Biological wax water purification self-adaptive control method based on runoff load prediction

Abstract

This invention relates to the field of water treatment control technology, and provides an adaptive control method for bio-wax water purification based on runoff load prediction. The method includes: acquiring a geographic information system (GIS) layer of the drainage network in the catchment area; dividing the catchment area into multiple sub-catchment units and calculating the hydraulic transmission time of each sub-catchment unit; calculating the pollution accumulation density score and corresponding contribution weight of each sub-catchment unit by combining the number of drought days in the previous period, the impermeable area ratio, the area of ​​the sub-unit, and the pollution intensity coefficient; calculating the local concentration time series using an exponential decay model, and superimposing the local concentration time series of the sub-catchment units that have reached the inlet according to their contribution weights to obtain a superimposed predicted concentration time series; calculating the target influent flow velocity based on the superimposed predicted concentration time series, and generating a feedforward command sequence. This invention accurately reconstructs the initial flushing concentration time series, effectively avoiding the problem of insufficient early shock treatment capacity.

Description

Technical Field

[0001] This application relates to the field of data processing technology, and more specifically, to an adaptive control method for bio-wax water purification based on runoff load prediction. Background Technology

[0002] With the acceleration of urbanization, non-point source pollution has become a significant factor affecting urban water quality. In the initial stages of rainfall, surface runoff washes away large amounts of pollutants accumulated on roads, forming highly concentrated initial runoff. To effectively treat this high-concentration runoff, bio-wax water purification units are widely used in the end-of-pipe interception and purification control of drainage systems.

[0003] However, traditional runoff pollution prediction models typically employ homogenization assumptions, treating the entire catchment area as a single pollution-generating unit and ignoring the spatial heterogeneity of pollution accumulation density and hydraulic transport time across different regions. This homogenization process results in the model masking out the premature concentration surges that occur in areas with high pollution density and close proximity during the initial stages of rainfall.

[0004] Because the model fails to accurately reflect the early concentration peaks caused by spatial heterogeneity, the feedforward control system severely underestimates the early concentration surges, resulting in excessively high influent flow rate commands. This leads to insufficient hydraulic retention time for the bio-wax purification unit during the initial flushing phase, making it unable to provide adequate adsorption and treatment capacity. Ultimately, this results in effluent water quality exceeding standards, making it difficult to effectively intercept and purify the initial flushing contaminants. Summary of the Invention

[0005] To address the problem that existing homogeneous prediction models, which smooth out spatial heterogeneity, lead to insufficient early shock treatment capacity, this application provides an adaptive control method for bio-wax water purification based on runoff load prediction.

[0006] This invention provides an adaptive control method for biowax-based water purification based on runoff load prediction, comprising: acquiring a geographic information system layer of the drainage network in the catchment area, dividing the catchment area into multiple sub-catchment units, and calculating the hydraulic transmission time of each sub-catchment unit; acquiring the number of drought days preceding a rainfall event, and calculating a pollution accumulation density score for each sub-catchment unit by combining the impermeable area ratio, sub-unit area, and pollution intensity coefficient of each sub-catchment unit, and calculating a corresponding contribution weight based on the pollution accumulation density score; calculating the local concentration time series of each sub-catchment unit using an exponential decay model based on the hydraulic transmission time and the contribution weight, and superimposing the local concentration time series of the sub-catchment units that have reached the inlet according to the contribution weight to obtain a superimposed predicted concentration time series; calculating the target inlet flow velocity based on the superimposed predicted concentration time series, and generating a feedforward command sequence for adaptive control of the biowax purification unit.

[0007] This invention transforms the spatial topological distribution of the catchment area into a defined hydraulic transport time and calculates the pollution accumulation density score of each region by combining multi-dimensional parameters. This breaks through the limitations of traditional homogeneous models and lays a solid data foundation for the subsequent reconstruction of high-precision initial flushing concentration time series.

[0008] Preferably, the step of obtaining the drainage network geographic information system layer of the catchment area, dividing the catchment area into multiple sub-catchment units, and calculating the hydraulic transmission time of each sub-catchment unit includes: importing the drainage network geographic information system layer, dividing the catchment area into multiple sub-catchment units using the confluence nodes on the drainage main as the dividing boundary; reading the pipe diameter, slope, and cross-sectional parameters of all pipe segments on the corresponding path of each sub-catchment unit, and calculating the full-flow velocity of each pipe segment; accumulating the flow time of each pipe segment and the surface confluence time to obtain the hydraulic transmission time of each sub-catchment unit, and arranging each sub-catchment unit in ascending order of the hydraulic transmission time to form a hydraulic arrival sequence.

[0009] By automatically transforming the complex physical topology of the pipeline network into an ordered time series, the arrival times of runoff at different spatial locations have a strict sequential logic, avoiding errors from manual calculation.

[0010] Preferably, after arranging each of the sub-catchment units in ascending order according to the hydraulic transmission time to form a hydraulic arrival sequence, the method further includes: extracting the surface functional zoning type label, the impermeable area ratio, and the sub-unit area of ​​each of the sub-catchment units from the drainage network geographic information system layer.

[0011] Preferably, the step of obtaining the number of drought days preceding the rainfall event, and calculating the pollution accumulation density score for each sub-catchment unit by combining the impervious area ratio, sub-unit area, and pollution intensity coefficient of each sub-catchment unit, includes: multiplying the number of drought days preceding the event, the impervious area ratio, and the pollution intensity coefficient, and dividing by the sub-unit area to obtain the original pollution accumulation density score; and dividing the original pollution accumulation density score by the arithmetic mean of the entire region to obtain a dimensionless normalized score as the pollution accumulation density score.

[0012] Preferably, the step of calculating the corresponding contribution weight based on the pollution accumulation density score includes: dividing the dimensionless normalized score of each sub-catchment unit by the sum of the dimensionless normalized scores of all sub-catchment units to obtain the contribution weight; and combining the hydraulic transmission time of each sub-catchment unit with the contribution weight to form a contribution matrix arranged in ascending order of the hydraulic transmission time.

[0013] Preferably, the step of calculating the local concentration time series of each of the sub-catchment units using the exponential decay model includes: multiplying the baseline concentration by the pollution accumulation density score to obtain the initial concentration; multiplying the baseline decay rate by the pollution accumulation density score to obtain the local decay rate; and calculating the local concentration time series using the exponential decay model based on the initial concentration, the local decay rate, and the hydraulic transmission time.

[0014] Preferably, the step of superimposing the local concentration time series of the sub-catchment units that have reached the inlet according to the contribution weight to obtain the superimposed predicted concentration time series includes: at any time, selecting the sub-catchment units that have reached the inlet whose hydraulic transmission time does not exceed that time; multiplying the local concentration time series of the sub-catchment units that have reached the inlet by the corresponding contribution weight and summing them to obtain the superimposed predicted concentration; and discretizing the superimposed predicted concentration according to a preset sampling period to obtain the superimposed predicted concentration time series.

[0015] Preferably, the step of calculating the target influent flow velocity based on the superimposed predicted concentration time series and generating a feedforward instruction sequence includes: obtaining the system rated flow velocity, maximum pollutant treatment capacity, and predicted runoff; dividing the maximum pollutant treatment capacity by the product of the superimposed predicted concentration time series and the predicted runoff, and multiplying by the rated flow velocity to calculate the target influent flow velocity; when the target influent flow velocity exceeds the rated flow velocity, limiting the target influent flow velocity to the rated flow velocity, and generating the feedforward instruction sequence.

[0016] Preferably, after generating the feedforward instruction sequence, the method further includes: when the absolute value of the deviation between the measured concentration and the superimposed predicted concentration exceeds a set threshold, proportionally correcting the target influent flow rate in the unexecuted interval according to the ratio of the measured concentration to the superimposed predicted concentration.

[0017] Preferably, after proportionally correcting the target influent flow velocity in the unexecuted interval according to the ratio of the measured concentration to the superimposed predicted concentration, the method further includes: after the rainfall event ends, calculating the mean residual between the measured concentration and the superimposed predicted concentration; allocating the mean residual according to the contribution weight ratio of each sub-catchment unit, and correcting the pollution intensity coefficient of the corresponding sub-catchment unit.

[0018] By inversely distributing the residuals of single-point observations according to a known weight structure, the underdetermined problem of multi-parameter inverse solutions is avoided, enabling the prediction model to automatically converge to the local true pollution characteristics as historical rainfall events occur.

[0019] This invention provides an adaptive control method for bio-wax water purification based on runoff load prediction. By transforming the spatial topological distribution of the catchment area into a hydraulic transport time series and combining it with the pollution accumulation density of each sub-unit to calculate independent weights, the spatial heterogeneity information that was originally smoothed out by the homogeneous model is effectively restored to the concentration time series at the inlet. This enables the system to accurately predict the phenomenon of premature concentration surge caused by high pollution in the near end area in the early stage of rainfall.

[0020] Based on the reconstructed high-precision concentration timing, this invention can actively reduce the influent flow rate in the early time window, allowing the biowax purification unit to have a longer hydraulic residence time. This not only ensures the full progress of the biowax adsorption process, but also effectively solves the problem of insufficient treatment capacity of the purification unit caused by the underestimation of the early concentration impact in traditional homogenization models, significantly improving the stability of the effluent water quality compliance. Attached Figure Description

[0021] The above and other objects, features, and advantages of exemplary embodiments of the present invention will become readily apparent upon reading the following detailed description with reference to the accompanying drawings. In the drawings, several embodiments of the invention are illustrated by way of example and not limitation, and like or corresponding reference numerals denote like or corresponding parts, wherein: Figure 1 This is a schematic flowchart illustrating an adaptive control method for bio-wax water purification based on runoff load prediction, as described in this invention. Detailed Implementation

[0022] 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, not all, of the embodiments of the present invention. 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.

[0023] The specific embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0024] This invention discloses an adaptive control method for bio-wax water purification based on runoff load prediction, referring to... Figure 1 This includes steps S1-S4: S1, Sub-collection unit division and hydraulic transmission time calibration.

[0025] During the system deployment phase, technicians imported the entire drainage network GIS layer of the catchment area into the parsing module. Based on the network topology, the parsing module automatically divided the catchment area into sections, using the confluence nodes on the main drainage pipes as the dividing boundaries. Each sub-catchment unit, numbered Each sub-collection unit corresponds to a unique drainage path, which eventually flows into the inlet of the bio-wax purification unit via several pipe segments. The division operation is automatically completed by the topological relationship of the GIS pipe network layer, without the need for manual specification of boundaries one by one.

[0026] In an optional embodiment, taking the urban catchment area as an example, the GIS layer was parsed and identified as follows: Each sub-catchment unit, for each sub-unit The system reads the pipe diameters of all pipe segments along the route from the pipeline network as-built drawing database. ,slope Based on the cross-sectional parameters of the water passage, the full-flow velocity of each pipe segment is calculated segment by segment according to the existing Manning relation. Then, the flow time of each pipe segment is accumulated, and the surface runoff time is added to obtain the sub-unit. Hydraulic transmission time (Unit: min). For example, the hydraulic transmission time for 8 sub-units. The times are: 8 min, 11 min, 14 min, 17 min, 22 min, 27 min, 33 min, and 41 min, arranged in ascending order to form the hydraulic arrival sequence.

[0027] Specifically, the system synchronously extracts the surface functional zoning type label of each sub-unit from the GIS layer. impermeable area ratio and sub-unit area (unit: ), for example, number Near-terminal unit ( The functional zoning type is commercial street block (corresponding to) (Higher value) , ;serial number The remote terminal unit ( The functional zoning type is green space mixed zone. , .

[0028] Furthermore, Subunits After ascending order, the nearest terminal unit ( The runoff from the smaller sub-units arrives at the inlet of the purification unit first, and the pollution load it carries constitutes the main body of the initial flushing concentration peak. When the runoff from the more distant sub-units arrives, the initial flushing peak has usually passed, and the concentration is in the decay stage. This sequencing transforms the spatial topological distribution information of the catchment area into a clear temporal sequence, so that the arrival time of the runoff from each sub-unit can be directly called by the subsequent concentration time-series reconstruction steps.

[0029] In this way, the drainage topology of the catchment area is transformed into a set of ordered hydraulic transport time calibration values, providing a structured temporal basis for the subsequent joint extrapolation of pollution density and arrival time.

[0030] S2, Pollution accumulation density score and contribution weight calculation.

[0031] After each rainfall event is triggered, the system reads the number of drought days preceding that rainfall event. This data is automatically compiled from precipitation records at weather stations, in daily units, combined with pre-calibrated data. , and For each subunit Calculate the raw score for pollution accumulation density This reflects the relative intensity of pollutants accumulated per unit area in this sub-unit during the preceding drought period: in, This is a uniform value across the entire region, reflecting the direct proportional impact of the duration of pre-rainfall accumulation on the total amount of pollutants on the road surface; The higher the value, the worse the permeability of the sub-unit and the greater the proportion of surface runoff eroding the road surface. This is a pollution intensity coefficient for surface functional zoning types, dimensionless. The initial value is determined based on the average concentration (EMC) ratio of typical events for each zoning type in the "Technical Specification for Urban Non-point Source Pollution Control," with a base value of 1 for green space mixed zones. Commercial, residential, and industrial zones are converted according to the corresponding typical EMC ratios in the standard. Subsequent values ​​are calculated after each rainfall event. The data is updated continuously through subsequent historical data regression steps. Located in the denominator, the larger the area, the lower the cumulative density per unit area.

[0032] Furthermore, Divide by the arithmetic mean of the whole region Normalization is performed to obtain a dimensionless normalized score. : thus, The average value for the entire region is always 1. Indicate subunit The pollution accumulation density is higher than the average level for the entire region; This indicates the subunit The pollution accumulation density is not higher than the average level of the entire region. For example, suppose there are 8 sub-units, when... The timing is right, and the commercial street-facing sub-units are located near the end of the street. )of The calculated result is approximately 2.31, for the remote green space mixed sub-unit ( )of The ratio is approximately 0.43, and the ratio directly reflects the significant differentiation in the density of spatial pollution accumulation.

[0033] Specifically, based on Calculate the contribution weight of each sub-unit to the initial flushing peak at the inlet of the purification unit. The weight magnitude is determined solely by the pollution accumulation density, satisfying the following relationship: Since both the numerator and denominator are A linear combination, where the sum of the weights across the entire region is always 1, and the hydraulic transmission time... It is not involved in the calculation of the weight magnitude, but is only used in subsequent steps to determine when the runoff of each sub-unit enters the superposition summation term.

[0034] Each subunit and Combine, form Contribution matrix , No. row storage ,according to Sort in ascending order for direct reading later.

[0035] In this way, the pollution accumulation density of each sub-unit is transformed into a dimensionless weight, and the heterogeneity of spatial distribution is fully preserved and transformed into a computable numerical structure.

[0036] S3, Spatial superposition weighted concentration temporal reconstruction.

[0037] Contribution matrix based on S2 output The initial flushing concentration sequence at the inlet of the purification unit is spatially superimposed and reconstructed to allow for the determination of any given time. The actual runoff received by the inlet is the total hydraulic transmission time not exceeding [a certain amount]. The mixture of subunit runoff, where the local concentration of each subunit that has reached decays exponentially according to its own characteristics.

[0038] Pair of sub-units The local concentration time series after the runoff reaches the inlet Described using an existing exponential decay model: In an optional embodiment, the subunit Initial concentration at the inlet (Unit: mg / L) Calculated using the following formula: in The EMC baseline concentration (unit: mg / L) corresponding to this catchment area type is taken from the typical EMC value of the corresponding zoning type in the "Technical Specification for Urban Non-point Source Pollution Control"; The resulting dimensionless normalized score is obtained by multiplying the two values ​​to achieve a dimensionless value of mg / L. The physical meaning is that sub-units with higher pollution accumulation density carry a higher initial concentration of pollutants when their runoff reaches the inlet, a clear distinction from the homogeneous model where all sub-units share the same initial concentration.

[0039] Specifically, sub-unit Local attenuation rate (Unit: 1 / min) is determined by the following relationship: in (Unit: 1 / min) represents the baseline attenuation rate for the entire region. The system extracts all measured inlet concentration attenuation curves from the historical rainfall event database, successively fits the attenuation rate sequence using the least squares method, and then takes the statistical median of this sequence as the baseline. The current value is automatically updated after each new rainfall event ends. It is a dimensionless quantity, and The dimensions remain 1 / min after multiplication, ensuring the exponent term is equal. Dimensionless High-pollution-density sub-units correspond to larger Physically, this corresponds to the characteristics of more intense initial release of pollutants and faster concentration decay.

[0040] Furthermore, the water inlet of the purification unit is constantly... Superimposed predicted concentration Calculate according to the following formula: The summation of this relation is limited to all... The summation only considers the contributions of sub-units that have actually reached the inlet; sub-units that have not yet reached the inlet are not included in the summation. Within a smaller early time window, the summation term only contains The shortest near-terminal unit; and precisely these near-terminal units correspond to and Both are relatively large, It is also relatively high, therefore The predicted values ​​in the early time window are significantly higher than the corresponding predicted values ​​of the original homogeneous single decay coefficient model, accurately reflecting the premature surge in concentration caused by the temporal and spatial superposition of high pollution density and short hydraulic transport.

[0041] Ultimately, Based on the system sampling period, for example, discretization is performed according to a system sampling period of once per minute to obtain the time series vector. ( (This is the sampling time sequence number), which serves as the direct input for subsequent feedforward instruction calculation.

[0042] In this way, the spatial heterogeneity information that was originally smoothed out by the homogeneous model is completely restored to the inlet concentration time series, enabling feedforward control to have the ability to predict the early surge in concentration during the initial flushing.

[0043] S4. Feedforward instruction sequence generation and adaptive correction.

[0044] Based on the discrete concentration time series of the output The system immediately calculates each time interval within the entire event cycle after a rainfall event is triggered. The corresponding target influent flow rate of the bio-wax purification unit In solving Previously, the system first determined the current time. Whether any sub-units have reached the inlet, i.e., determining the set Is it non-empty? If the set is empty, then... If no runoff reaches the inlet, the pollutants have not yet entered the purification unit, and the system directly... Operating at the rated flow rate, no further calculations are required. Only when the set is not empty, the target influent flow rate is calculated using the following formula: in The rated flow rate of the system is the factory-calibrated value, in m³ / h. This represents the maximum pollutant treatment capacity of the purification unit per unit time, which is the factory rated value, in g / h. For the first The predicted runoff for the interval is determined by the real-time rainfall intensity. The total impervious area of ​​the catchment area is calculated using the existing runoff coefficient formula, in m³ / h. For the output of the first Interval superposition prediction concentration, unit: g / m³, equivalent to mg / L. The physical meaning of this relationship is: ensuring that the total amount of pollutants treated by the purification unit per unit time does not exceed... Under the constraints, the inverse solution yields the maximum allowable inflow velocity for that interval. When the solution result exceeds... At that time, take The rated flow rate of the system is used as the upper limit of the hard constraint.

[0045] In an optional embodiment, when within an earlier time window When the value is significantly higher due to the contribution of the highly polluted subunits near the edge, the corresponding The solution is calculated to be lower than the value instructed by the homogeneous model, so the purification unit obtains a longer hydraulic retention time, and the biowax adsorption process is fully carried out. This effectively avoids the defect of the homogeneous model in this window, which gives too high a flow rate instruction due to underestimating the concentration, resulting in insufficient treatment capacity of the purification unit.

[0046] Furthermore, feedforward instruction sequence The execution adopts an advance distribution method: in each time interval Before the arrival of the flow, a corresponding command is issued with a delay time for the actuator action (calibrated by actual measurements of the inlet valve and flow regulation mechanism during the system commissioning phase, typically 30 to 90 seconds) to ensure that the control action is completed before the runoff actually reaches the inlet. Biowax replenishment trigger flag. The determination is based on the current estimated remaining capacity of the bio-wax adsorption capacity of the purification unit. When the estimated remaining capacity is lower than a set percentage of the rated capacity, Set to 1 to trigger a supplementary operation. This threshold is set by the system administrator during the deployment phase based on the characteristic parameters of the biowax material.

[0047] Specifically, based on the existing 3 sigma standard, the actual concentration measured by the inlet sensor... The predicted concentration is superimposed on the corresponding time. absolute value of deviation When the residual value exceeds three standard deviations of the system's historical prediction residual sequence, the current time interval and all subsequent unexecuted intervals are affected. The model is corrected proportionally to the ratio of measured to predicted concentrations, and the corrected command is issued immediately. This fallback correction handles model prediction bias in a conservative manner, providing limited correction for abnormal deviations without relying on feedback control of the main path.

[0048] After each rainfall event, the system will record the time series of the measured concentrations at the inlet for that event. With overlay prediction time series Calculate the residual sequence for each sampling point and take the mean of the residuals. ,Will Based on the contribution weight of each sub-unit Pollution intensity coefficient proportionally allocated to the corresponding sub-unit The adjustment is made according to the principle of adjusting upwards if the actual measurement is higher than the forecast and downwards if the actual measurement is lower than the forecast. The adjustment step size is a fixed small percentage of the absolute value of the residual mean, for example, 5%, to avoid the influence of residuals from a single event. The violent oscillations. This allocation update method distributes the information from a single inlet scalar observation back to each sub-unit according to a known weight structure, avoiding the underdetermined problem of resolving multi-sub-unit parameters from single-point observations, thus... As historical events accumulate, the pollution gradually converges to the actual local pollution characteristics.

[0049] In this way, the feedforward instructions are highly matched with the actual pollution concentration impact within the critical early time window of the initial flushing, and the operating parameters of the biowax purification unit are fully preset, effectively avoiding the insufficient early impact treatment capacity caused by the homogeneous prediction model.

[0050] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.

[0051] While this specification has shown and described numerous embodiments of the invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Many modifications, alterations, and alternatives will occur to those skilled in the art without departing from the spirit and essence of the invention. It should be understood that various alternatives to the embodiments of the invention described herein may be employed in the practice of this invention.

Claims

1. A bio-wax-based adaptive control method for water purification based on runoff load prediction, characterized in that, include: Obtain the geographic information system layer of the drainage pipe network in the catchment area, divide the catchment area into multiple sub-catchment units, and calculate the hydraulic transmission time of each sub-catchment unit; The number of drought days preceding the rainfall event is obtained. The impermeable area ratio, sub-unit area, and pollution intensity coefficient of each sub-catchment unit are combined to calculate the pollution accumulation density score of each sub-catchment unit. The corresponding contribution weight is calculated based on the pollution accumulation density score. Based on the hydraulic transmission time and the contribution weight, the local concentration time series of each sub-catchment unit is calculated using the exponential decay model, and the local concentration time series of the sub-catchment units that have reached the inlet are superimposed according to the contribution weight to obtain the superimposed predicted concentration time series. Based on the superimposed predicted concentration time series calculation of the target influent flow rate, a feedforward command sequence is generated to adaptively control the biowax purification unit.

2. The adaptive control method for bio-wax water purification based on runoff load prediction according to claim 1, characterized in that, The process of acquiring the geographic information system layer of the drainage network in the catchment area, dividing the catchment area into multiple sub-catchment units, and calculating the hydraulic transmission time of each sub-catchment unit includes: Import the drainage network geographic information system layer, and divide the catchment area into multiple sub-catchment units using the confluence nodes on the drainage main as the dividing boundary; Read the pipe diameter, slope and cross-sectional parameters of all pipe segments on the path corresponding to each sub-collection unit, and calculate the full-flow velocity of each pipe segment; The flow time of each pipe segment and the surface runoff time are summed to obtain the hydraulic transmission time of each sub-catchment unit, and the sub-catchment units are arranged in ascending order of the hydraulic transmission time to form a hydraulic arrival sequence.

3. The adaptive control method for bio-wax water purification based on runoff load prediction according to claim 2, characterized in that, After arranging the sub-collection units in ascending order according to the hydraulic transmission time to form a hydraulic arrival sequence, the method further includes: Extract the surface functional zoning type label, the impermeable area ratio, and the area of ​​each sub-catchment unit from the geographic information system layer of the drainage network.

4. The adaptive control method for bio-wax water purification based on runoff load prediction according to claim 1, characterized in that, The method of obtaining the number of drought days preceding the rainfall event, combined with the impervious area ratio, sub-unit area, and pollution intensity coefficient of each sub-catchment unit, calculates the pollution accumulation density score for each sub-catchment unit, including: Multiply the number of drought days in the early stage, the ratio of impermeable area, and the pollution intensity coefficient, and divide by the area of ​​the sub-unit to obtain the original score of pollution accumulation density; Divide the original pollution accumulation density score by the arithmetic mean of the entire region to obtain a dimensionless normalized score, which is used as the pollution accumulation density score.

5. The adaptive control method for bio-wax water purification based on runoff load prediction according to claim 4, characterized in that, The contribution weight calculated based on the pollution accumulation density score includes: The contribution weight is obtained by dividing the dimensionless normalized score of each sub-catchment unit by the sum of the dimensionless normalized scores of all sub-catchment units. The hydraulic transmission time of each sub-catchment unit is combined with the contribution weight to form a contribution matrix arranged in ascending order of hydraulic transmission time.

6. The adaptive control method for bio-wax water purification based on runoff load prediction according to claim 1, characterized in that, The calculation of the local concentration time series of each sub-catchment unit using the exponential decay model includes: The initial concentration is obtained by multiplying the baseline concentration by the pollution accumulation density score; Multiply the baseline decay rate by the pollution accumulation density score to obtain the local decay rate; Based on the initial concentration, the local decay rate, and the hydraulic transport time, the local concentration time series is calculated using the exponential decay model.

7. The adaptive control method for bio-wax water purification based on runoff load prediction according to claim 6, characterized in that, The step of superimposing the local concentration time series of the sub-collection units that have reached the inlet according to the contribution weight to obtain the superimposed predicted concentration time series includes: At any given time, select the arrived sub-collection units whose hydraulic transmission time does not exceed that time. The local concentration time series that have reached the sub-catchment unit are multiplied by their respective contribution weights and then summed to obtain the superimposed predicted concentration; The superimposed predicted concentration is discretized according to a preset sampling period to obtain the superimposed predicted concentration time series.

8. The adaptive control method for bio-wax water purification based on runoff load prediction according to claim 1, characterized in that, The step of calculating the target influent flow velocity based on the superimposed predicted concentration time series and generating a feedforward instruction sequence includes: Obtain the system's rated flow velocity, maximum pollutant treatment capacity, and predicted runoff volume; The target influent flow rate is calculated by dividing the maximum pollutant treatment capacity by the product of the superimposed predicted concentration time series and the predicted runoff, and then multiplying by the rated flow rate. When the target inlet flow rate exceeds the rated flow rate, the target inlet flow rate is limited to the rated flow rate, and the feedforward instruction sequence is generated.

9. The adaptive control method for bio-wax water purification based on runoff load prediction according to claim 8, characterized in that, After generating the feedforward instruction sequence, the method further includes: When the absolute value of the deviation between the measured concentration and the superimposed predicted concentration exceeds a set threshold, the target influent flow rate in the unexecuted interval is proportionally corrected according to the ratio of the measured concentration to the superimposed predicted concentration.

10. The adaptive control method for bio-wax water purification based on runoff load prediction according to claim 9, characterized in that, After proportionally correcting the target influent flow rate in the unexecuted interval according to the ratio of the measured concentration to the superimposed predicted concentration, the method further includes: After the rainfall event ends, calculate the mean residual between the measured concentration and the superimposed predicted concentration; The residual mean is allocated according to the contribution weight ratio of each of the sub-catchment units, and the pollution intensity coefficient of the corresponding sub-catchment unit is corrected.

Images