantities. .

[0077] Furthermore, considering the electromagnetic interference caused by the operation of high-power frequency converters and the pulsating characteristics of the fluid itself, the data acquisition module 121 introduces a moving average filtering algorithm to extract effective reference values. This filtering mechanism maintains a fixed-length data queue on the time axis. As the latest sampled data enters, the oldest data is automatically removed, and the average state of the data in the current queue is calculated, thereby smoothing out transient high-frequency noise. The corresponding formula expression is:

[0078] ;

[0079] In the formula, This is the baseline engineering quantity output after filtering and timestamp alignment; The set sliding window depth parameter; This refers to the physical engineering quantities obtained by converting data at corresponding historical sampling times; This is the set high-frequency sampling period.

[0080] In this embodiment, to balance the computational load of the programmable logic controller with the need to capture hydraulic fluctuations, The value range is set to 100ms to 1000ms; window depth parameter The value is then selected between 10 and 50, depending on the frequency of interference in the on-site environment. To avoid algorithm logic dead zones, this should be avoided during system cold start initialization or when the data queue accumulation is insufficient. At any given time, the data acquisition module 121 performs a dynamic average calculation based on the actual accumulated total number of samples. Furthermore, if the system identifies... If the value exceeds the upper or lower threshold range, it is determined that there is an anomaly in the corresponding physical channel, and the current cycle temporarily maintains the valid output value of the previous moment to prevent the model from crashing.

[0081] Through the aforementioned multiple processing mechanisms, the complex physical state collected by the perception execution layer 110 is reliably converted into clean and aligned baseline engineering quantity data and stored in the shared memory area of ​​the control layer 120, providing accurate data input for the subsequent system drive volume displacement matrix.

[0082] See attached document Figure 4 In this embodiment, after the control layer 120 completes the acquisition of the underlying physical data, it needs to transform the complex physical space of the biochemical reactor into a logical data structure that the control system can address efficiently.

[0083] Traditional control schemes typically rely on a fixed time constant to estimate the flow location of pollutants. However, when the influent flow rate fluctuates drastically, the actual residence time of water in the structure changes accordingly, leading to severe distortion of time-based prediction models. Therefore, the matrix construction module 122 in this invention abandons time-dimensional extrapolation and adopts a purely volumetric discrete grid partitioning model. The principle of this model is that the reactor volume is divided into multiple standard volumetric micro-elements. Regardless of fluctuations in the influent flow rate, the fluid always moves forward in physical space at a fixed volumetric step size, thereby completely eliminating the interference of uncertain fluid penetration time.

[0084] To achieve the conversion from the physical space to the logical space, the matrix construction module 122 obtains the civil engineering dimensions and design operating liquid levels of each structure in the biochemical treatment section. Through geometric multiplication and cumulative calculation, the system calculates the total effective physical volume currently participating in the biochemical reaction. Subsequently, the matrix construction module 122 divides this total effective physical volume into equal parts. An incompressible fluid element. In this process, the standard volume constant of the single fluid element is calculated using the following formula:

[0085] ;

[0086] In the formula, The standard volume constant of a single fluid element; The total effective physical volume of the biochemical reactor at the currently set operating liquid level; The total number of spatial discretization grids is set. As a preferred approach, the total number of grids... The value range is determined by a combination of the upper limit of the control system's memory capacity and the spatial control precision required by the process. In the application scenario of conventional municipal wastewater treatment plants, The value is usually configured in the range of positive integers from 100 to 2000 to balance computational load and addressing resolution.

[0087] After determining the mesh size, the system needs to store these micro-data elements in the computational domain. The matrix construction module 122 dynamically allocates contiguous storage blocks in the programmable logic controller's memory space based on the calculated total number of meshes N, thereby constructing a one-dimensional array of length N, which serves as the volumetric displacement matrix for subsequent tracking of fluid motion. Each node element in this matrix corresponds to a fluid micro-element in physical space, and each node encapsulates a specific data structure for storing feature vectors of instantaneous pollutant concentrations such as chemical oxygen demand, ammonia nitrogen, nitrate nitrogen, and / or total nitrogen. For continuous memory allocation and structure array instantiation operations in the control system, those skilled in the art can perform conventional calls based on the selected underlying hardware and configuration software environment. The principles of memory allocation and addressing are well-known in the field and will not be elaborated upon here.

[0088] As a preferred approach, during the system cold start initialization phase, the matrix construction module 122 initializes each node of the volume displacement matrix to the baseline pollutant concentration vector under the design conditions, the average pollutant concentration vector of the most recent stable operating cycle, or the concentration vector composed of the first batch of effective influent water quality data. Before the matrix has completed the first round of effective filling, the control commands output by the feedforward execution module 125 can adopt preset conservative values ​​or the most recent stable operating condition values ​​to avoid control fluctuations caused by the lack of initial boundary conditions during the cold start phase.

[0089] After establishing the volumetric displacement matrix, it is also necessary to anchor the key process operation points along the biochemical reactor into this data structure. Since the installation positions of physical equipment such as dosing pumps and blowers are fixed, the matrix construction module 122 maps them into a volumetric displacement matrix based on the physical volume accumulation of each piece of equipment from the inlet. A fixed integer index in the context. Let a physical device node... The cumulative effective volume along the feedwater path from the inlet of the biochemical reactor is Then the mapping formula for the static constant index of the node in the matrix is:

[0090] ;

[0091] In the formula, The calculated integer constant index number; This is the rounding function; This refers to the cumulative fluid volume from the physical inlet to the target equipment node. This is the standard volume constant obtained from the aforementioned calculations.

[0092] Based on this static mapping algorithm, the system can automatically calculate the constant addressing coordinates of various core execution and dispensing mechanisms. Specifically, the dosing point in the anoxic zone where the dosing execution unit 114 is located is mapped to the anoxic zone index. The physical span of the aerobic zone where the aeration execution unit 115 is located is mapped to a set of continuous index intervals. ,in and These correspond to the physical volume accumulation positions at the beginning and end of the aerobic zone, respectively. Similarly, the extraction point of the internal reflux pipeline is usually located at the end of the aerobic zone, and it is mapped to the reflux extraction index. The internal recirculation access point located at the beginning of the anoxic zone is mapped as a recirculation access index. .

[0093] To ensure the algorithmic integrity of the addressing process and avoid array out-of-bounds dead zones caused by computational truncation or abnormal liquid levels, the matrix construction module 122 is configured with boundary protection logic. When the calculated constant index value is less than 1, the system forcibly limits it to 1; when the index value is greater than the total number of grid cells... When this happens, the forced amplitude limit is... This static anchoring mechanism accurately projects continuous physical structures and their attached electromechanical equipment into a one-dimensional digital space, providing a reliable data layer foundation for subsequent dynamic tracking and equipment drive control based on real fluid accumulation.

[0094] See attached document Figure 5 After completing the static mapping from physical space to logical matrix, the system needs to establish an update mechanism for this data structure that is synchronized with the actual physical process. Existing feedforward control typically relies on the system clock as a trigger source, extrapolating the diffusion of pollutants in the biological treatment tank under the assumption of constant influent flow. However, when faced with the dynamic changes in influent flow in municipal wastewater treatment plants, time-based extrapolation often results in significant spatiotemporal misalignment.

[0095] To address this issue, the control layer 120 in this embodiment is equipped with a volume drive module 123, which uses the accumulated influent volume instead of the system clock as the shift drive source for the main control program. The principle is to convert the time-fluctuating flow rate into a definite spatial displacement; specifically, as long as the total volume of wastewater entering the reactor reaches the capacity of a discrete grid, regardless of the time elapsed, the system logically pushes the internal tracking data forward by a micro-element step.

[0096] Based on the above principles, the system needs to calculate the amount of water entering the structure in real time within the control cycle. The volume drive module 123 acquires the instantaneous inflow flow rate transmitted by the inflow monitoring unit 111 and performs continuous time integration on it within the operation cycle of the programmable logic controller. In discrete industrial control systems, this integration operation is specifically manifested as the accumulation of the product of flow rate and time according to the sampling period, used to accurately measure the total amount of water actually flowing into the biochemical reactor. The corresponding continuous integral equation expression is:

[0097] ;

[0098] In the formula, From the previous trigger time Up to the current system time The cumulative influent volume; The instantaneous inflow rate collected by the inflow monitoring unit 111; This serves as the reference time for system initialization or the last displacement trigger. As a preferred approach, the discrete step length of this integral calculation is kept consistent with the sampling period of the control system, set between 100ms and 1000ms, to ensure effective capture of transient hydraulic fluctuations.

[0099] As the integration process continues, accurately determining the timing of data displacement becomes crucial for achieving synchronous tracking. The volume drive module 123 continuously compares the current accumulated influent volume within each program scan cycle. The standard volume constant determined by matrix construction module 122 When detected When the system determines that the actual inflowing sewage has equivalently filled a physical grid cell, a global reference shift event is triggered.

[0100] Furthermore, to ensure the completeness of the algorithm logic, when a certain control cycle detects... ,in When the value is a positive integer greater than or equal to 1, the system calculates the number of shifts required in this cycle. and execute continuously Sub-global reference shift. To avoid truncating and discarding minute excess volumes generated during the high-frequency scanning cycle, the system completes... After the shift, the integral accumulator is not absolutely cleared. Instead, the system performs a modulo subtraction operation, setting the new accumulator base to 0. This eliminates the dead zone of volume calculation omission during long-term operation.

[0101] After a global reference shift event is triggered, the data structure needs to synchronously reflect the spatial displacement state of the water body. The volume drive module 123 adjusts the volume displacement matrix in memory. Execute the instruction to extend the array elements in ascending order of their indices. Specifically, the system starts from the last node of the matrix. Start by reversing the order of the previous node. Assign the entire data structure to the current node This process continues until node 2 receives data from node 1. Through this reverse assignment process, the pollutant characteristics stored in each element of the matrix complete a forward shift in logical space. When multiple global reference shifts need to be performed within a certain control cycle, the reverse assignment process is repeated. Next, or equivalently, sequentially continue in the direction of index increment. Each node has a step size. For batch sequential operations of array elements in a control system, those skilled in the art can call the block move instruction or the first-in-first-out queue function of the programmable logic controller. The underlying storage and addressing logic is well-known in the art and will not be described in detail here.

[0102] As the original data continues, the first and second nodes of the matrix... The storage space is released. To form a complete data flow loop, the system needs to inject physical parameters representing the latest water inflow status into this node. The volume drive module 123 extracts this integration period. The influent pollutant concentration vector is used to calculate the total mass of material input during the cycle, combined with the instantaneous flow rate, to obtain the characteristic average concentration of the new fluid micro-element entering the system's inlet. The corresponding mathematical calculation process is as follows:

[0103] ;

[0104] In the formula, This is the new fluid element feature concentration vector assigned to the first and second nodes; This represents the instantaneous concentration vector of pollutants in the influent collected at the corresponding time.

[0105] When multiple shifts are triggered within a certain control cycle, the volume drive module 123 calculates the characteristic concentration of multiple new fluid micro-elements according to the influent flow rate and influent pollutant concentration in each sub-integral interval, and assigns the values ​​to the nodes released at the head of the matrix in chronological order; the above formula corresponds to the characteristic concentration calculation method of a single new fluid micro-element.

[0106] In practical industrial applications, online water quality probes require regular maintenance. As a preferred approach, if the online water quality analyzer of the influent monitoring unit 111 is in automatic cleaning or calibration mode during the integration cycle, resulting in missing real-time concentration data, the system will automatically recall the data from the previous valid cycle. Numerical substitution calculations are performed to prevent the computational link of the volume tracking model from crashing due to signal interruption. After the calculation is completed, the volume drive module 123 outputs the obtained data. Assigned to the first node of the matrix And update the system's starting integration time simultaneously. This allows for a single data addressing and model update cycle driven by the actual physical volume.

[0107] When performing multiple shifts, the system updates the system's starting integration time uniformly after assigning values ​​to the head node for the corresponding number of shifts.

[0108] See attached document Figure 6In this embodiment, the system completes the basic linear spatial extrapolation through the volume-driven module 123. Under the ideal flow assumption, fluid micro-elements move unidirectionally only according to the physical spatial sequence. Since biochemical reaction processes typically include pipe network topologies such as internal reflux of nitrification liquid and external reflux of sludge, some of the mixed liquid at the end of the structure will be extracted and re-transported to the front end of the structure. If only the basic array sequential update is maintained, the volume displacement matrix will not be able to characterize the material concentration dilution and mixing process caused by the reflux. To this end, the control layer 120 is configured with a flow state correction module 124 to synchronously introduce a nonlinear data crossing and re-injection mechanism in the discretized mapping model. The principle of this mechanism is that, based on the actual physical pipe network connectivity, at the corresponding index node of the logical matrix, the data of the crossing node and the sequential node are weighted and fused according to the law of mass conservation.

[0109] The operational logic of the flow correction module 124 is strictly synchronized with the global reference shift event. In specific implementation, the cross-node data re-injection process is executed according to the following steps:

[0110] Step S501: Extract dynamic reflux parameters in real time and perform dimensionless processing.

[0111] When the system triggers a global reference shift event, the flow correction module 124 synchronously retrieves the historical acquisition records of the return flow monitoring unit 113 within that integration period. To maintain spatial and temporal consistency in hydraulic calculations, the flow correction module 124 performs continuous-time integration on the instantaneous return flow rate transmitted by the return flow monitoring unit 113, calculating the cumulative return flow volume within that triggering period. The corresponding integral equation is:

[0112] ;

[0113] In the formula, The time since the last shift trigger Up to the current system time The cumulative reflux volume; The instantaneous return flow rate is collected in real time by the return flow monitoring unit 113.

[0114] After obtaining the cumulative backflow volume, the flow regime correction module 124 converts it into a universal calculation coefficient that is not limited by the absolute size of the structure. Specifically, the system calls the standard volume constant determined during the initialization phase of the matrix construction module 122. Calculate the dynamic reflux ratio constant at the current moment. The formula is:

[0115] ;

[0116] In the formula, This is a dimensionless dynamic reflux ratio constant. As a preferred method, in the nitrogen and phosphorus removal processes of conventional municipal wastewater treatment plants, The effective value range is between 0.5 and 4.0. The specific value is dynamically determined by the actual operating frequency of the on-site return pump and the hydraulic resistance of the pipeline network.

[0117] Step S502: Perform cross-node data re-injection calculation based on the principle of quality balance.

[0118] After the underlying data structure completes the array traversal from beginning to end, it enters the backflow access index. The data node at the location actually stores the fluid micro-element features extending from the upstream node. Considering that the water body at this physical location is actually formed by the confluence of upstream downstream sewage and the mixed liquid extracted by the return pump, the flow correction module 124 extracts the corresponding return extraction point index from the volume displacement matrix. The data vector at the specified location, combined with the previously calculated dynamic backflow ratio constant, is used to perform a weighted summation and overlay on the data at the target node. The mathematical model for this re-injection update process is expressed as:

[0119] ;

[0120] In the formula, The updated target node stores the fluid micro-element feature concentration vector; This is the upstream feature concentration vector temporarily stored at this node after the underlying array is extended; This is to directly extract the feature concentration vector at the index of the self-reflux extraction point.

[0121] To ensure the integrity of the control algorithm, the system is equipped with an adaptive mechanism for abnormal start-up and shutdown conditions of field equipment. When the backflow monitoring unit 113 detects an instantaneous flow rate of zero due to backflow pump shutdown or pipeline maintenance, the system calculates the cumulative backflow volume. and dynamic reflux ratio constant Automatically resets to zero. At this point, the above heavy betting formula is equivalent to... This means that the system automatically maintains the original push-streaming state, thereby effectively avoiding algorithm dead zones caused by interruptions in the computational logic.

[0122] Furthermore, if multiple cross-pipelines, such as internal nitrification liquid recirculation and external sludge recirculation, exist simultaneously in the biochemical reaction system, the flow regime correction module 124 will perform independent mixed weighted calculations on the corresponding constant index nodes according to the actual physical order of water flow into the structures. Through the above-mentioned nonlinear data re-injection mechanism, the volume displacement matrix can accurately reflect the material circulation distribution under complex pipe network topology, eliminating the systematic bias of the pure plug flow space model.

[0123] See attached document Figure 7 In this embodiment, in addition to the macroscopic backflow of materials caused by the topological network, the microscopic hydraulic turbulence caused by the operation of the aeration equipment inside the bioreactor also alters the ideal plug flow state of the system. In the aerobic section of the biochemical reaction process, the bubbles released by the bottom microporous aeration disc form a gas-liquid two-phase flow. This gas-liquid mixed flow induces eddies and axial mixing of the liquid phase during its upward movement, leading to physical infiltration of fluids and pollutants in adjacent spatial regions. If the system maintains a completely isolated fluid micro-element unidirectional extension model without introducing physical space material diffusion compensation, a step deviation in the characteristic concentration of local logic nodes will occur.

[0124] To correct this local model distortion, the flow correction module 124 in the control layer 120 introduces a lateral data smoothing mechanism based on the reverse representation of the execution end parameters after data re-injection. The principle of this mechanism is to use the actual air supply of the bottom aeration execution device as a characterizing variable of the hydrodynamic disturbance intensity, and thereby infer the degree of data diffusion between adjacent micro-elements, thus realizing the logical reconstruction of the physical backmixing state without the need for additional flow velocity sensors.

[0125] To establish a mathematical relationship between equipment actions and flow regime changes, the flow regime correction module 124 performs lateral flow regime correction calculations according to the following steps:

[0126] Step S601: Extract the operating parameters of the aeration unit and perform dimensionless mapping calculation of the mixing intensity.

[0127] The aeration intensity is positively correlated with the degree of fluid backmixing within the area; the flow correction module 124 acquires the actual instantaneous air supply flow rate of each aerobic zone air supply branch in the aeration execution unit 115 in real time; in some embodiments, if the branch is not equipped with a separate air flow detection device, the flow correction module 124 reads the valve opening, blower frequency, main pipe pressure and / or main pipe flow rate of the branch, and calculates the estimated air supply flow rate of the corresponding branch in combination with the preset valve flow characteristic relationship, which is used as... This parameter is then used in subsequent calculations. To convert this physical parameter into a diffusion index that can be accessed via a volumetric displacement matrix, the system invokes a preset nonlinear mapping function to convert the instantaneous air supply flow rate into a dynamic backmixing coefficient that characterizes the axial mixing intensity of adjacent fluid micro-elements. The mapping calculation formula is as follows:

[0128] ;

[0129] In the formula, The dimensionless dynamic backmixing coefficient is calculated. The instantaneous air supply flow rate of the aeration execution unit 115 corresponding to the aerobic zone is read in real time by the system. The maximum permissible physical air supply flow rate for this aerobic zone is determined by the hardware specifications. The set maximum backmixing coefficient boundary threshold; For nonlinear morphological parameters. As a preferred approach, considering the numerical convergence of the multi-point weighted algorithm and the actual fluid dynamics boundary, The effective value range is set between 0.05 and 0.30 for the morphological parameters. Based on an empirical model of gas-liquid turbulent diffusion, the value is set between 1.2 and 1.5.

[0130] As a preferred approach, the maximum back-mixing coefficient boundary threshold in the aforementioned nonlinear mapping function and morphological parameters The dynamic backmixing coefficient can be calibrated based on the results of tracer tests during the commissioning period, the fitting results of historical operating data, and / or the results of fluid dynamics simulation, so that the calculated results of the dynamic backmixing coefficient match the actual axial mixing characteristics of the aerobic zone.

[0131] Step S602: Perform a horizontal weighted smoothing operation on the target node within the volume displacement matrix.

[0132] After obtaining the real-time dynamic back-mixing coefficient, the flow regime correction module 124 performs traversal calculations on the logical nodes within the aerobic zone span. The system determines whether the volume displacement matrix is ​​located within the mapping interval. All consecutive nodes within this index interval are located in aerated, disturbed water. For any target node within this index interval... The system employs a weighted algorithm of adjacent three points to correct the diffusion simulation of the concentration distribution within the state matrix. The specific discrete mathematical expression is as follows:

[0133] ;

[0134] In the formula, To smooth the target node Internally update the stored concentration vector of fluid micro-element features; This is the original basic feature concentration vector of the target node before smoothing; and These are the basic feature concentration vectors of the upstream and downstream nodes that are adjacent to the target node, respectively.

[0135] To ensure the integrity of the control algorithm under boundary conditions, the flow regime correction module 124 is equipped with out-of-bounds compensation logic. This applies when the target node is located at the physical starting point of the aerobic zone. or physical end When the adjacent node on one side is in a non-aeration disturbance area, the system sets the back-mixing weight coefficient of the adjacent node on the boundary side to zero, and adds this part of the weight value to the target node's own retained weight coefficient. In order to ensure quality conservation between nodes and avoid logical dead zones caused by address overflow, the system is designed to adapt to abnormal operating conditions. For example, when the aeration execution unit 115 receives an intermittent shutdown command, resulting in a decrease in the actual instantaneous air supply flow rate... When reduced to zero, the dynamic backmixing coefficient calculated by the system Synchronous zeroing. At this point, the above diffusion smoothing formula is equivalent to... The system model logic automatically switches back to the basic push-flow extension state. This mechanism enables the system to directly map the flow field evolution caused by changes in the operating equipment state, reconstructing the static volumetric displacement matrix into a fluid dynamics calculation framework with dynamic diffusion characterization capabilities.

[0136] See attached document Figure 8 In this embodiment, the volumetric displacement matrix processed by the flow correction module 124 can accurately map the physicochemical parameter distribution of the fluid micro-elements inside the bioreactor. Traditional control schemes cannot accurately quantify the penetration time of water flow in the structure when the influent flow rate fluctuates, which can easily lead to the dosing or aeration commands being ahead of or behind the actual arrival point of the pollutants. To solve the above-mentioned control lag problem, a feedforward execution module 125 is configured inside the control layer 120. Its principle is to use the constructed spatial discretization model to directly trigger the execution action of the underlying equipment based on precise spatial addressing, so that the control command issuance is synchronized with the physical flow of materials in time and space.

[0137] As a preferred approach, after each global baseline shift event, the control layer 120 sequentially performs reflux data re-injection, aerobic zone lateral smoothing, and feedforward control calculations in a fixed order. For dosing control, the feedforward execution module 125 performs event-driven calculations for the fixed-point constant index of the anoxic zone after the shift event; for aeration control, the feedforward execution module 125 performs interval integration calculations for the continuous index interval of the aerobic zone after the shift event or according to a preset control cycle.

[0138] To ensure precise issuance of control commands, the feedforward execution module 125 executes the following logical control steps:

[0139] Step S701: Perform spatial fixed-point data extraction to obtain real-time material load.

[0140] During each control cycle, the system continuously monitors the state of specific constant indices in the volume displacement matrix. Since the matrix construction module 122 has already anchored the underlying execution device to the digital space, the feedforward execution module 125 directly addresses the dosing point indexes in the hypoxic zone from the memory space. and the continuous index interval of the aerobic zone When a global reference shift event is triggered at the lower level, the system determines that the fluid micro-element carrying the corresponding pollutant load has been physically moved to the location of the actuator. At this time, the feedforward execution module 125 immediately extracts the feature concentration vector that has been shifted to the aforementioned constant index node. This vector structure encapsulates key water quality prediction parameters such as chemical oxygen demand, ammonia nitrogen, and nitrate nitrogen within the fluid micro-element, forming the basic data source for subsequent material demand calculation.

[0141] In step S702, the feedforward execution module 125 performs absolute material requirement calculation and drive signal conversion.

[0142] After acquiring the data features extracted through spatial addressing, they need to be converted into specific equipment execution instructions. In the carbon source dosing control stage in the anoxic zone, the system retrieves the dosing point index. Matrix data vector at location Combined with the internally solidified stoichiometric constant matrix A denitrification carbon source gap assessment is performed. The principle of this calculation process is to convert the relative pollutant concentration difference within a fluid micro-element into the absolute mass of material requiring the removal of that pollutant. The corresponding carbon source demand calculation model is as follows:

[0143] ;

[0144] In the formula, The absolute carbon source mass required to eliminate the carbon source gap in the denitrification reaction within this fluid micro-element; and These are the data vectors extracted from the matrix. Instantaneous predicted concentrations of nitrate nitrogen and chemical oxygen demand; The theoretical carbon-to-nitrogen consumption ratio constant is set between 4.0 and 6.0, with the specific value depending on the type of carbon source added on-site (such as sodium acetate, glucose, etc.). is the standard volume constant for a single fluid element. This formula introduces... This is a limit boundary function, mainly used to deal with the condition that the influent itself has sufficient carbon source, and to avoid the system outputting negative value addition instructions causing algorithm dead zones or equipment errors.

[0145] Complete absolute carbon source quality After calculation, the feedforward execution module 125 converts it into a hardware-recognizable drive signal. Based on the relationship between the calibrated reagent concentration and stroke displacement of the variable frequency dosing pump in the dosing execution unit 114, the system calculates the corresponding pulse frequency. The feedforward execution module 125 then sends an electrical control signal to the dosing execution unit 114, driving the variable frequency dosing pump to reach the next... Previously, an equal amount of liquid carbon source was precisely and quantitatively added. This execution logic, which uses volumetric steps for quantitative drug delivery, effectively prevents drastic fluctuations in influent flow rate from interfering with the timing of drug administration.

[0146] In step S703, in the air supply control stage of the aerobic zone, the feedforward execution module 125 uses a spatial integration strategy to process continuous data.

[0147] In addition to single-point dosing control, the biochemical reaction also involves the continuous air supply requirement across a large water area; given that the aerobic zone covers a relatively long physical grid, the system extracts index intervals. The characteristic concentration vectors of all continuous fluid elements within the water body are used to calculate the overall theoretical oxygen demand of the microbial community in that water area. The formula for calculating this theoretical oxygen demand is as follows:

[0148] ;

[0149] In the formula, This represents the total theoretical oxygen demand of all fluid micro-elements within the aerobic zone. This refers to the index number of the specific node within the aerobic interval; and These correspond to the constant index numbers at the beginning and end of the aerobic zone, respectively. and These are the predicted concentrations of ammonia nitrogen and chemical oxygen demand extracted from the corresponding node jjj, respectively. The theoretical oxygen consumption equivalent coefficient for the nitration reaction is set to 4.57 as a preferred method. The theoretical oxygen consumption equivalent coefficient for the carbonization reaction is between 0.5 and 1.2, and its value is affected by the microbial cell yield.

[0150] Obtain Total Oxygen Demand Subsequently, the feedforward execution module 125, combining the standard oxygen transfer efficiency of the aeration equipment with the current water temperature parameters, calculates the required actual airflow rate and converts it into a standard opening command and frequency signal. As a preferred embodiment, the current water temperature parameter is provided by a temperature detection device installed inside the bioreactor or in related pipelines; in embodiments without a temperature detection device, the current water temperature parameter can also be a preset seasonal empirical value or a manually set value within the control system. The standard oxygen transfer efficiency is the calibration parameter of the aeration equipment, or an effective oxygen transfer efficiency parameter corrected by considering the operating liquid level, wastewater temperature, and equipment operating status.

[0151] The system synchronously sends instructions to the aeration execution unit 115, which controls the valve opening of the electric regulating valve and the speed of the variable frequency blower. Regarding the underlying PID closed-loop follow-up control mechanism of the frequency converter and electric valve after receiving the control signal, those skilled in the art can deploy it using conventional electrical automation drive logic. Its underlying follow-up principle is well-known in the field and will not be elaborated upon here.

[0152] See attached document Figure 9 In this embodiment, the feedforward execution module achieves spatiotemporal synchronization between control commands and the physical flow of materials based on the volume displacement matrix. However, the activity of the microbial community inside the bioreactor changes slowly with seasonal variations in water temperature, influent composition, and sludge age. This means that the stoichiometric constants solidified in the early stages of the system are difficult to maintain absolute accuracy over a long period. If the system relies solely on fixed theoretical constants for feedforward calculations for an extended period, static model biases will inevitably accumulate. Traditional feedback control schemes, when faced with fluctuations in effluent quality, typically convert measurement errors directly into the inverter's opening degree or frequency compensation signal. This direct intervention in the underlying hardware output disrupts the material balance established by the feedforward control based on volume calculations, leading to frequent fluctuations in system control commands.

[0153] To address this issue, the control layer 120 is equipped with a feedback adjustment module 126. Its principle is to construct a closed-loop data link and use the measurement deviation at the water outlet to dynamically fine-tune the stoichiometric constants called by the underlying layer of the feedforward model. This achieves adaptive convergence of the digital mapping model without interfering with the hardware execution rhythm.

[0154] To achieve the closed-loop mechanism based on model constant fine-tuning, the specific operation of the feedback adjustment module 126 includes the following control steps:

[0155] Step S801: Extract the predicted residual sequence and perform spatiotemporal alignment.

[0156] As the cumulative influent volume increases, the fluid micro-elements within the volume displacement matrix are gradually pushed to the end of the matrix. The pollutant concentration characteristics encapsulated within the end nodes represent the theoretically predicted water quality of this portion of the water flowing out of the biochemical reactor effluent. The feedback adjustment module 126 extracts the characteristic concentration vector output by the end nodes of the volume displacement matrix in real time and compares the pollutant concentration components corresponding to the monitoring items of the effluent monitoring unit 112 with the actual measured target pollutant concentration data of the effluent monitoring unit 112. Considering that online chemical analysis instruments in wastewater treatment plants (such as ammonia nitrogen analyzers, nitrate nitrogen analyzers, and total nitrogen analyzers) typically have measurement delays, the system needs to perform spatiotemporal alignment processing on the data to ensure the scientific validity of the error calculation. The system allocates a data buffer queue in memory to temporarily store the predicted output values ​​of the end nodes, ensuring that the predicted water body at the time of difference corresponds strictly to the water body actually sampled by the analysis instrument in terms of physical batch. The corresponding residual calculation formula is:

[0157] ;

[0158] In the formula, For the current system time The generated prediction residual values; The actual effluent pollutant concentration output in real time by the effluent monitoring unit; This is the instrument response delay time set according to the sampling and measurement cycle of the on-site online analytical instrument; Delay for data buffer queue The theoretically predicted concentration of pollutants corresponds to the terminal node output after a certain time.

[0159] As a preferred embodiment, if the effluent monitoring unit 112 is located at the effluent outlet of the biochemical reactor, the aforementioned spatiotemporal alignment mainly compensates for the sampling and measurement delay of the online analytical instrument itself. If, in other embodiments, the effluent monitoring unit 112 is located at the effluent outlet of a subsequent structure, the system further introduces the equivalent residence delay parameter of the subsequent structure to perform additional delay compensation on the predicted value at the end of the matrix before participating in the residual calculation.

[0160] To ensure the integrity of the closed-loop algorithm and prevent model collapse due to external sensor failure, the feedback adjustment module 126 is equipped with input anomaly shielding logic. As a preferred method, when the effluent monitoring unit 112 is in automatic cleaning or calibration mode, or when the output... When the measured value exceeds the set reasonable physical range, the system automatically determines that the current measurement value is invalid and forces the predicted residual value of the current period to be invalidated. This mechanism effectively prevents erroneous error signals from entering subsequent control stages and causing algorithm divergence.

[0161] Step S802: Dynamic fine-tuning of chemical reaction constants is performed based on discrete PID algorithm.

[0162] After acquiring a continuous and aligned prediction residual sequence, the system transforms this error signal into a driving force for correcting the underlying model. The feedback adjustment module 126 inputs the generated prediction residual sequences for each target pollutant into one or more system-level proportional-integral-derivative (PID) controllers. For the ammonia nitrogen residual sequence, the system uses the compensation output from the PID controller to correct the nitrification-related oxygen consumption equivalent coefficient and / or nitrification reaction constant called in the feedforward execution module 125; for the nitrate nitrogen residual sequence and / or total nitrogen residual sequence, the system uses the compensation output from the PID controller to correct the theoretical carbon-nitrogen consumption ratio constant called in the denitrification carbon source addition calculation; in implementations employing multi-pollutant joint control, the feedback adjustment module 126 can also combine multiple PID outputs into a parameter correction vector according to preset weights to correct multiple stoichiometric constants in parallel. The mathematical model for its discrete positional operation is as follows:

[0163] ;

[0164] In the formula, This is the dynamic compensation amount of the model constants calculated at the current moment; and These are the prediction residuals between the current sampling time and the previous sampling time, respectively; This refers to the sampling operation cycle of the control system. This represents the total cumulative number of samples since the system was started or reset. The gain is controlled proportionally. For integral control gain; This refers to the differential control gain. For the engineering tuning of the controller's internal parameters, those skilled in the art can use empirical trial-and-error methods or the critical proportional gain method for conventional debugging, for example, by setting... The initial value range is set between 0.01 and 0.1 to ensure a smooth transition during constant fine-tuning. Furthermore, to eliminate the integral saturation dead zone caused by long-term operation, anti-integral saturation logic is enabled internally in the controller. When the compensation amount reaches the set limit, the system automatically stops further integration of the error.

[0165] The dynamic compensation amount is calculated. Subsequently, the feedback adjustment module 126 directly applies it to the stoichiometric constant matrix called by the feedforward execution module. Taking the control of denitrification carbon source addition in the anoxic zone as an example, the system superimposes this compensation amount onto the theoretical carbon-nitrogen consumption ratio constant. The specific update mechanism is as follows:

[0166] ;

[0167] In the formula, This is the updated actual execution carbon and nitrogen consumption ratio constant used for the current feedforward calculation cycle; This is the basic carbon-nitrogen consumption ratio constant called in the previous calculation cycle.

[0168] Similarly, for other key reaction parameters called in the aeration control of the aerobic zone, such as the oxygen consumption equivalent coefficient and nitrification reaction constant, the system can also implement the above-mentioned PID constant correction logic in parallel based on the corresponding residual sequence.

[0169] To prevent excessive compensation from the PID calculations under extreme conditions from causing system instability, the feedback control module 126 sets a hard protection boundary for the updated constants. As a preferred approach, the updated carbon-nitrogen consumption ratio constant... The system is forcibly clamped within a reasonable biochemical reaction threshold range of 3.0 to 8.0. If the calculated result is below 3.0, the output is limited to 3.0; if it is above 8.0, the output is limited to 8.0. Similarly, for other key reaction parameters such as the oxygen consumption equivalent coefficient used in the aeration control of the aerobic zone, the system also sets limit ranges that match their physical meanings. Through this control mechanism that intervenes in the underlying model constants rather than directly intervening in the hardware, the system retains the rapid response capability of feedforward control to sudden changes in influent flow rate, while also enabling the digital spatial mapping model to acquire long-term optimization capabilities that adapt to the slow evolution of on-site microbial activity.

[0170] Furthermore, when any key feedback channel has no valid data for more than a preset time period, the system keeps the stoichiometric constant of the most recent stable period unchanged and makes the feedforward execution module 125 continue to operate according to the most recent stable operating parameters or preset conservative parameters, while outputting a fault alarm signal to avoid model divergence caused by erroneous feedback signals.

[0171] Through the above settings, the control layer 120 uses the volume displacement matrix in memory as a unified data carrier. After each global reference shift event triggered by the accumulated influent volume, it sequentially completes data updates, refluxing, backmixing smoothing, feedforward execution, and feedback constant correction. This ensures closed-loop coupling of influent monitoring, refluxing, aeration feedback, chemical dosing execution, and effluent feedback within the same spatial discretization framework. Therefore, the system can maintain spatiotemporal consistency between pollutant spatial location prediction and actuator actions when the influent flow rate dynamically changes, and can gradually correct model parameters through feedback adjustment, ensuring control accuracy and stability during long-term operation.

[0172] To further verify the feasibility and technical effectiveness of the wastewater treatment plant operation support system based on dynamic influent flow and real-time water quality monitoring described in this invention in actual operating scenarios, the following description is provided in conjunction with specific application embodiments, comparative verification embodiments, and related drawings. It should be understood that the operating parameters, statistical results, and accompanying drawings in the following embodiments are only used to illustrate that the solution of this invention can be implemented and has the expected technical effects, and do not constitute a limitation on the scope of protection of this invention.

[0173] See attached document Figure 10 To further illustrate the specific application of the system described in this invention, the following example uses a system employing A... 2 This explanation will take a municipal wastewater treatment plant using the / O biochemical treatment process as an example.

[0174] In this embodiment, the wastewater treatment plant is designed to treat 50,000 m³ of wastewater. 3 The biochemical treatment unit comprises an anaerobic zone, an anoxic zone, and an aerobic zone. The anoxic zone has a carbon source addition point, the aerobic zone has multiple gas supply branches, and the aerobic zone has an internal reflux extraction point at its end. The anoxic zone has an internal reflux inlet point at its front end. The total effective physical volume of the biochemical reactor at normal operating liquid level is 10,000 m³. 3 The matrix construction module 122 divides the total effective physical volume into 500 incompressible, equal-volume fluid micro-elements, thus defining the standard volume constant of a single fluid micro-element. 20m 3 .

[0175] In terms of hardware deployment, the influent monitoring unit 111 is located at the main influent channel of the biochemical reactor to continuously collect instantaneous influent flow rate, influent chemical oxygen demand, influent ammonia nitrogen, and influent total nitrogen concentration; the effluent monitoring unit 112 is located at the end of the biochemical reactor to continuously collect effluent ammonia nitrogen and effluent total nitrogen concentration; the reflux monitoring unit 113 is located in the internal reflux pipeline to continuously collect the instantaneous internal reflux flow rate; the dosing execution unit 114 is located at the front end of the anoxic zone to add external carbon sources; and the aeration execution unit 115 is located in each air supply zone of the aerobic zone to adjust the actual air supply flow rate of each zone.

[0176] In this embodiment, the sampling period of the data acquisition module 121 is set to 1 second, and the moving average window depth parameter W is set to 20. The matrix construction module 122 maps the dosing points in the anoxic zone to constant indices based on the actual spatial relationships of the pool. =120, mapping the beginning and end of the aerobic zone to respectively. =180 and =420, mapping the internal reflux extraction point to =410, mapping the internal return access point to =105.

[0177] During system operation, the volume drive module 123 continuously integrates the instantaneous influent flow rate over time. When the cumulative influent volume reaches 20m³... 3 When the cumulative inflow volume reaches 40m³, a global reference shift event is triggered. 3 60m 3 If the value is higher, a global reference shift event is triggered the corresponding number of times based on the integer multiple relationship reached, and the remaining volume is retained as the initial integral value for the next control cycle. (Appendix) Figure 10 The diagram illustrates the correlation between dynamic changes in influent flow rate and the number of shifts per unit time. As the influent flow rate increases, the cumulative influent volume reaches the volume of one or more fluid micro-elements more quickly per unit time, thus triggering a corresponding increase in the number of global baseline shifts. Conversely, as the influent flow rate decreases, the shift frequency decreases accordingly. This process demonstrates that the present invention does not drive model updates based on a fixed time interval, but rather updates the volume displacement matrix synchronously based on the actual cumulative influent volume.

[0178] When the internal recirculation pipeline is in operation, the flow correction module 124 calculates the dynamic recirculation ratio constant based on the instantaneous recirculation flow rate collected by the recirculation monitoring unit 113, and indexes the end extraction point of the aerobic zone. The data vector at that location is re-injected into the index of the front-end access point in the hypoxic zone according to the mass conservation principle. The flow pattern is characterized by the remixing process caused by internal backflow. Simultaneously, the flow pattern correction module 124 calculates the dynamic backmixing coefficient based on the actual or estimated air supply flow rate of each air supply branch in the aerobic zone, and adjusts the flow pattern between the aerobic zones. The continuous nodes within the system perform a transverse weighted smoothing operation to characterize the local axial mixing effect caused by aeration disturbance.

[0179] See attached document Figure 11 This illustrates the distribution of typical pollutant concentrations such as chemical oxygen demand (COD), ammonia nitrogen, and nitrate nitrogen along the index direction in the volume displacement matrix during a single runtime period. Figure 11 It is evident that after introducing recirculation and backmixing smoothing correction, the distribution of pollutant concentration along the spatial index direction of the pool no longer exhibits a simple step migration under ideal plug flow conditions, but rather more closely resembles the concentration distribution state formed by the combined effects of recirculation and aeration under actual process conditions.

[0180] The current feedforward execution module 125 detected the index of the dosing point located in the hypoxia zone. When the fluid micro-element experiences an increase in nitrate nitrogen concentration and a shortage of usable organic matter, the system calculates the required external carbon source mass for the current fluid micro-element based on the currently invoked theoretical carbon-nitrogen consumption ratio constant, and converts it into a control frequency signal for the dosing execution unit 114; when the system operates in the aerobic zone... After traversing all nodes, if the calculated total theoretical oxygen demand in the interval increases, the operating frequency of the blower and the opening of the corresponding branch regulating valve are increased simultaneously to achieve feedforward aeration regulation that matches the spatial arrival position of pollutants.

[0181] The feedback adjustment module 126 performs spatiotemporal alignment of the predicted effluent ammonia nitrogen and total nitrogen concentrations at the end nodes of the volume displacement matrix with the measured values ​​of the effluent monitoring unit 112, and then calculates the difference to form ammonia nitrogen residual sequences and total nitrogen residual sequences, respectively. The PID controller then slowly corrects the nitrification-related oxygen consumption equivalent coefficient and carbon-nitrogen consumption ratio constant to adapt to long-term model deviations caused by changes in sludge activity, water temperature, and influent composition.

[0182] In this embodiment, after the system runs continuously, it can dynamically update the predicted spatial location of pollutants in the tank according to the changes in the influent flow rate. This makes the carbon source addition and aeration adjustment actions no longer dependent on a fixed hydraulic retention time, but directly dependent on the actual cumulative influent volume and real-time water quality status, thereby improving the consistency between the control timing and the actual operating conditions.

[0183] See attached document Figure 12 To further verify the feasibility and operational effectiveness of the system described in this invention under dynamic water inflow conditions, the wastewater treatment plant was selected as the verification object, and the scheme of this invention was compared with the traditional operation assistance method based on fixed time delay calculation.

[0184] In this embodiment, two operating schemes were set up: a control group and an experimental group. The control group adopted a traditional operation assistance method based on a fixed time delay. This method calculates the arrival time of pollutants at the anoxic zone dosing point and each aeration zone in the aerobic zone based on the average residence time under the design average flow conditions, and adjusts carbon source addition and aeration accordingly. The experimental group adopted the wastewater treatment plant operation assistance system described in this invention, based on dynamic influent flow and real-time water quality monitoring. The system updates the volume displacement matrix according to the cumulative influent volume, and combines recirculation re-injection, backmixing correction, feedforward execution, and feedback constant correction for operation assistance. Both schemes operated under the same tank structure, equipment foundation, and influent conditions, with the control objective of ensuring stable compliance of effluent ammonia nitrogen and total nitrogen.

[0185] To demonstrate the impact of influent fluctuations on operational support effects, stable operating conditions, intraday peak-valley conditions, and rainfall-induced disturbance conditions were selected for comparison during the validation period. Under stable operating conditions, the influent flow rate fluctuation range did not exceed ±5% of the design average flow rate; under intraday peak-valley conditions, the influent flow rate fluctuation range reached ±25% of the design average flow rate; under rainfall-induced disturbance conditions, the influent flow rate significantly increased within a short period, reaching a maximum flow rate of 1.45 times the design average flow rate, while the influent chemical oxygen demand (COD) concentration showed a significant dilution and decrease. Under each operating condition, both schemes were operated continuously for 72 hours, and data such as influent flow rate, influent water quality, effluent water quality, chemical dosage, air supply, blower energy consumption, model predictions, and control response time were recorded during operation.

[0186] Appendix Figure 12 The curves showing the changes in ammonia nitrogen and total nitrogen in the effluent over time for the experimental and control groups during the validation period are presented. (See attached image.) Figure 12 It is evident that under stable operating conditions, both schemes were able to maintain effluent compliance. However, under intraday peak and valley conditions and rainfall disturbance conditions, the control group, which still calculated the arrival time of pollutants at the process location using a fixed time delay method, experienced some lead-up or lag in its dosing and aeration actions relative to the actual operating conditions, resulting in larger fluctuations in the effluent curve. The experimental group, on the other hand, dynamically updated the volume displacement matrix based on the cumulative influent volume, making its control actions more consistent with the actual spatial location of the pollutants. Consequently, the fluctuations in effluent ammonia nitrogen and total nitrogen were significantly reduced, and the recovery was faster after disturbances occurred.

[0187] See attached document Figure 13 and attached Figure 14 To evaluate the effectiveness of the two operating schemes, this embodiment selects the average value of effluent ammonia nitrogen, the average value of effluent total nitrogen, the standard deviation of effluent ammonia nitrogen, the standard deviation of effluent total nitrogen, the carbon source unit treatment capacity consumption, the aeration unit treatment capacity energy consumption, the average absolute error between the predicted effluent and the measured effluent, and the control response lag time after the disturbance occurs as evaluation indicators.

[0188] The comparative results show that, under the conditions of this embodiment, after adopting the scheme of the present invention, the average ammonia nitrogen concentration in the effluent decreased from 1.85 mg / L to 1.32 mg / L, and the average total nitrogen concentration decreased from 11.8 mg / L to 9.6 mg / L; the standard deviation of ammonia nitrogen decreased from 0.62 mg / L to 0.34 mg / L, and the standard deviation of total nitrogen decreased from 1.95 mg / L to 1.08 mg / L. These results demonstrate that the scheme of the present invention can reduce the average concentration level of the effluent under dynamic influent conditions and significantly reduce effluent fluctuations.

[0189] Regarding operational consumption, after adopting the solution of this invention, carbon source consumption is reduced from 118 kg / 10 4 m 3 Reduced to 103kg / 10 4 m3 Aeration energy consumption was reduced from 468 kWh / 10 4 m 3 Reduced to 421kWh / 10 4 m 3 This demonstrates that the present invention can not only improve the stability of effluent, but also reduce carbon source addition and aeration energy consumption while ensuring that the effluent meets the standards.

[0190] Regarding model prediction accuracy and response speed, after adopting the scheme of this invention, the average absolute error of effluent ammonia nitrogen prediction decreased from 0.71 mg / L to 0.29 mg / L, the average absolute error of effluent total nitrogen prediction decreased from 1.86 mg / L to 0.94 mg / L, and the control response lag time after disturbance was shortened from 26 min to 8 min. (Appendix) Figure 13 The differences between the experimental group and the control group on multiple evaluation indicators are shown in bar chart format. Figure 14 The fitted relationship between the predicted effluent concentration and the measured effluent concentration is shown. Figure 14 It is evident that the fitting data points of the experimental group are closer to the reference fitting relationship overall, indicating that the volume displacement matrix model constructed in this invention can more accurately reflect the migration and transformation process of pollutants in the pool under real working conditions.

[0191] The main reasons why the present invention can achieve the above-mentioned effects are as follows: First, the present invention uses the cumulative influent volume instead of fixed time as the main driving force of the spatial tracking model, which fundamentally reduces the spatial position prediction deviation caused by the dynamic fluctuation of influent flow rate; Second, the present invention introduces a backflow re-injection mechanism, which can characterize the cross-space redistribution process of pollutants caused by process backflow; Third, the present invention introduces a back-mixing correction mechanism based on actual air supply, which can reflect the local axial mixing effect caused by aeration disturbance; Fourth, the present invention adopts a control structure that combines feedforward control and feedback correction, which can not only respond quickly to instantaneous disturbances, but also slowly correct the model constants through feedback, thereby improving the adaptive capability during long-term operation.

[0192] Therefore, the above comparative verification results show that the wastewater treatment plant operation support system based on dynamic influent flow and real-time water quality monitoring described in this invention has good feasibility, operational stability and engineering application value under dynamic influent and complex process flow conditions.

[0193] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A wastewater treatment plant operation support system based on dynamic monitoring of influent flow and real-time water quality, characterized in that, include: The sensing and execution layer, located on-site, includes monitoring and execution units, used to collect water quality and quantity data and perform actions. The control layer, which is communicatively connected to the perception and execution layer, includes: The data acquisition module is used to receive instantaneous influent flow rate, influent characteristic data, instantaneous return flow rate, actual gas supply flow rate, and actual effluent concentration data. The matrix construction module is used to divide the effective physical volume of the biochemical reactor into a set number of equal-volume fluid micro-elements to establish a volume displacement matrix, and to map the physical spatial positions of the field actuators and pipelines to the constant index of the volume displacement matrix. The volume driving module is used to integrate the instantaneous influent flow rate to obtain the cumulative influent volume. When the cumulative influent volume reaches the set value of the volume of a single equal-volume fluid micro-element, a reference shift command is triggered to extend the nodes in the volume displacement matrix and assign the influent characteristic data to the nodes released at the head. The flow correction module is used to perform data re-injection between the constant indices of the corresponding return pipeline according to the instantaneous return flow rate, calculate the dynamic back-mixing coefficient according to the actual gas supply flow rate, and perform lateral smoothing operation on the continuous nodes of the corresponding aerobic zone. The feedforward execution module is used to calculate the material requirements by combining the stoichiometric constant matrix when the data node is shifted to the corresponding constant index, and to generate a feedforward control command to be sent to the execution unit. The feedback adjustment module is used to spatiotemporally align the actual effluent concentration data with the predicted concentration at the end node of the volume displacement matrix and generate a residual sequence by subtraction, so as to output correction parameters to update the stoichiometric constant matrix.

2. The wastewater treatment plant operation support system based on dynamic influent flow and real-time water quality monitoring according to claim 1, characterized in that, The volume driving module is specifically used for: The integer part of the ratio of the current cumulative influent volume to the set volume value is used as the number of shifts n to be performed in the current cycle; By continuously triggering the reference shift command n times, the data of each node in the volume displacement matrix is ​​extended by n steps in the direction of index increment, and the average water inflow characteristic data in the corresponding integration interval is sequentially assigned to the n nodes released from the head of the volume displacement matrix. The margin of the current cumulative influent volume relative to the set volume value is retained as the initial integral value for the next control cycle.

3. The wastewater treatment plant operation support system based on dynamic influent flow and real-time water quality monitoring according to claim 1, characterized in that, The process of the flow correction module performing the data re-injection specifically includes: The instantaneous return flow rate is integrated over time to obtain the cumulative return volume, and the cumulative return volume is divided by the volume set value to calculate the dimensionless dynamic return ratio constant. The data vector at the reflux extraction index in the volume displacement matrix is ​​extracted, and combined with the dynamic reflux ratio constant, the feature concentration vector extending to the reflux access index is weighted and summed to cover it.

4. The wastewater treatment plant operation support system based on dynamic influent flow and real-time water quality monitoring according to claim 1, characterized in that, The process by which the flow correction module performs lateral smoothing calculations specifically includes: The actual air supply flow rate of each air supply branch in the aerobic zone is converted into the dynamic back-mixing coefficient using a preset nonlinear mapping function; For the target node in the volume displacement matrix that is within the aerobic zone span, the adjacent three-point weighted algorithm is invoked, and the original basic feature concentration vector of the target node and the basic feature concentration vectors of the upstream and downstream nodes adjacent to the target node are weighted and smoothed using the dynamic back-mixing coefficient.

5. The wastewater treatment plant operation support system based on dynamic influent flow and real-time water quality monitoring according to claim 1, characterized in that, The specific process by which the feedforward execution module generates and sends the dosing control command to the execution unit includes: When the reference shift command is triggered, the matrix data vector is extracted and shifted to the constant index corresponding to the drug delivery point in the hypoxia zone. The absolute carbon source mass is calculated by using the instantaneous predicted concentration in the matrix data vector, combined with the theoretical carbon-nitrogen consumption ratio constant called in the stoichiometric constant matrix and the volume set value. The absolute carbon source mass is converted into a corresponding pulse frequency and sent to the execution unit.

6. The wastewater treatment plant operation support system based on dynamic influent flow and real-time water quality monitoring according to claim 1, characterized in that, The specific process by which the feedforward execution module generates and sends the aeration control command to the execution unit includes: Extract the feature concentration vectors of all continuous fluid micro-elements within the continuous index interval of the aerobic zone in the volume displacement matrix; By combining the theoretical oxygen consumption equivalent coefficients called from the stoichiometric constant matrix, the predicted concentrations of all nodes within the continuous index interval are calculated by interval integration to obtain the theoretical total oxygen demand. Combining the standard oxygen transfer efficiency of the aeration equipment with the current water temperature parameters, the theoretical total oxygen demand is converted into an actual target air flow rate signal and sent to the execution unit.

7. The wastewater treatment plant operation support system based on dynamic influent flow and real-time water quality monitoring according to claim 1, characterized in that, The process by which the feedback adjustment module generates the residual sequence specifically includes: A data buffer queue is created in memory to temporarily store the predicted concentration output by the end node of the volume displacement matrix with a delay time set according to the instrument response delay time based on the sampling and measurement cycle of the on-site online analyzer. The actual effluent concentration data acquired in real time is subtracted from the theoretically predicted concentration output after the delay and temporary storage to generate the residual sequence; when the actual effluent concentration data exceeds the set range, the residual value of the current period is forced to be zero.

8. The wastewater treatment plant operation support system based on dynamic influent flow and real-time water quality monitoring according to claim 1, characterized in that, The process by which the feedback adjustment module updates the chemosimetric constant matrix specifically includes: The residual sequence is input into the discrete proportional-integral-derivative controller, and the dynamic compensation amount of the model constant at the current time is calculated by combining the enabled anti-integral saturation logic. The dynamic compensation amount of the model constant is superimposed on the basic stoichiometric constant called in the previous calculation cycle, and after setting a hard protection boundary to clamp and limit its amplitude within a physical threshold range, the current execution constant is generated for the feedforward execution module to call.

9. The wastewater treatment plant operation support system based on dynamic influent flow and real-time water quality monitoring according to claim 1, characterized in that, The perception execution layer specifically includes: The influent monitoring unit includes an electromagnetic flow meter and an online water quality analyzer; The effluent monitoring unit includes an ammonia nitrogen sensor and a nitrate nitrogen sensor or a total nitrogen analyzer; The backflow monitoring unit includes a flow meter device; The dosing unit includes a variable frequency dosing pump; The aeration unit includes a variable frequency blower, an air regulating valve, and an air flow detection device.

10. The wastewater treatment plant operation support system based on dynamic influent flow and real-time water quality monitoring according to claim 1, characterized in that, The data acquisition module is specifically used for: The original register values ​​output by the sensing and execution layer are read from the bus network according to the preset sampling period, and the analog signal is converted into a physical engineering quantity value by linear engineering quantity conversion. A unified system timestamp is assigned to each set of converted physical engineering quantity values, and a data queue of fixed length is maintained on the time axis. A moving average filtering algorithm is used to extract the baseline engineering quantity to eliminate transient high-frequency noise.

Images