Multi-parameter coordinated energy-saving aeration control system for municipal sewage plant

Abstract

The application discloses a kind of municipal sewage plant multi-parameter synergic energy-saving aeration control systems, specifically relates to system energy-saving optimization control technical field, for solving the energy efficiency implicit degradation problem caused by the online energy efficiency index interference of existing industrial control system due to self-regulation action;Aeration energy efficiency sequence, blower frequency sequence and valve opening sequence are obtained by sequence acquisition module, and steady-state interval is determined by interval determination module according to the direction of control action and energy efficiency change Different proportion determines steady-state interval, and the preliminary lag relationship is identified by conditional entropy reduction calculation by relationship identification module in steady-state interval, and the forward lag relationship is obtained by excluding reverse causal components by relationship extraction module, and the inherent energy efficiency reference of process is obtained by reference reconstruction module from the residual component in current energy efficiency value by stripping, and the optimization direction of multi-parameter synergic controller is corrected according to the deviation of inherent energy efficiency reference of process and current energy efficiency value by direction correction module.

Description

Technical Field

[0001] This invention relates to the field of system energy-saving optimization control technology, and more specifically, to a multi-parameter collaborative energy-saving aeration control system for municipal wastewater treatment plants. Background Technology

[0002] Existing municipal wastewater treatment plants generally employ multi-parameter collaborative control schemes based on industrial control systems in their aerobic aeration stages. These industrial control systems collect process variables such as dissolved oxygen, ammonia nitrogen, and water temperature online, dynamically calculate the aeration rate using a pre-set optimization model, and send adjustment commands to blowers and valves via a control network. The aim is to reduce aeration energy consumption while meeting effluent quality requirements. Some industrial control systems also incorporate real-time energy efficiency indicators as feedback, gradually adjusting the control strategy to bring the system closer to a preset energy-saving operating point.

[0003] However, in the actual operation of such industrial control systems, the online energy efficiency indicators on which collaborative optimization relies already include the cumulative effects of the existing regulation of the control system. This makes it impossible for the indicators to objectively reflect the true energy consumption benchmark of the process. When the control strategy continuously seeks optimization based on feedback information that is disturbed by its own actions, it is easy to form a self-misleading control cycle, causing the system to gradually deviate from the true optimal operating condition. Ultimately, this manifests as a hidden increase in aeration energy consumption during long-term operation, and such energy efficiency degradation is difficult to detect and correct in a timely manner through conventional monitoring methods. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, the present invention provides a multi-parameter collaborative energy-saving aeration control system for municipal wastewater treatment plants to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A multi-parameter coordinated energy-saving aeration control system for municipal wastewater treatment plants includes: The sequence acquisition module is used to acquire the aeration efficiency sequence, blower frequency sequence, and valve opening sequence of the industrial control system at the current moment and within the previous preset time period. The interval determination module is used to compare the direction of adjustment action in the blower frequency sequence and valve opening sequence with the direction of change of the aeration energy efficiency sequence in the corresponding time period. When the proportion of different directions exceeds the preset proportion threshold, the aeration energy efficiency sequence is determined to be in the steady state interval. The relationship identification module is used to symbolize the adjustment direction of the blower frequency sequence and valve opening sequence, as well as the change direction of the aeration energy efficiency sequence, within the steady-state range. It calculates the conditional entropy reduction from the adjustment direction symbol sequence to the change direction symbol sequence under different time delays, and uses the time delay of the maximum conditional entropy reduction as the preliminary lag relationship between historical control actions and energy efficiency response. The relationship extraction module is used to perform causal direction testing on the preliminary lag relationship, remove the reverse causal components in which the energy efficiency response leads the control action, and obtain the positive lag relationship; The baseline reconstruction module is used to estimate the residual component caused by historical control actions in the current energy efficiency value using the positive hysteresis relationship, and to subtract the residual component from the current energy efficiency value to obtain the process-inherent energy efficiency baseline. The direction correction module is used to correct the optimization direction of the multi-parameter collaborative controller by utilizing the deviation between the inherent energy efficiency baseline of the process and the current energy efficiency value.

[0006] Furthermore, the aeration efficiency sequence, blower frequency sequence, and valve opening sequence of the industrial control system at the current moment and within the previously preset time period are obtained, including: Load the original sequence of aeration efficiency, the original sequence of blower frequency, and the original sequence of valve opening for the preset time period before the current moment into the historical database of the industrial control system. Align the timestamps in the original sequences of aeration energy efficiency, blower frequency, and valve opening, and remove records with no corresponding timestamp values ​​to obtain aeration energy efficiency aligned sequences, blower frequency aligned sequences, and valve opening aligned sequences with a unified time base. From the aeration efficiency alignment sequence, blower frequency alignment sequence, and valve opening alignment sequence with a unified time base, segments that trace back a preset time from the current moment are extracted as the aeration efficiency sequence, blower frequency sequence, and valve opening sequence.

[0007] Furthermore, by comparing the direction of the adjustment actions in the blower frequency sequence and valve opening sequence with the direction of change in the aeration efficiency sequence within the corresponding time period, when the proportion of dissimilar directions exceeds a preset threshold, the aeration efficiency sequence is determined to be in a steady-state range, including: By comparing the frequency values ​​of two adjacent moments in the blower frequency sequence point by point, the moment when the frequency value increases is marked as the positive adjustment action point, and the moment when the frequency value decreases is marked as the negative adjustment action point; By comparing the opening values ​​of two adjacent moments in the valve opening sequence point by point, the moment when the opening value increases is marked as the positive adjustment action point, and the moment when the opening value decreases is marked as the negative adjustment action point; The adjustment direction marks at the same moment in the blower frequency sequence and valve opening sequence are merged. If both are positive adjustment action points at the same moment, they are merged into a positive adjustment action point. If both are negative adjustment action points at the same moment, they are merged into a negative adjustment action point. If the two directions are inconsistent at the same moment, they are not recorded as adjustment action points, thus obtaining the adjustment action direction sequence. By comparing the energy efficiency values ​​of two adjacent moments in the aeration energy efficiency sequence point by point, the moments when the energy efficiency value increases are marked as positive change points, and the moments when the energy efficiency value decreases are marked as negative change points, thus obtaining the energy efficiency change direction sequence; Align the sequence of adjustment action directions with the sequence of energy efficiency change directions by timestamp, and count the proportion of adjustment action points whose directions differ from the corresponding energy efficiency change points to the total number of adjustment action points. When the proportion of opposite directions exceeds the preset proportion threshold, the aeration efficiency sequence is determined to be in the steady state range.

[0008] Furthermore, within the steady-state range, the adjustment directions of the blower frequency sequence and valve opening sequence, as well as the change direction of the aeration energy efficiency sequence, are symbolized respectively. The conditional entropy reduction from the adjustment direction symbol sequence to the change direction symbol sequence under different time delays is calculated. The time delay of the maximum conditional entropy reduction is taken as the preliminary lag relationship between historical control actions and energy efficiency responses, including: In the sequence of adjustment action directions, positive adjustment action points are denoted as positive symbols, negative adjustment action points are denoted as negative symbols, and moments with no adjustment action points are denoted as resting symbols, thus obtaining the sequence of adjustment direction symbols; In the energy efficiency change direction sequence, positive change points are denoted as positive signs, and negative change points are denoted as negative signs, thus obtaining the change direction sign sequence; The direction symbol sequence is shifted according to different time delays, and the conditional entropy of the change direction symbol sequence under each shift amount is calculated when the direction symbol sequence is adjusted after a known shift. Calculate the unconditional entropy of the change direction symbol sequence when the adjustment direction symbol sequence is unknown, and subtract the conditional entropy of each translation amount from the unconditional entropy to obtain the conditional entropy reduction for each translation amount. The shift that results in the largest reduction in conditional entropy is determined as the initial lag relationship between historical control actions and energy efficiency response.

[0009] Furthermore, a causal direction test is performed on the initial lag relationship, eliminating the reverse causal component where the energy efficiency response precedes the control action, resulting in a positive lag relationship, including: The adjustment direction symbol sequence and the change direction symbol sequence are time-aligned according to the initial lag relationship to obtain the aligned adjustment-change symbol pair sequence; Check the proportion of the change direction sign leading the adjustment direction sign in the aligned adjustment-change sign pair sequence; When the proportion by which the change direction sign leads the adjustment direction sign exceeds a preset proportion, it is determined that the preliminary lag relationship contains a reverse causal component, and the preliminary lag relationship is marked as invalid. When the proportion by which the direction of change precedes the direction of adjustment does not exceed the preset proportion, the initial lag relationship is determined to be a positive lag relationship.

[0010] Furthermore, the proportion in which the change direction sign leads the adjustment direction sign in the aligned sequence of adjustment-change sign pairs is checked pairwise, including: Examine each symbol pair in the aligned adjustment-change symbol pair sequence, and determine the symbol pair whose timestamp of the adjustment direction symbol is later than the timestamp of the change direction symbol as the symbol pair whose change direction symbol is leading; Divide the number of symbol pairs that lead the change direction symbol by the total number of symbol pairs in the aligned adjustment-change symbol pair sequence to obtain the ratio of the change direction symbol leading the adjustment direction symbol.

[0011] Furthermore, the residual component caused by historical control actions in the current energy efficiency value is estimated using the positive hysteresis relationship, and the residual component is subtracted from the current energy efficiency value to obtain the process-specific energy efficiency benchmark, including: Determine the starting point at which historical control actions affect the current energy efficiency value based on the positive lag relationship; Extract the amplitude of each adjustment action from the blower frequency sequence and valve opening sequence within the interval from the start time to the current time. The amplitude of each adjustment action is attenuated according to a preset attenuation coefficient to obtain the sub-residual component of each adjustment action on the current energy efficiency value. The residual components caused by historical control actions are obtained by summing up the sub-residual components of all adjustment actions within the interval from the start time to the current time. The residual component is subtracted from the current energy efficiency value to obtain the inherent energy efficiency benchmark of the process.

[0012] Furthermore, the amplitude of each adjustment action is extracted from the blower frequency sequence and valve opening sequence within the interval from the start time to the current time, including: Calculate the difference between frequency values ​​at adjacent moments in the blower frequency sequence point by point, and record the non-zero difference as the action amplitude of the blower frequency adjustment action; Calculate the difference between the opening values ​​at adjacent moments in the valve opening sequence point by point, and record the non-zero difference as the action amplitude of the valve opening adjustment action; The amplitude values ​​of the blower frequency adjustment action and the valve opening adjustment action are uniformly included in the adjustment action amplitude set.

[0013] Furthermore, by utilizing the deviation between the inherent energy efficiency baseline of the process and the current energy efficiency value, the optimization direction of the multi-parameter collaborative controller is corrected, including: The self-interference bias is obtained by calculating the difference between the inherent energy efficiency baseline of the process and the current energy efficiency value. Determine whether the self-interference deviation exceeds the preset deviation range; When the self-interference deviation exceeds the preset deviation range, obtain the current original optimization direction of the multi-parameter collaborative controller; The directional component of the self-interference deviation along the original optimization direction is used as the correction amount. The correction amount is subtracted from the original optimization direction to obtain the reconstructed optimization direction. The refactoring optimization direction is used as the optimization direction for the corrected multi-parameter collaborative controller.

[0014] Furthermore, the directional component of the self-interference deviation along the original optimization direction is used as a correction factor, including: The self-interference bias is multiplied by the unit vector of the original optimization direction to obtain the projection scalar of the self-interference bias in the original optimization direction. The correction is obtained by multiplying the projected scalar by the unit vector of the original optimization direction.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. Through progressive collaboration of modules, an energy efficiency self-interference stripping mechanism was constructed, operating independently within the industrial control system. This mechanism utilizes the anisotropic characteristics of control actions and energy efficiency changes within the steady-state range to lock in the system's stable operating window. Within this window, the initial lag relationship between control and energy efficiency is identified by the time delay of maximum conditional entropy reduction. After causal direction verification to eliminate inverse components, a positive lag relationship is obtained. Based on this positive lag relationship, the benchmark reconstruction module subtracts the accumulated residual attenuation of historical control actions from the current energy efficiency value, restoring the inherent energy efficiency benchmark of the process, unaffected by the self-regulating behavior of the control system. All of the above modules are embedded in the data acquisition and command generation chain of the industrial control system. Without changing the original control network architecture or adding external sensors, the industrial control system now possesses the ability to distinguish between the actual energy consumption level of the process and control self-interference components for the first time.

[0016] 2. The direction correction module uses the deviation between the inherent energy efficiency baseline of the process and the current energy efficiency value as the correction basis. It subtracts the component of the deviation along the original optimization direction of the multi-parameter collaborative controller from the original optimization direction to obtain the reconstructed optimization direction. This ensures that the optimization direction of the industrial control system is no longer constrained by the reflective influence of historical control actions. It avoids the online energy efficiency index from gradually guiding the controller to a control cycle that deviates from the true optimal operating condition due to the presence of self-adjustment traces. This suppresses the implicit drift of aeration energy consumption during long-term operation. The entire process uses the time-series data inside the industrial control system as the processing object and symbolic computation and conditional entropy reduction analysis as processing methods. The computational load is matched with the existing hardware resources of the industrial control system, making it practically deployable in engineering. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the structure of a multi-parameter collaborative energy-saving aeration control system for municipal wastewater treatment plants according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example: Figure 1 A schematic diagram of a multi-parameter synergistic energy-saving aeration control system for a municipal wastewater treatment plant is provided. The multi-parameter synergistic energy-saving aeration control system for a municipal wastewater treatment plant includes: The sequence acquisition module is used to acquire the aeration efficiency sequence, blower frequency sequence, and valve opening sequence of the industrial control system at the current moment and within the previous preset time period. The interval determination module is used to compare the direction of adjustment action in the blower frequency sequence and valve opening sequence with the direction of change of the aeration energy efficiency sequence in the corresponding time period. When the proportion of different directions exceeds the preset proportion threshold, the aeration energy efficiency sequence is determined to be in the steady state interval. The relationship identification module is used to symbolize the adjustment direction of the blower frequency sequence and valve opening sequence, as well as the change direction of the aeration energy efficiency sequence, within the steady-state range. It calculates the conditional entropy reduction from the adjustment direction symbol sequence to the change direction symbol sequence under different time delays, and uses the time delay of the maximum conditional entropy reduction as the preliminary lag relationship between historical control actions and energy efficiency response. The relationship extraction module is used to perform causal direction testing on the preliminary lag relationship, remove the reverse causal components in which the energy efficiency response leads the control action, and obtain the positive lag relationship; The baseline reconstruction module is used to estimate the residual component caused by historical control actions in the current energy efficiency value using the positive hysteresis relationship, and to subtract the residual component from the current energy efficiency value to obtain the process-inherent energy efficiency baseline. The direction correction module is used to correct the optimization direction of the multi-parameter collaborative controller by utilizing the deviation between the inherent energy efficiency baseline of the process and the current energy efficiency value.

[0020] In a specific implementation of the sequence acquisition module, the original sequence of aeration efficiency, the original sequence of blower frequency, and the original sequence of valve opening within a preset time period before the current moment are loaded into the historical database of the industrial control system.

[0021] The industrial control system's historical database is a time-series data storage unit deployed in the plant-level monitoring network. It records the timestamps and values ​​of feedback signals from various online instruments and frequency converters at a fixed sampling period. Each data point in the aeration energy efficiency raw sequence represents the aeration energy consumption for removing a unit mass of pollutants in the aerobic tank at the corresponding time. Each data point in the aeration energy efficiency raw sequence is calculated in real time by the industrial control system based on the blower power metering signal and the online detection value of effluent pollutant concentration and stored in the historical database. The data points in the blower frequency raw sequence are taken from the output value of the frequency feedback register of the blower frequency converter at the current time. The data points in the valve opening raw sequence are taken from the valve position percentage signal of the valve position feedback transmitter of the aeration branch electric valve at the current time.

[0022] The preset time period is a fixed time window that traces backward from the current time. The duration of the preset time period is determined based on the magnitude of the hydraulic retention time in the aerobic tank and the nitrifying bacteria growth rate in the activated sludge system. Both the hydraulic retention time in the aerobic tank and the nitrifying bacteria growth rate are obtained from the process design parameters stored in the industrial control system. Specifically, the lower limit of the preset time period is set as an integer multiple of the hydraulic retention time in the aerobic tank and an integer multiple of the reciprocal of the nitrifying bacteria growth rate. Several integer multiples of this lower limit are then used as the final value of the preset time period. For example, if the hydraulic retention time in the aerobic tank is 6 hours and the reciprocal of the nitrifying bacteria growth rate is approximately 24 hours, the preset time period can be set to, for example, 24 hours, 48 ​​hours, or 72 hours. Setting the preset time period ensures that the sequence contains sufficient coupling information between control actions and energy efficiency responses, while avoiding the introduction of slow-drift interference on the sludge age timescale due to an excessively long sequence.

[0023] Align the timestamps in the original sequences of aeration energy efficiency, blower frequency, and valve opening, and remove records with no corresponding timestamp values ​​to obtain aeration energy efficiency aligned sequences, blower frequency aligned sequences, and valve opening aligned sequences with a unified time base.

[0024] The specific process of timestamp alignment is as follows: using the timestamp of the original blower frequency sequence as the reference time grid, nearest neighbor matching is performed on the original aeration efficiency sequence and the original valve opening sequence. When there is no corresponding sampled value in the neighboring interval of the original aeration efficiency sequence or the original valve opening sequence at a certain reference timestamp, it is determined that there is a record with no corresponding value at that reference timestamp, and the record at the reference timestamp determined to have no corresponding value is deleted as a whole. The length of the neighboring interval is set according to the sampling period of the industrial control system, and the length of the neighboring interval is taken as 1 / 2 of the sampling period of the industrial control system to ensure that the maximum deviation between the nearest neighbor matching result and the matched timestamp does not exceed half a sampling period. After the deletion operation, each remaining record has a valid value in the three dimensions of aeration efficiency, blower frequency, and valve opening, and shares the same timestamp, forming a unified time base.

[0025] From the aeration efficiency alignment sequence, blower frequency alignment sequence, and valve opening alignment sequence with a unified time base, a segment of a preset time duration is extracted from the current moment and used as the aeration efficiency sequence, blower frequency sequence, and valve opening sequence.

[0026] The truncation method is as follows: taking the current time as the truncation endpoint and the time obtained by subtracting a preset duration from the current time as the truncation start point, all records whose timestamps fall within the closed interval from the truncation start point to the truncation endpoint are retained in the aeration efficiency alignment sequence, blower frequency alignment sequence, and valve opening alignment sequence with a unified time base, forming the aeration efficiency sequence, blower frequency sequence, and valve opening sequence. The i-th element in the aeration efficiency sequence, blower frequency sequence, and valve opening sequence shares the same timestamp, and the lengths of the aeration efficiency sequence, blower frequency sequence, and valve opening sequence are the same.

[0027] The aeration efficiency sequence, blower frequency sequence, and valve opening sequence obtained in this step provide strictly time-synchronized foundational data for subsequent interval determination, relationship identification, and benchmark reconstruction. Compared with existing technologies that do not perform timestamp alignment and unified time base processing on the original sequences, this step eliminates the time misalignment caused by the delay differences in each signal acquisition link from the source. This avoids the introduction of false hysteresis estimation errors in subsequent conditional entropy reduction calculations due to time misalignment, thereby ensuring that the preliminary hysteresis relationship and positive hysteresis relationship identified from the symbol sequence truly reflect the physical transmission delay between control actions and energy efficiency responses, rather than artifacts introduced by the measurement system.

[0028] In a specific implementation of the interval determination module, the frequency values ​​of two adjacent moments in the blower frequency sequence are compared point by point. The moment when the frequency value increases is marked as the positive adjustment action point, and the moment when the frequency value decreases is marked as the negative adjustment action point.

[0029] Each frequency value in the blower frequency sequence corresponds to a timestamp. During point-by-point comparison, starting from the second timestamp of the blower frequency sequence, the frequency value corresponding to the current timestamp is subtracted from the frequency value corresponding to the previous timestamp to obtain the frequency difference. When the frequency difference is greater than 0, the current timestamp is marked as a positive adjustment action point. When the frequency difference is less than 0, the current timestamp is marked as a negative adjustment action point. When the frequency difference is equal to 0, no marking is performed.

[0030] By comparing the opening values ​​of two adjacent moments in the valve opening sequence point by point, the moment when the opening value increases is marked as the positive adjustment action point, and the moment when the opening value decreases is marked as the negative adjustment action point.

[0031] Each valve opening value in the sequence corresponds to a timestamp. During point-by-point comparison, starting from the second timestamp in the sequence, the opening value corresponding to the current timestamp is subtracted from the opening value corresponding to the previous timestamp to obtain the opening difference. When the opening difference is greater than 0, the current timestamp is marked as a positive adjustment action point. When the opening difference is less than 0, the current timestamp is marked as a negative adjustment action point. When the opening difference is equal to 0, no marking is performed.

[0032] The adjustment direction markers at the same moment in the blower frequency sequence and valve opening sequence are merged. If both are positive adjustment action points at the same moment, they are merged into a positive adjustment action point. If both are negative adjustment action points at the same moment, they are merged into a negative adjustment action point. If the two directions are inconsistent at the same moment, they are not recorded as adjustment action points, thus obtaining the adjustment action direction sequence.

[0033] The merging process uses the common moment of the blower frequency sequence and the valve opening sequence as the merging node. At the same merging node, if both the blower frequency sequence and the valve opening sequence are marked as positive adjustment action points, a positive adjustment action point is recorded at that merging node. If both the blower frequency sequence and the valve opening sequence are marked as negative adjustment action points, a negative adjustment action point is recorded at that merging node. If the blower frequency sequence is marked as a positive adjustment action point and the valve opening sequence is marked as a negative adjustment action point, it indicates that the blower and valve are in a counter-regulation state at that merging node. Action data in a counter-regulation state is not conducive to identifying the unidirectional coupling relationship between control and energy efficiency, and therefore, no adjustment action point is recorded at that merging node. Similarly, if the blower frequency sequence is marked as a negative adjustment action point and the valve opening sequence is marked as a positive adjustment action point, no adjustment action point is recorded at that merging node. If neither the blower frequency sequence nor the valve opening sequence is marked at the merging node, no adjustment action point is recorded at that merging node. After processing all merging nodes, the recorded positive and negative adjustment action points are arranged in timestamp order to obtain the adjustment action direction sequence.

[0034] By comparing the energy efficiency values ​​of two adjacent moments in the aeration energy efficiency sequence point by point, the moments when the energy efficiency value increases are marked as positive change points, and the moments when the energy efficiency value decreases are marked as negative change points, thus obtaining the energy efficiency change direction sequence.

[0035] Each energy efficiency value in the aeration energy efficiency sequence corresponds to a timestamp. During point-by-point comparison, starting from the second timestamp of the aeration energy efficiency sequence, the energy efficiency value corresponding to the current timestamp is subtracted from the energy efficiency value corresponding to the previous timestamp to obtain the energy efficiency difference. When the energy efficiency difference is greater than 0, the current timestamp is marked as a positive change point. When the energy efficiency difference is less than 0, the current timestamp is marked as a negative change point. When the energy efficiency difference is equal to 0, no marking is performed. After completing the point-by-point comparison, the marked positive and negative change points are arranged in timestamp order to obtain the energy efficiency change direction sequence.

[0036] Align the sequence of adjustment action directions with the sequence of energy efficiency change directions by timestamp, and count the proportion of adjustment action points whose directions differ from the corresponding energy efficiency change points to the total number of adjustment action points.

[0037] When aligning by timestamp, for each timestamp of a regulation action point in the regulation action direction sequence, find the energy efficiency change point with the same timestamp in the energy efficiency change direction sequence. For each pair of regulation action points and energy efficiency change points with the same timestamp, compare their directions: if the regulation action point is positive and the energy efficiency change point is negative, the directions are determined to be different; if the regulation action point is negative and the energy efficiency change point is positive, the directions are determined to be different; if both the regulation action point and the energy efficiency change point are positive or both are negative, the directions are determined to be the same. Count the number of regulation action point and energy efficiency change point pairs with different directions, and divide the number of such pairs by the total number of regulation action points in the regulation action direction sequence to obtain the proportion of pairs with different directions.

[0038] When the proportion of opposite directions exceeds the preset proportion threshold, the aeration efficiency sequence is determined to be in the steady state range.

[0039] The preset proportional threshold is set based on the control behavior characteristics of the industrial control system under steady-state conditions. Under steady-state conditions, the adjustment actions issued by the multi-parameter collaborative controller are mainly small-scale corrections. The direction of the adjustment action is mostly inversely related to the direction of energy efficiency change; that is, increasing aeration is to counteract the downward trend of energy efficiency, and decreasing aeration is to suppress the upward trend of energy efficiency. Therefore, the proportions of opposite directions within the steady-state range are close to a relatively high value. The specific method for setting the preset proportional threshold is as follows: Select manually labeled steady-state range samples from the historical database of the industrial control system. The basis for manual labeling is that the dissolved oxygen fluctuation within the steady-state range does not exceed the upper or lower limit of the target value, and the effluent ammonia nitrogen concentration is continuously stable within the compliance line. Calculate the lowest value of the proportions of opposite directions within each steady-state range sample, and multiply the lowest value by a preset coefficient to obtain the preset proportional threshold. The preset coefficient ranges from 0.7 to 0.9. For example, if the lowest proportion of opposite directions in the manually labeled steady-state range samples is 0.85, and the preset coefficient is 0.8, then the preset proportional threshold is set to 0.68. When the calculated proportion of dissimilar directions exceeds a preset threshold, the aeration efficiency sequence is determined to be in a steady-state range. A dissimilar proportion exceeding the preset threshold means that the direction of the control action and the direction of energy efficiency change have a stable inverse relationship within the current time period, consistent with the control behavior characteristics of a steady-state range. When the dissimilar proportion does not exceed the preset threshold, it means that the directional relationship between control and energy efficiency is chaotic within the current time period, belonging to a transient range.

[0040] By comparing the directional relationship between control actions and energy efficiency changes to determine the steady-state interval, this study reveals the inherent inverse relationship between the directions of control actions and energy efficiency changes under steady-state conditions. Under transient conditions, the control system performs large adjustments to track changes in operating conditions. Due to response lag and overshoot, the direction of the adjustment action and the direction of energy efficiency change exhibit a chaotic proportional relationship, and the proportion of dissimilar directions cannot be stably maintained at a high level. Using the proportion of dissimilar directions as a criterion, the steady-state interval and transient interval can be distinguished statistically. This provides a reliable data foundation for subsequently identifying the initial lag relationship only within the steady-state interval, avoiding spurious lag estimations introduced by the non-stationary coupling of control and energy efficiency within the transient interval.

[0041] In a specific implementation of the relationship identification module, positive adjustment action points in the adjustment action direction sequence are recorded as positive symbols, negative adjustment action points are recorded as negative symbols, and moments without adjustment action points are recorded as resting symbols, thus obtaining an adjustment direction symbol sequence.

[0042] The adjustment action direction sequence is generated by the interval determination step and includes positive adjustment action points, negative adjustment action points, and times when no adjustment action point is recorded, arranged in timestamp order. The symbolization process traverses each timestamp position in the adjustment action direction sequence: when a positive adjustment action point exists at a timestamp position, a positive symbol is written into the adjustment direction symbol sequence; when a negative adjustment action point exists at a timestamp position, a negative symbol is written into the adjustment direction symbol sequence; when no adjustment action point exists at a timestamp position, a rest symbol is written into the adjustment direction symbol sequence. The positive, negative, and rest symbols each take distinct discrete values; for example, the positive symbol takes +1, the negative symbol takes -1, and the rest symbol takes 0. The length of the adjustment direction symbol sequence is the same as the timestamp length of the adjustment action direction sequence, and each symbol in the adjustment direction symbol sequence corresponds to a timestamp.

[0043] In the energy efficiency change direction sequence, positive change points are denoted as positive signs, and negative change points are denoted as negative signs, thus obtaining the change direction sign sequence.

[0044] The energy efficiency change direction sequence is generated by an interval determination step and contains positive and negative change points arranged in timestamp order. The symbolization process traverses each timestamp position in the energy efficiency change direction sequence: when a positive change point exists at a timestamp position, a positive symbol is written into the change direction symbol sequence; when a negative change point exists at a timestamp position, a negative symbol is written into the change direction symbol sequence. The values ​​of the positive and negative symbols are consistent with the adjustment direction symbol sequence; for example, the positive symbol is +1, and the negative symbol is -1. No resting symbol is set in the change direction symbol sequence. When the energy efficiency difference is equal to 0 in the energy efficiency change direction sequence and is not marked, the symbol value of the previous time is written into the corresponding timestamp in the change direction symbol sequence as a maintenance signal. The length of the change direction symbol sequence is the same as the length of the adjustment direction symbol sequence, and the change direction symbol sequence and the adjustment direction symbol sequence correspond one-to-one in terms of timestamps.

[0045] The direction symbol sequence is shifted according to different time delays, and the conditional entropy of the change direction symbol sequence under each shift amount is calculated when the direction symbol sequence is adjusted after a known shift.

[0046] The time delay is measured by the number of sampling points of sequence offset, and the shift value ranges from 0 to the preset maximum shift. The preset maximum shift is set based on the ratio of the hydraulic retention time in the aerobic tank to the sampling period of the industrial control system. The preset maximum shift is an integer value of the ratio of the hydraulic retention time in the aerobic tank to the sampling period of the industrial control system multiplied by a preset multiple. The hydraulic retention time in the aerobic tank is obtained from the process design parameters stored in the industrial control system, and the sampling period of the industrial control system is read from the configuration parameters in the historical database of the industrial control system. The preset multiple is set based on the ratio of the response time of dissolved oxygen to aeration adjustment in the activated sludge system to the hydraulic retention time in the aerobic tank. The preset multiple is the upper limit of the ratio of dissolved oxygen response time to the hydraulic retention time in the aerobic tank, rounded down to ensure that the maximum shift covers the entire possible delay range from control action to energy efficiency response. For example, if the hydraulic retention time of the aerobic tank is 6 hours, the sampling cycle of the industrial control system is 1 minute, the ratio is 360, and the dissolved oxygen response time is about 15 minutes, then the preset multiple is 15 and the ratio of 360 is rounded to 1. The preset maximum translation amount is 360×2=720 points, with an additional margin of 1 to ensure coverage.

[0047] For each translation, the adjustment direction symbol sequence is shifted in the direction of increasing time by the number of points corresponding to the current translation. Symbols that are shifted out of the range of the adjustment direction symbol sequence are discarded, and the empty positions at the beginning of the adjustment direction symbol sequence are filled with resting symbols. The translated adjustment direction symbol sequence and the change direction symbol sequence are aligned by timestamps to form a symbol pair sequence. The conditional frequency of the change direction symbol value under each adjustment direction symbol value in the symbol pair sequence is statistically analyzed, and the conditional frequency is used as an estimate of the conditional probability. The conditional entropy is calculated as: H(Y|Xshifted) = -∑[p(x)×∑p(y|x)×logp(y|x)]; where H(Y|Xshifted) represents the conditional entropy of the direction-shifting symbol sequence when the direction-shifting symbol sequence is known after translation; p(x) represents the probability of a certain symbol appearing in the direction-shifting symbol sequence after translation, represented by the frequency of that symbol appearing in the direction-shifting symbol sequence after translation; p(y|x) represents the conditional probability of the direction-shifting symbol being y given that the direction-shifting symbol is x, represented by the statistically obtained conditional frequency; and log represents the logarithmic operation to base 2. When a certain conditional frequency is 0, the corresponding conditional entropy term is 0.

[0048] Calculate the unconditional entropy of the change direction symbol sequence when the adjustment direction symbol sequence is unknown, and subtract the conditional entropy for each translation amount from the unconditional entropy to obtain the conditional entropy reduction for each translation amount.

[0049] The unconditional entropy of the direction-changing symbol sequence is calculated as: H(Y) = -∑p(y) × logp(y); where H(Y) represents the unconditional entropy of the direction-changing symbol sequence, and p(y) represents the probability of a certain symbol appearing in the direction-changing symbol sequence, replaced by the frequency of that symbol appearing in the direction-changing symbol sequence. For each shift, the conditional entropy reduction is calculated as: ΔH(t) = H(Y) - H(Y|Xshifted); where ΔH(t) represents the conditional entropy reduction when the shift is t. The conditional entropy reduction characterizes the amount of information transmitted from the shifted direction-changing symbol sequence to the direction-changing symbol sequence. The greater the amount of information, the stronger the explanatory power of the control action on the energy efficiency response when the shift is t.

[0050] The shift that results in the largest reduction in conditional entropy is determined as the initial lag relationship between historical control actions and energy efficiency response.

[0051] After traversing all translation values, the conditional entropy reduction values ​​corresponding to each translation value are compared. The translation value with the maximum conditional entropy reduction is converted into time units by multiplying the translation value with the maximum conditional entropy reduction by the sampling period of the industrial control system, thus obtaining the preliminary lag relationship in time units. The preliminary lag relationship reflects the average propagation delay experienced from the issuance of the control action to the observable change in energy efficiency.

[0052] This step utilizes conditional entropy reduction to identify preliminary lag relationships. Compared to existing methods that use cross-correlation functions or fixed time delay parameters, conditional entropy reduction inherently possesses asymmetry, enabling it to distinguish between the causal path of control-driven energy efficiency and the pseudo-path of energy efficiency changes reacting to control, based on the direction of information transmission. Simultaneously, by introducing resting symbols to mark unregulated moments, interference from noise symbols during control inactivity is avoided in information calculations, resulting in a sharper conditional entropy reduction peak and improving the accuracy and noise resistance of preliminary lag relationship identification.

[0053] In a specific implementation of the relationship extraction module, the adjustment direction symbol sequence and the change direction symbol sequence are time-aligned according to the initial lag relationship to obtain the aligned adjustment-change symbol pair sequence.

[0054] The preliminary lag relation is output from the relation identification step. It is a time-based delay, representing the length of time from the issuance of a control action to the resulting observable change in energy efficiency. During time alignment, the entire adjustment direction symbol sequence is shifted in the direction of decreasing time, using the timestamps of the change direction symbol sequence as a reference. During this shift, symbols in the adjustment direction symbol sequence whose timestamps exceed the range of the change direction symbol sequence are discarded, and the empty timestamp positions at the end of the adjustment direction symbol sequence due to the shift are filled with resting symbols. After the shift, symbols with the same timestamps in the adjustment direction symbol sequence and the change direction symbol sequence are grouped into a symbol pair. All symbol pairs are arranged in timestamp order to obtain the aligned adjustment-change symbol pair sequence. In the aligned adjustment-change symbol pair sequence, each symbol pair contains one adjustment direction symbol and one change direction symbol. The adjustment direction symbol comes from the adjusted direction symbol sequence shifted according to the preliminary lag relation, and the change direction symbol comes from the unshifted change direction symbol sequence.

[0055] Check the proportion of the change direction symbol leading the adjustment direction symbol in the aligned adjustment-change symbol pair sequence.

[0056] In the aligned regulation-change symbol pair sequence, each symbol pair is examined. Symbol pairs where the timestamp of the regulation direction symbol is later than that of the change direction symbol are identified as leading symbol pairs. After time alignment according to the initial lag relationship, the regulation direction symbol sequence has shifted in the direction of decreasing time. In an ideal positive causal situation, the timestamp of the regulation direction symbol should be equal to or earlier than that of the change direction symbol. When a symbol pair in the aligned regulation-change symbol pair sequence has a timestamp of the regulation direction symbol later than that of the change direction symbol, it indicates that the change direction symbol leads the regulation direction symbol, meaning that the energy efficiency change occurs earlier than the control action, which is a manifestation of reverse causality. Dividing the number of leading symbol pairs by the total number of symbol pairs in the aligned regulation-change symbol pair sequence yields the proportion of change direction symbols leading regulation direction symbols. This proportion reflects the percentage of events in all symbol pairs where the energy efficiency response leads the control action under the initial lag relationship.

[0057] When the proportion by which the direction of change precedes the direction of adjustment exceeds a preset proportion, it is determined that the preliminary lag relationship contains a reverse causal component, and the preliminary lag relationship is marked as invalid.

[0058] The preset ratio is set based on the unidirectional causal characteristics of control and energy efficiency in industrial control systems under steady-state conditions. Within the steady-state range, control actions are the driving factor for energy efficiency changes, and energy efficiency changes are the response result of control actions. Ideally, the timestamp of the adjustment direction symbol should not be later than the timestamp of the change direction symbol. Due to random fluctuations in online instruments of industrial control systems caused by sampling jitter and signal transmission delays, a small number of symbol pairs may have timestamps of adjustment direction symbols later than those of change direction symbols. The ratio caused by such random fluctuations is usually in a low value range. The preset ratio is used to distinguish between sporadic reverse symbol pairs caused by random fluctuations and reverse symbol pairs caused by systematic reverse causal components. The specific setting method for the preset ratio is as follows: Manually labeled sample segments with a stable positive causal relationship between control and energy efficiency are selected from the historical database of the industrial control system. For each sample segment, the ratio of the change direction symbol leading the adjustment direction symbol is calculated according to the alignment and verification process in this step. The maximum value among all sample segments is multiplied by a preset margin coefficient to obtain the preset ratio. The preset margin coefficient ranges from 1.1 to 1.3, used to leave a reasonable margin for random fluctuations. For example, if the maximum value of the corresponding proportion in all sample segments is 0.08, and the preset margin coefficient is 1.25, then the preset proportion is set to 0.10. When the proportion in which the sign of change leads the sign of adjustment exceeds the preset proportion, it indicates that the number of opposite sign pairs exceeds the range that random fluctuations can explain, and there is a systematic reverse causal component in the preliminary lag relationship; therefore, the preliminary lag relationship is marked as invalid. When the proportion in which the sign of change leads the sign of adjustment does not exceed the preset proportion, it indicates that the number of opposite sign pairs is within the range of random fluctuations; therefore, the preliminary lag relationship is determined to be a positive lag relationship.

[0059] Compared to existing technologies that directly use identification results as lag relationships, this approach adds a causal direction screening step based on the sequence of timestamps using symbols. The initial lag relationship is determined by the peak value of the conditional entropy reduction. While the peak value indicates the maximum time delay of information transmission intensity, it doesn't inherently distinguish whether the information transmission direction is driven by control efficiency or vice versa. By aligning the initial lag relationship with time and examining the proportion where the sign of change precedes the sign of adjustment, we can determine whether a reverse causal component is mixed into the initial lag relationship based on the temporal sequence. Preliminary lag relationships containing reverse causal components are eliminated, retaining only positive lag relationships that satisfy the causal direction, providing a more accurate lag basis for subsequent baseline reconstruction steps.

[0060] In a specific implementation of the benchmark reconfiguration module, the starting moment when historical control actions affect the current energy efficiency value is determined based on the positive hysteresis relationship.

[0061] The positive hysteresis relation is output from the relation extraction step. It is a time-based delay, representing the length of time from the issuance of a control action to the resulting observable change in energy efficiency. The starting time at which historical control actions affect the current energy efficiency value is determined by subtracting the time length corresponding to the positive hysteresis relation from the current time. The starting time signifies that control actions issued at or after the starting time have already produced and are observable in terms of energy efficiency at the current time; control actions issued before the starting time have had their effects decay completely before the current time and are no longer included in the residual component of the current energy efficiency value.

[0062] Extract the amplitude of each adjustment action from the blower frequency sequence and valve opening sequence within the interval from the start time to the current time.

[0063] The frequency difference between adjacent timestamps in the blower frequency sequence is calculated point-by-point. Adjacent timestamps refer to consecutive timestamps one before and one after. The frequency difference is obtained by subtracting the frequency value of the previous time from the frequency value of the next time. Non-zero differences are recorded as the amplitude of the blower frequency adjustment action. The amplitude of the blower frequency adjustment action and its corresponding time are included in the adjustment action amplitude set. Similarly, the valve opening difference between adjacent timestamps in the valve opening sequence is calculated point-by-point. The opening difference is obtained by subtracting the opening value of the previous time from the opening value of the next time. Non-zero differences are recorded as the amplitude of the valve opening adjustment action. The amplitude of the valve opening adjustment action and its corresponding time are included in the adjustment action amplitude set. Each record in the adjustment action amplitude set contains two fields: action time and action amplitude. The action time is the timestamp of the time following the action amplitude.

[0064] The amplitude of each adjustment action is attenuated according to a preset attenuation coefficient to obtain the sub-residual component of the current energy efficiency value for each adjustment action.

[0065] For each record in the set of adjustment action amplitudes, the time interval from the action moment to the current moment is calculated, and the action amplitude is multiplied by a decay coefficient to obtain the sub-residual component. The decay coefficient is calculated by multiplying the action amplitude by a decay function value with the time interval as the independent variable. The decay function used is an exponential decay function, in the form e^(-∞ / ∞). (-t / T)denoted as an exponential function with the natural constant e as its base, t represents the time interval from the moment of action to the current moment, and T represents the decay time constant. The decay time constant T is set according to the response characteristics of dissolved oxygen in the aerobic tank to aeration adjustment. The decay time constant T is calibrated from the historical database of the industrial control system through step response test data. The specific calibration method is as follows: under steady-state conditions, a step increment is applied to the blower frequency, and the time taken for dissolved oxygen to rise from the moment the step is applied to a specified proportion of the stable value of the step response is recorded. The recorded time is used as the value of the decay time constant T. The specified proportion is taken as, for example, 0.63 of the stable value of the step response, that is, the time taken for dissolved oxygen to rise to 63% of the stable value of the step response is recorded. If the recorded time is, for example, 8 minutes, then the decay time constant T is set to 8 minutes.

[0066] The dimensions of the residual component are consistent with those of the aeration efficiency sequence. The larger the time interval t from the action moment to the current moment, the larger the decay function value e. (-t / T) The smaller the value, the smaller the corresponding residual component, indicating that the earlier the adjustment action occurs, the weaker the residual impact on the current energy efficiency value.

[0067] The residual components caused by historical control actions are obtained by summing up the sub-residual components of all adjustment actions within the interval from the start time to the current time.

[0068] The residual components corresponding to all records in the set of adjustment action amplitudes are summed. The summation result is the residual component caused by the historical control actions. The residual component represents the cumulative effect left in the energy efficiency value at the current time by all adjustment actions in the blower frequency sequence and valve opening sequence from the start time to the current time.

[0069] The residual component is subtracted from the current energy efficiency value to obtain the inherent energy efficiency benchmark of the process.

[0070] The current energy efficiency value is the energy efficiency value at the current time in the aeration energy efficiency sequence. Subtracting the residual component from the current energy efficiency value yields the process-specific energy efficiency benchmark. The process-specific energy efficiency benchmark eliminates the self-interference components accumulated in the energy efficiency index by historical control actions, reflecting the true energy consumption level of the process at the current time.

[0071] By identifying the starting point of influence through positive hysteresis, and using a decay function to perform time-weighted decay processing on historical adjustment actions and then accumulating the residual components, the self-interference component of the control action is extracted from the current energy efficiency value. Compared with the existing technology that directly uses the original energy efficiency value containing self-interference components as the feedback indicator, the obtained process-inherent energy efficiency benchmark can objectively reflect the true energy consumption level of the process. It eliminates the self-misleading control cycle caused by the online energy efficiency index being contaminated by its own actions in the background technology, and provides an accurate basis for the correction of the subsequent optimization direction.

[0072] In a specific implementation of the direction correction module, the difference between the inherent energy efficiency baseline of the process and the current energy efficiency value is calculated to obtain the self-interference deviation.

[0073] The inherent energy efficiency benchmark is output by the benchmark reconstruction module. This benchmark eliminates the self-interference components accumulated in the energy efficiency indicators from historical control actions, reflecting the true energy consumption level of the process at the current moment. The current energy efficiency value is the energy efficiency value with the current timestamp in the aeration energy efficiency sequence. Subtracting the inherent energy efficiency benchmark from the current energy efficiency value yields the self-interference deviation. This deviation quantifies the degree of pollution to online energy efficiency indicators caused by the control system's own adjustments. A positive self-interference deviation indicates that the current energy efficiency value is higher than the inherent energy efficiency benchmark, and historical control actions have a positive effect on improving energy efficiency; a negative self-interference deviation indicates that the current energy efficiency value is lower than the inherent energy efficiency benchmark, and historical control actions have a negative effect on lowering energy efficiency.

[0074] Determine whether the self-interference deviation exceeds the preset deviation range.

[0075] The preset deviation range is set based on the statistical characteristics of energy efficiency fluctuations in the industrial control system under conditions of no self-interference. Under ideal conditions of no self-interference, the deviation between the current energy efficiency value and the inherent energy efficiency benchmark of the process is caused only by the measurement noise of online instruments and random fluctuations in the process, and the deviation value is symmetrically distributed around zero. The specific setting method of the preset deviation range is as follows: select manually marked stable periods of process operation from the historical database of the industrial control system, extract the inherent energy efficiency benchmark of the process obtained after processing by the benchmark reconstruction module within the stable period and the current energy efficiency value at the corresponding time, calculate the self-interference deviation at each time, and statistically analyze the standard deviation of the self-interference deviation sequence. Set the upper limit of the preset deviation range to a positive standard deviation multiplied by a preset multiple, and set the lower limit of the preset deviation range to a negative standard deviation multiplied by a preset multiple. For example, the preset multiple is 3, that is, the preset deviation range is a closed interval from negative 3 times the standard deviation to positive 3 times the standard deviation. When the self-interference deviation is greater than the upper limit of the preset deviation range or less than the lower limit of the preset deviation range, it is determined that the self-interference deviation exceeds the preset deviation range, indicating that the current energy efficiency value has been significantly contaminated by historical control actions, and optimization direction correction needs to be initiated. When the self-interference deviation is within the preset deviation range, it indicates that the self-interference deviation is within the range of random fluctuations, and the original optimization direction of the multi-parameter collaborative controller remains unchanged.

[0076] When the self-interference deviation exceeds the preset deviation range, the current original optimization direction of the multi-parameter collaborative controller is obtained.

[0077] A multi-parameter cooperative controller is a control unit in an industrial control system responsible for solving multivariable optimization problems. In each control cycle, the multi-parameter cooperative controller outputs an initial optimization direction, which is a multi-dimensional vector. Each dimension of the multi-dimensional vector corresponds to the adjustment direction and magnitude of a control variable. This direction is obtained by reading the initial optimization direction vector for the current control cycle from the output interface of the multi-parameter cooperative controller.

[0078] The directional component of the self-interference deviation along the original optimization direction is used as the correction amount. The correction amount is subtracted from the original optimization direction to obtain the reconstructed optimization direction.

[0079] The self-interference deviation is multiplied by the unit vector of the original optimization direction to obtain the projected scalar of the self-interference deviation in the original optimization direction. The unit vector of the original optimization direction is calculated by dividing the original optimization direction vector by its own magnitude, which is the square root of the sum of the squares of the components of the original optimization direction vector. The projected scalar is calculated as: P = D·U; where P represents the projected scalar, D represents the self-interference deviation, U represents the unit vector of the original optimization direction, and · represents the dot product operation. The projected scalar is multiplied by the unit vector of the original optimization direction to obtain the correction amount, calculated as: C = P × U; where C represents the correction amount. The correction amount represents the vector component of the self-interference deviation in the original optimization direction, i.e., the misleading offset of the optimization direction caused by the historical control action. The correction amount is subtracted from the original optimization direction to obtain the reconstructed optimization direction, calculated as: Vrecon = Vorig - C; where Vrecon represents the reconstructed optimization direction, and Vorig represents the original optimization direction.

[0080] The refactoring optimization direction is used as the optimization direction for the corrected multi-parameter collaborative controller.

[0081] The reconstructed optimization direction is written into the optimization direction output interface of the multi-parameter collaborative controller, replacing the original optimization direction. In the next control cycle, the multi-parameter collaborative controller executes the adjustment commands for the blower frequency and valve opening according to the reconstructed optimization direction.

[0082] This step corrects the optimization direction by removing the component along the original optimization direction from the self-interference deviation. Compared with the existing technology that directly uses contaminated energy efficiency indicators to drive optimization calculations, the corrected optimization direction no longer contains misleading deviations induced by the self-induction of historical control actions. This avoids the observer effect control degradation caused by the self-feedback contamination of online energy efficiency indicators in the background technology, and ensures that the multi-parameter collaborative controller always converges towards the true optimal operating condition of the process, suppressing the implicit degradation of energy efficiency during long-term operation.

[0083] All calculations involved in the embodiments are performed using dimensionless numerical values, and the preset parameters and thresholds in the calculations can be set by those skilled in the art according to actual conditions.

[0084] This technical solution can be flexibly deployed, for example, as embedded software running on device hardware, or installed on personal computers or other smart terminals with user interfaces, thus adapting to various hardware environments and usage requirements.

[0085] The above solutions can be implemented in software, hardware, firmware, or a combination thereof. When implemented in software, they are presented, in whole or in part, as a computer program product, containing one or more computer instructions or programs. When these instructions or programs are loaded and executed on a computer, results are produced corresponding to the processes or functions of the embodiments of this application. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. Computer instructions can be stored in a computer-readable storage medium or transmitted from one medium to another via wired or wireless means, such as from a website, server, or data center via wired means such as fiber optic cables, twisted-pair cables, or coaxial cables, or wireless means such as infrared or microwaves to another site. A computer-readable storage medium refers to any usable medium that a computer can access or a data storage device such as a server or data center that contains one or more usable media, including magnetic media such as floppy disks, hard disks, and magnetic tapes, optical media such as DVDs, and semiconductor media such as solid-state drives.

[0086] The specific working process of the system, device and module can be found in the method embodiment, and will not be repeated here.

[0087] The disclosed systems, devices, and methods can be implemented in other ways. The device embodiments are for illustrative purposes only, and the module division is only a logical division. In practice, different divisions can be implemented, such as merging or integrating multiple modules or components, or omitting some features. Coupling, direct coupling, or communication connections between the components can be achieved through interfaces, while indirect coupling or communication connections can take electrical, mechanical, or other forms.

[0088] The modules described as separate components may or may not be physically separated. The components shown as modules can be physical hardware or software, and can be deployed centrally or distributed across multiple network nodes. Some or all of the modules can be selected to achieve the purpose of this embodiment, depending on actual needs.

[0089] The functional modules in each embodiment can be integrated into one processing module, or they can exist independently, or two or more modules can be integrated into one.

[0090] If the functionality is implemented as a software module and used as an independent product, it can be stored in a computer-readable storage medium. Under this understanding, the substantial contribution of the technical solution of this application can be embodied in the form of a software product. This computer software product is stored in a storage medium and contains instructions to cause a computer device, such as a personal computer, server, or network device, to execute all or part of the steps of the methods in the embodiments of this application. The storage medium includes any medium capable of storing program code, such as a USB flash drive, portable hard drive, read-only memory, random access memory, magnetic disk, or optical disk.

[0091] The above are merely specific embodiments of this application, and the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should fall within the scope of protection of this application.

Claims

1. A multi-parameter collaborative energy-saving aeration control system for municipal wastewater treatment plants, characterized in that, include: The sequence acquisition module is used to acquire the aeration efficiency sequence, blower frequency sequence, and valve opening sequence of the industrial control system at the current moment and within the previous preset time period. The interval determination module is used to compare the direction of adjustment action in the blower frequency sequence and valve opening sequence with the direction of change of the aeration energy efficiency sequence in the corresponding time period. When the proportion of different directions exceeds the preset proportion threshold, the aeration energy efficiency sequence is determined to be in the steady state interval. The relationship identification module is used to symbolize the adjustment direction of the blower frequency sequence and valve opening sequence, as well as the change direction of the aeration energy efficiency sequence, within the steady-state range. It calculates the conditional entropy reduction from the adjustment direction symbol sequence to the change direction symbol sequence under different time delays, and uses the time delay of the maximum conditional entropy reduction as the preliminary lag relationship between historical control actions and energy efficiency response. The relationship extraction module is used to perform causal direction testing on the preliminary lag relationship, remove the reverse causal components in which the energy efficiency response leads the control action, and obtain the positive lag relationship; The baseline reconstruction module is used to estimate the residual component caused by historical control actions in the current energy efficiency value using the positive hysteresis relationship, and to subtract the residual component from the current energy efficiency value to obtain the process-inherent energy efficiency baseline. The direction correction module is used to correct the optimization direction of the multi-parameter collaborative controller by utilizing the deviation between the inherent energy efficiency baseline of the process and the current energy efficiency value.

2. The multi-parameter coordinated energy-saving aeration control system for municipal wastewater treatment plants according to claim 1, characterized in that, Acquire the aeration efficiency sequence, blower frequency sequence, and valve opening sequence of the industrial control system at the current moment and within the previously preset time period, including: Load the original sequence of aeration efficiency, the original sequence of blower frequency, and the original sequence of valve opening for the preset time period before the current moment into the historical database of the industrial control system. Align the timestamps in the original sequences of aeration energy efficiency, blower frequency, and valve opening, and remove records with no corresponding timestamp values ​​to obtain aeration energy efficiency aligned sequences, blower frequency aligned sequences, and valve opening aligned sequences with a unified time base. From the aeration efficiency alignment sequence, blower frequency alignment sequence, and valve opening alignment sequence with a unified time base, segments that trace back a preset time from the current moment are extracted as the aeration efficiency sequence, blower frequency sequence, and valve opening sequence.

3. The multi-parameter coordinated energy-saving aeration control system for municipal wastewater treatment plants according to claim 1, characterized in that, By comparing the direction of adjustment actions in the blower frequency sequence and valve opening sequence with the direction of change in the aeration efficiency sequence within the corresponding time period, when the proportion of dissimilar directions exceeds a preset proportion threshold, the aeration efficiency sequence is determined to be in a steady-state range, including: By comparing the frequency values ​​of two adjacent moments in the blower frequency sequence point by point, the moment when the frequency value increases is marked as the positive adjustment action point, and the moment when the frequency value decreases is marked as the negative adjustment action point; By comparing the opening values ​​of two adjacent moments in the valve opening sequence point by point, the moment when the opening value increases is marked as the positive adjustment action point, and the moment when the opening value decreases is marked as the negative adjustment action point; The adjustment direction marks at the same moment in the blower frequency sequence and valve opening sequence are merged. If both are positive adjustment action points at the same moment, they are merged into a positive adjustment action point. If both are negative adjustment action points at the same moment, they are merged into a negative adjustment action point. If the two directions are inconsistent at the same moment, they are not recorded as adjustment action points, thus obtaining the adjustment action direction sequence. By comparing the energy efficiency values ​​of two adjacent moments in the aeration energy efficiency sequence point by point, the moments when the energy efficiency value increases are marked as positive change points, and the moments when the energy efficiency value decreases are marked as negative change points, thus obtaining the energy efficiency change direction sequence; Align the sequence of adjustment action directions with the sequence of energy efficiency change directions by timestamp, and count the proportion of adjustment action points whose directions differ from the corresponding energy efficiency change points to the total number of adjustment action points. When the proportion of opposite directions exceeds the preset proportion threshold, the aeration efficiency sequence is determined to be in the steady state range.

4. The multi-parameter coordinated energy-saving aeration control system for municipal wastewater treatment plants according to claim 1, characterized in that, Within the steady-state range, the adjustment directions of the blower frequency sequence and valve opening sequence, as well as the change direction of the aeration energy efficiency sequence, are symbolized. The conditional entropy reduction from the adjustment direction symbol sequence to the change direction symbol sequence under different time delays is calculated. The time delay of the maximum conditional entropy reduction is taken as the initial lag relationship between historical control actions and energy efficiency responses, including: In the sequence of adjustment action directions, positive adjustment action points are denoted as positive symbols, negative adjustment action points are denoted as negative symbols, and moments with no adjustment action points are denoted as resting symbols, thus obtaining the sequence of adjustment direction symbols; In the energy efficiency change direction sequence, positive change points are denoted as positive signs, and negative change points are denoted as negative signs, thus obtaining the change direction sign sequence; The direction symbol sequence is shifted according to different time delays, and the conditional entropy of the change direction symbol sequence under each shift amount is calculated when the direction symbol sequence is adjusted after a known shift. Calculate the unconditional entropy of the change direction symbol sequence when the adjustment direction symbol sequence is unknown, and subtract the conditional entropy of each translation amount from the unconditional entropy to obtain the conditional entropy reduction for each translation amount. The shift that results in the largest reduction in conditional entropy is determined as the initial lag relationship between historical control actions and energy efficiency response.

5. The multi-parameter coordinated energy-saving aeration control system for municipal wastewater treatment plants according to claim 1, characterized in that, A causal orientation test was performed on the initial lag relationship, and the reverse causal components where the energy efficiency response precedes the control action were eliminated, resulting in a positive lag relationship, including: The adjustment direction symbol sequence and the change direction symbol sequence are time-aligned according to the initial lag relationship to obtain the aligned adjustment-change symbol pair sequence; Check the proportion of the change direction sign leading the adjustment direction sign in the aligned adjustment-change sign pair sequence; When the proportion by which the change direction sign leads the adjustment direction sign exceeds a preset proportion, it is determined that the preliminary lag relationship contains a reverse causal component, and the preliminary lag relationship is marked as invalid. When the proportion by which the direction of change precedes the direction of adjustment does not exceed the preset proportion, the initial lag relationship is determined to be a positive lag relationship.

6. The multi-parameter coordinated energy-saving aeration control system for municipal wastewater treatment plants according to claim 5, characterized in that, The proportion in which the change direction sign leads the adjustment direction sign in the aligned sequence of adjustment-change sign pairs is checked one by one, including: Examine each symbol pair in the aligned adjustment-change symbol pair sequence, and determine the symbol pair whose timestamp of the adjustment direction symbol is later than the timestamp of the change direction symbol as the symbol pair whose change direction symbol is leading; Divide the number of symbol pairs that lead the change direction symbol by the total number of symbol pairs in the aligned adjustment-change symbol pair sequence to obtain the ratio of the change direction symbol leading the adjustment direction symbol.

7. The multi-parameter coordinated energy-saving aeration control system for municipal wastewater treatment plants according to claim 1, characterized in that, The residual component caused by historical control actions in the current energy efficiency value is estimated using the positive hysteresis relationship, and the residual component is subtracted from the current energy efficiency value to obtain the process-specific energy efficiency benchmark, including: Determine the starting point at which historical control actions affect the current energy efficiency value based on the positive lag relationship; Extract the amplitude of each adjustment action from the blower frequency sequence and valve opening sequence within the interval from the start time to the current time. The amplitude of each adjustment action is attenuated according to a preset attenuation coefficient to obtain the sub-residual component of each adjustment action on the current energy efficiency value. The residual components caused by historical control actions are obtained by summing up the sub-residual components of all adjustment actions within the interval from the start time to the current time. The residual component is subtracted from the current energy efficiency value to obtain the inherent energy efficiency benchmark of the process.

8. A multi-parameter coordinated energy-saving aeration control system for municipal wastewater treatment plants according to claim 7, characterized in that, Extract the amplitude of each adjustment action from the blower frequency sequence and valve opening sequence within the interval from the start time to the current time, including: Calculate the difference between frequency values ​​at adjacent moments in the blower frequency sequence point by point, and record the non-zero difference as the action amplitude of the blower frequency adjustment action; Calculate the difference between the opening values ​​at adjacent moments in the valve opening sequence point by point, and record the non-zero difference as the action amplitude of the valve opening adjustment action; The amplitude values ​​of the blower frequency adjustment action and the valve opening adjustment action are uniformly included in the adjustment action amplitude set.

9. A multi-parameter coordinated energy-saving aeration control system for municipal wastewater treatment plants according to claim 1, characterized in that, By utilizing the deviation between the inherent energy efficiency baseline of the process and the current energy efficiency value, the optimization direction of the multi-parameter collaborative controller is corrected, including: The self-interference bias is obtained by calculating the difference between the inherent energy efficiency baseline of the process and the current energy efficiency value. Determine whether the self-interference deviation exceeds the preset deviation range; When the self-interference deviation exceeds the preset deviation range, obtain the current original optimization direction of the multi-parameter collaborative controller; The directional component of the self-interference deviation along the original optimization direction is used as the correction amount. The correction amount is subtracted from the original optimization direction to obtain the reconstructed optimization direction. The refactoring optimization direction is used as the optimization direction for the corrected multi-parameter collaborative controller.

10. A multi-parameter coordinated energy-saving aeration control system for municipal wastewater treatment plants according to claim 9, characterized in that, The directional component of the self-interference deviation along the original optimization direction is used as the correction amount, including: The self-interference bias is multiplied by the unit vector of the original optimization direction to obtain the projection scalar of the self-interference bias in the original optimization direction. The correction is obtained by multiplying the projected scalar by the unit vector of the original optimization direction.

Images