Intelligent control method and system for aerobic granular sludge process

Abstract

The present application relates to the field of sewage biological treatment process control, and particularly relates to an intelligent control method and system of an aerobic granular sludge process, which method comprises: collecting an aerobic granular sludge image, extracting a granular size distribution, a surface roughness and a boundary integrity feature to obtain a granular structure state index; collecting multi-point dissolved oxygen data and combining an oxygen demand stable interval to determine an oxygen demand index; dividing a space partition according to a measurement point position to obtain a dissolved oxygen distribution index, obtaining a solid phase concentration profile and partitioning and corresponding, normalizing and calculating a matching deviation index; generating a partition aeration control instruction output to an aeration execution device according to the same, and updating the oxygen demand stable interval when the oxygen demand index and the dissolved oxygen distribution index change in the same direction, so as to realize precise oxygen supply distribution, inhibit local anoxia and over-aeration and reduce energy consumption.

Description

Technical Field

[0001] This invention relates to the field of wastewater biological treatment process control, specifically to an intelligent control method and system for an aerobic granular sludge process. Background Technology

[0002] Aerobic granular sludge processes offer advantages such as small footprint, good settling performance, and resistance to shock loads in the upgrading and renovation of municipal and industrial wastewater. Their large-scale, stable operation directly impacts effluent compliance, energy consumption, and operation and maintenance costs. Current operation and management methods largely rely on single-point dissolved oxygen settings or empirical parameter adjustments, which struggle to reflect the mass transfer differences in particle structure under varying loads and shear conditions. Furthermore, dissolved oxygen within the reactor exhibits spatial heterogeneity, with localized anoxic and over-aeration coexisting, leading to energy waste and the risk of particle disintegration. Moreover, the mismatch between solid-phase stratification and oxygen supply distribution often lacks quantifiable indicators, making it difficult to maintain consistent control strategies across different seasons and water quality conditions, resulting in insufficient portability and verifiability of industrial control systems. Summary of the Invention

[0003] This invention provides an intelligent control method and system for aerobic granular sludge processes, which at least solves the problem of how to achieve zoned aeration closed-loop regulation and adaptively update the oxygen supply demand boundary under conditions of granular structure changes and spatial non-uniformity of dissolved oxygen.

[0004] In a first aspect, the present invention provides an intelligent control method for an aerobic granular sludge process, the method comprising:

[0005] Acquire aerobic granular sludge image data, extract particle size distribution features, surface roughness features, and boundary integrity features to obtain particle structure state indicators;

[0006] Obtain dissolved oxygen measurement data from multiple points, and determine oxygen demand indicators based on particle structure state indicators, multi-point dissolved oxygen measurement data, and oxygen demand stability intervals;

[0007] Based on multi-point dissolved oxygen measurement data, spatial partitions are divided according to the location of the measurement points to determine the dissolved oxygen distribution index, obtain the solid phase concentration profile in the direction of reactor height, and perform partition correspondence and normalization processing on the dissolved oxygen distribution index and the solid phase concentration profile to determine the matching deviation index.

[0008] Based on the oxygen demand index, dissolved oxygen distribution index, and matching deviation index, a zoned aeration control command is generated and output to the aeration actuator to adjust the zoned aeration volume. When the oxygen demand index and dissolved oxygen distribution index change in the same direction, the oxygen demand stability range is updated.

[0009] In one possible implementation, acquiring aerobic granular sludge image data and extracting particle size distribution features, surface roughness features, and boundary integrity features includes: denoising and normalizing the brightness of the aerobic granular sludge image data; performing particle segmentation on the denoised and normalized aerobic granular sludge image data to obtain particle segmentation results; and extracting particle size distribution features, surface roughness features, and boundary integrity features based on the particle segmentation results.

[0010] In one possible implementation, the stable oxygen demand range is determined based on historical operating data, and the stable oxygen demand range includes a lower limit and an upper limit for the oxygen demand index.

[0011] In one possible implementation, determining the oxygen demand index includes: determining the oxygen transfer resistance index based on the particle structure state index; determining the oxygen consumption intensity index based on multi-point dissolved oxygen measurement data; and determining the oxygen demand index based on the oxygen transfer resistance index and the oxygen consumption intensity index.

[0012] In one possible implementation, the oxygen consumption intensity index is determined by the rate of change of dissolved oxygen measurement data at multiple points within a preset sampling time window; the oxygen transfer resistance index is determined based on the mapping relationship between particle size distribution characteristics, surface roughness characteristics, and boundary integrity characteristics, and the mapping relationship is obtained through experimental calibration.

[0013] In one possible implementation, determining the dissolved oxygen distribution index includes: calculating the average dissolved oxygen value for each spatial partition; determining the dissolved oxygen concentration gradient value between spatial partitions based on the average dissolved oxygen value corresponding to each spatial partition; the dissolved oxygen distribution index includes the average dissolved oxygen value and the dissolved oxygen concentration gradient value.

[0014] In one possible implementation, obtaining the solid concentration profile includes: acquiring measurement data from a suspended solids concentration sensor arranged along the height direction of the reactor; and resampling the measurement data from the suspended solids concentration sensor according to the height coordinate to obtain the solid concentration profile.

[0015] In one possible implementation, obtaining the solid phase concentration profile includes: acquiring differential pressure measurement data from differential pressure sensors arranged along the height direction of the reactor; determining the liquid column density distribution based on the differential pressure measurement data; and converting the liquid column density distribution with a preset conversion relationship to obtain the solid phase concentration profile, wherein the preset conversion relationship is obtained by experimental calibration.

[0016] In one possible implementation, the partitioned correspondence and normalization process includes: performing correspondence processing on the solid phase concentration profile according to spatial partitions, and normalizing the correspondence-processed solid phase concentration profile to obtain a solid phase concentration distribution vector; normalizing the average dissolved oxygen value to obtain a dissolved oxygen distribution vector; determining the matching deviation index based on the sum of the absolute values ​​of the differences between the dissolved oxygen distribution vector and the solid phase concentration distribution vector in each spatial partition; and generating partitioned aeration control instructions when the matching deviation index is greater than the matching deviation threshold, including increasing the partitioned aeration rate of the spatial partition with a lower average dissolved oxygen value and decreasing the partitioned aeration rate of the spatial partition with a higher average dissolved oxygen value, wherein the matching deviation threshold is determined by historical operating data.

[0017] Secondly, the present invention provides an intelligent control system for an aerobic granular sludge process, for implementing an intelligent control method for the aerobic granular sludge process, the system comprising:

[0018] The structure extraction module is used to acquire aerobic granular sludge image data, extract particle size distribution features, surface roughness features, and boundary integrity features, and obtain particle structure state indicators.

[0019] The demand determination module is used to acquire multi-point dissolved oxygen measurement data and determine oxygen demand indicators based on particle structure state indicators, multi-point dissolved oxygen measurement data and oxygen demand stability intervals.

[0020] The deviation calculation module is used to divide the spatial region according to the location of the measurement point based on multi-point dissolved oxygen measurement data, determine the dissolved oxygen distribution index, obtain the solid phase concentration profile in the direction of reactor height, and perform partition correspondence and normalization processing between the dissolved oxygen distribution index and the solid phase concentration profile to determine the matching deviation index.

[0021] The aeration control module is used to generate zone aeration control commands based on oxygen demand indicators, dissolved oxygen distribution indicators, and matching deviation indicators, and output them to the aeration actuator to adjust the zone aeration volume. It also updates the oxygen demand stability range when the oxygen demand indicators and dissolved oxygen distribution indicators change in the same direction.

[0022] Compared with the prior art, the advantages and beneficial effects of the present invention are as follows:

[0023] By extracting particle image features and constructing particle structure state indicators, early characterization of mass transfer limitation trends was achieved; by jointly determining oxygen demand indicators based on multi-point dissolved oxygen and stable oxygen demand intervals, total oxygen supply demand could be quantified; by correlating dissolved oxygen zoning indicators with solid phase concentration profiles and calculating matching deviation indicators, verifiable identification of mismatch between oxygen supply distribution and solid phase stratification was achieved; and by generating stable interval updates under trend-consistent conditions through zoning aeration control commands triggered by matching deviations, closed-loop regulation of zoning aeration and adaptive adaptation of long-term operating boundaries were realized. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the execution flow of the method of the present invention;

[0025] Figure 2 This is a structural block diagram of the system of the present invention. Detailed Implementation

[0026] The embodiments of the present disclosure will now be described with reference to the accompanying drawings. However, it should be understood that these descriptions are exemplary only and are not intended to limit the scope of the disclosure. In the following detailed description, numerous specific details are set forth to provide a thorough understanding of the embodiments of the present disclosure for ease of explanation. However, it will be apparent that one or more embodiments may be practiced without these specific details. Furthermore, descriptions of well-known structures and techniques are omitted in the following description to avoid unnecessarily obscuring the concepts of the present disclosure.

[0027] The terminology used herein is for the purpose of describing particular embodiments only and is not intended to limit this disclosure. The terms “comprising,” “including,” etc., as used herein indicate the presence of the stated features, steps, operations, and / or components, but do not exclude the presence or addition of one or more other features, steps, operations, or components.

[0028] All terms used herein (including technical and scientific terms) have the meanings commonly understood by those skilled in the art, unless otherwise defined. It should be noted that the terms used herein are to be interpreted in a manner consistent with the context of this specification, and not in an idealized or overly rigid way.

[0029] Intelligent control typically refers to the linkage and organization of field sensing, state representation, decision calculation, and execution mechanisms into a repeatable and verifiable closed-loop control process within the framework of an industrial control system. Unlike relying solely on fixed setpoints or human experience, intelligent control emphasizes online identification of process states and dynamic maintenance of control boundaries. On the one hand, it utilizes multi-source online data to construct state indicators that reflect key mechanisms, expanding the controlled object from a "single measurement point" to a combined representation of "structural state and spatial distribution state." On the other hand, it forms verifiable judgment rules and control command generation logic based on state indicators, enabling the control strategy to adaptively adjust with changes in operating stages and disturbances, while maintaining safety boundaries and controllability. Based on these ideas, this invention targets aerobic granular sludge processes, constructing a closed-loop control path driven by image structural state, multi-point dissolved oxygen spatial distribution, and solid phase stratification information, and outputting zoned aeration control commands to achieve process regulation.

[0030] like Figure 1 As shown, an intelligent control method for an aerobic granular sludge process includes:

[0031] Acquire aerobic granular sludge image data, extract particle size distribution features, surface roughness features, and boundary integrity features to obtain particle structure state indicators;

[0032] In one embodiment, the system acquires aerobic granular sludge image data through the reactor observation window or sampling imaging chamber. During acquisition, the camera focal length and exposure parameters are fixed, and a continuous image sequence is formed according to a preset sampling period. For each frame, the granular region is first located, and then particle size distribution features, surface roughness features, and boundary integrity features are extracted. Particle size distribution features reflect the concentration and dispersion of particle diameter, surface roughness features reflect surface texture undulations, and boundary integrity features reflect contour gaps and signs of breakage. Sliding statistics are performed on the features from multiple frames and then fused to obtain granular structure state indicators. These granular structure state indicators serve as inputs for subsequent oxygen demand calculations and control decisions.

[0033] Acquiring aerobic granular sludge image data and extracting particle size distribution features, surface roughness features, and boundary integrity features includes: denoising and normalizing the brightness of the aerobic granular sludge image data; performing particle segmentation on the denoised and normalized aerobic granular sludge image data to obtain particle segmentation results; and extracting particle size distribution features, surface roughness features, and boundary integrity features based on the particle segmentation results.

[0034] In one embodiment, to ensure that particle size distribution characteristics, surface roughness characteristics, and boundary integrity characteristics remain comparable under different lighting and water turbidity conditions, noise reduction processing, brightness normalization processing, and particle segmentation processing are added after aerobic granular sludge image data are acquired.

[0035] Denoising is used to suppress the interference of bubble reflections and random noise on texture calculation; brightness normalization is used to reduce the overall brightness shift caused by light attenuation and viewport contamination; particle segmentation is used to separate particle regions from background regions, providing particle segmentation results for subsequent feature calculations. Denoising can employ any of the following: median filtering, bilateral filtering, or nonlocal mean filtering. The filter kernel size is set to a range of three to seven pixels based on the minimum resolvable diameter of the particles.

[0036] Brightness normalization first converts the image to grayscale or a brightness channel, then performs linear stretching based on the average brightness of the background region, and truncates saturated regions. The background region is obtained by reserving grain-free areas around the image or by detecting low-texture areas. Grain segmentation employs a combination of thresholding and morphological processing. Thresholding uses adaptive or Otsu thresholding to obtain an initial binary image; then, opening operations are performed to remove small-area noise, followed by closing operations to fill small holes; finally, connected components with areas smaller than a preset threshold are removed based on connected component analysis to obtain the grain segmentation result.

[0037] For adherent particles, a distance transform is calculated on the particle segmentation results, and watershed segmentation is performed to separate the contours of contacting particles. Each connected region after separation is treated as a particle sample for subsequent feature statistics. After obtaining the particle segmentation results, the particle size distribution characteristics are realized by calculating the equivalent diameter of each particle and forming a diameter histogram. The equivalent diameter is obtained by converting the particle area. At the same time, the average diameter, median diameter, and coefficient of variation are calculated to reflect the overall level and dispersion of particle size.

[0038] Surface roughness features are used to calculate the statistical measures of gray-level gradient amplitude and local contrast within the particle region. These statistics include the mean, standard deviation, and quantiles. Boundary integrity features are determined by the relationship between the contour perimeter and area, the proportion of abrupt changes in contour curvature, and the proportion of contour gap lengths. The contour gap is obtained by the difference between the contour and the contour convex hull. Particle structure state indices are obtained by fusing particle size distribution features, surface roughness features, and boundary integrity features according to preset weights, which are determined through experimental calibration. During experimental calibration, particle samples from different operating stages are selected, and the corresponding image features and operating phenomenon labels are recorded. Based on this, the contribution direction and weight range of each feature to the particle structure state indices are determined. The particle structure state indices are updated using a sliding time window, with the time window length set to one to ten minutes based on the aeration adjustment response time.

[0039] Obtain dissolved oxygen measurement data from multiple points, and determine oxygen demand indicators based on particle structure state indicators, multi-point dissolved oxygen measurement data, and oxygen demand stability intervals;

[0040] In one embodiment, multi-point dissolved oxygen measurement data is simultaneously collected by dissolved oxygen sensors at different spatial locations within the reactor, with the acquisition period aligned with the image acquisition period or using timestamp alignment. The collected data undergoes outlier removal and short-term smoothing to generate a dissolved oxygen time-series curve suitable for calculation. The system uses particle structure state indicators as input to the mass transfer state, the dissolved oxygen time-series curve as input to the process response, and combines this with the oxygen demand stability interval to determine the current operating state, thus obtaining an oxygen demand index. The oxygen demand index characterizes the oxygen supply level required to maintain the target operating state under the current particle structure and load conditions, and provides a baseline for subsequent zoned aeration control.

[0041] The stable oxygen demand range is determined based on historical operating data and includes the lower limit and upper limit of the oxygen demand index.

[0042] In one embodiment, to ensure comparable boundaries for oxygen demand indicators, the stable oxygen demand range is determined based on historical operating data. This historical operating data can be derived from records of the same reactor during its stable operation phase. The stable operation phase is selected based on objective conditions such as effluent quality meeting standards, minimal dissolved oxygen fluctuations, and low frequency of aeration adjustments. During the selection process, candidate intervals are selected based on periods where the dissolved oxygen time-series curve remains within a stable fluctuation range without significant drift over multiple consecutive sampling periods, while periods of sensor calibration, maintenance, or sudden changes in influent load are excluded.

[0043] Statistical analysis of oxygen demand indicators within the candidate interval yields a typical range for these indicators. The stable oxygen demand interval is defined as the range between the lower and upper limits of the oxygen demand indicators. The lower limit defines the boundary of insufficient oxygen supply risk, while the upper limit defines the boundary of over-aeration risk. The lower and upper limits can be determined using quantiles, for example, using the lower quantile of the candidate interval as the lower limit and the higher quantile as the upper limit, to avoid a small number of outliers affecting the interval boundaries; alternatively, they can be determined by adding or subtracting a multiple of the standard deviation from the mean, to accommodate scenarios with approximately symmetrical data distributions.

[0044] Once the stable oxygen demand range is established, the coverage of historical operating data can be reviewed periodically during operation. The stable oxygen demand range is then slowly updated when the stable operation criteria are met; otherwise, it remains unchanged, avoiding the introduction of unreasonable boundaries during periods of fluctuation. Through this method, the stable oxygen demand range is aligned with reactor operating conditions, sensor ranges, and process objectives, providing clear upper and lower limits for determining oxygen demand indicators.

[0045] Determining oxygen demand indicators includes: determining oxygen transfer resistance indicators based on particle structure state indicators; determining oxygen consumption intensity indicators based on multi-point dissolved oxygen measurement data; and determining oxygen demand indicators based on both oxygen transfer resistance indicators and oxygen consumption intensity indicators.

[0046] In one embodiment, the oxygen demand index is jointly determined by the oxygen transport resistance index and the oxygen consumption intensity index. The oxygen transport resistance index characterizes the degree to which the particle structure hinders oxygen transport, while the oxygen consumption intensity index characterizes the intensity of dissolved oxygen consumption per unit time. The oxygen transport resistance index is derived from the particle structure index, and its derivation logic is related to particle size distribution characteristics, surface roughness characteristics, and boundary integrity characteristics: when the average particle size is large, the surface roughness is reduced, and the boundary integrity is decreased, the oxygen transport path is more easily restricted, and the oxygen transport resistance index increases; when the particle size is more uniform and the boundary integrity is better, the oxygen transport resistance index decreases. The oxygen consumption intensity index is derived from multi-point dissolved oxygen measurement data and reflects the strength of the dissolved oxygen decreasing trend within a preset sampling time window. The larger the oxygen consumption intensity index, the stronger the oxygen consumption by the biological reaction.

[0047] The oxygen demand index is obtained by fusing the oxygen transport resistance index and the oxygen consumption intensity index. This fusing can be linear or regular. In the case of linear fusing, the oxygen demand index satisfies:

[0048]

[0049] in, As an indicator of oxygen demand, As an indicator of oxygen transport resistance, As an indicator of oxygen consumption intensity, As the weight for oxygen transport resistance, The weight is the oxygen consumption intensity.

[0050] Weights can be obtained through experimental calibration or slowly self-tuned during the stable operation phase based on the control effect. However, the update rate of self-tuning is limited by a preset upper limit to avoid oscillations. After obtaining the oxygen demand index, it is compared with the stable oxygen demand range: when the oxygen demand index is below the lower limit, it is determined that there is a trend of insufficient oxygen supply; when the oxygen demand index is above the upper limit, it is determined that there is a trend of excessive oxygen supply; when the oxygen demand index is within the stable oxygen demand range, it is determined that the oxygen supply demand is within an acceptable range. By simultaneously incorporating particle structure state indicators and multi-point dissolved oxygen measurement data, the oxygen demand index can distinguish between "decreased oxygen supply efficiency due to structural mass transfer limitation" and "increased oxygen demand due to enhanced biochemical oxygen consumption," providing a more interpretable basis for subsequent control strategies.

[0051] The oxygen consumption intensity index is determined by the rate of change of dissolved oxygen measurement data at multiple points within a preset sampling time window; the oxygen transfer resistance index is determined based on the mapping relationship between particle size distribution characteristics, surface roughness characteristics and boundary integrity characteristics, and the mapping relationship is obtained by experimental calibration.

[0052] In one embodiment, the oxygen consumption intensity index is determined by the rate of change of dissolved oxygen measurement data at multiple points within a preset sampling time window. The preset sampling time window is selected to cover the time length of a stable aeration control action response, such as 30 seconds to 5 minutes.

[0053] To ensure that the rate of change over time reflects the true oxygen consumption trend, the control state of the aeration actuator is kept unchanged or allowed only small fluctuations within the preset sampling time window, and time periods with relatively linear changes in dissolved oxygen are preferentially selected. The rate of change over time can be calculated using linear fitting: least squares fitting is performed on the dissolved oxygen sampling points within the preset sampling time window, and the absolute value of the fitting slope is used as the oxygen consumption intensity index; when the dissolved oxygen sampling points contain peaks caused by bubble interference, the sampling points are first subjected to median filtering or amplitude limiting before fitting, in order to avoid distortion of the oxygen consumption intensity index due to single-point anomalies.

[0054] When dissolved oxygen (DO) measurements are taken from different spatial locations, the oxygen consumption intensity index can be calculated as a weighted average of the oxygen consumption intensities at each measurement point. The weighting coefficient is related to the aeration influence range of the spatial partition where the measurement point is located, so that the oxygen consumption intensity index more closely reflects the overall process oxygen consumption level. The oxygen transfer resistance index is determined based on the mapping relationship between particle size distribution characteristics, surface roughness characteristics, and boundary integrity characteristics. This mapping relationship is obtained through experimental calibration. During experimental calibration, samples with different particle states are selected, and dissolved oxygen rise curves or dissolved oxygen recovery curves are recorded under the same aeration intensity and the same liquid volume conditions. The recovery rate of the curve characterizes the degree of oxygen transfer limitation. Simultaneously, corresponding particle images are acquired, and particle size distribution characteristics, surface roughness characteristics, and boundary integrity characteristics are calculated to establish a mapping relationship between image features and the degree of oxygen transfer limitation.

[0055] The mapping relationship can be established using either a lookup table or a regression model. The lookup table method discretizes the feature space into several intervals and records the corresponding resistance levels, while the regression model method outputs continuous resistance levels and limits the output range to avoid exceeding physically reasonable limits. During operation, when imaging conditions change, the bias or scale term of the mapping relationship is updated using a small number of calibration samples to ensure that the oxygen transfer resistance index remains consistent with the actual mass transfer response. Through these methods, both the oxygen consumption intensity index and the oxygen transfer resistance index have clear data sources, calculation paths, and calibration paths, enabling stable calculation of the oxygen demand index on-site and supporting subsequent closed-loop control.

[0056] Based on multi-point dissolved oxygen measurement data, spatial partitions are divided according to the location of the measurement points to determine the dissolved oxygen distribution index, obtain the solid phase concentration profile in the direction of reactor height, and perform partition correspondence and normalization processing on the dissolved oxygen distribution index and the solid phase concentration profile to determine the matching deviation index.

[0057] Multiple dissolved oxygen sensors are arranged within the reactor, dividing the reactor into spatial zones according to their spatial locations. After collecting dissolved oxygen measurement data from multiple points, the spatial zones are assigned, and then the dissolved oxygen distribution index is calculated. Simultaneously, the solid phase concentration profile along the reactor's height is acquired, and the dissolved oxygen distribution index is correlated with the solid phase concentration profile within the same spatial zone. The two types of distribution data are then normalized to obtain comparable distribution sequences. Based on these two distribution sequences, a matching deviation index is calculated. This index characterizes the degree of mismatch between the oxygen supply spatial distribution and the solid phase stratification distribution, serving as a key input for subsequent zoned aeration adjustment.

[0058] Determining dissolved oxygen distribution indicators includes: calculating the average dissolved oxygen value for each spatial zone; determining the dissolved oxygen concentration gradient value between spatial zones based on the average dissolved oxygen value for each spatial zone; the dissolved oxygen distribution indicators include the average dissolved oxygen value and the dissolved oxygen concentration gradient value.

[0059] In one embodiment, the newly added limiting point in this paragraph clarifies the calculation path of the dissolved oxygen distribution index. This limiting point transforms multi-point dissolved oxygen measurement data from a "set of point values" into a "regional index" that can be directly used for comparison in spatial partitioning.

[0060] In practice, the mapping relationship between sensor coordinates and spatial partitions is first established. This mapping is completed during the installation and commissioning phase, recording the height, radial position, and circumferential position of each sensor, and determining the spatial partition number accordingly. During sampling, the dissolved oxygen measurements from each sensor are aggregated according to the spatial partition number, resulting in a dissolved oxygen sampling sequence for each spatial partition. To reduce occasional bubble interference, outlier points are first removed from the dissolved oxygen sampling sequence. Outlier removal can employ a limiting rule, with the limiting threshold determined by the sensor range and normal fluctuation range; alternatively, a three-point median rule can be used, replacing spikes with the median of neighboring points. Subsequently, the average dissolved oxygen value for each spatial partition is calculated. The average dissolved oxygen value is calculated using the average within the same sampling time window, which is consistent with the control cycle, typically ranging from thirty seconds to two minutes.

[0061] If a spatial partition has missing values ​​within the sampling time window, the minimum number of valid sampling points is used as a condition; spatial partitions that do not meet this condition are not included in the gradient calculation and are marked as spatial partitions to be supplemented. The dissolved oxygen concentration gradient value between spatial partitions is used to reflect the degree of difference in spatial distribution. The gradient value is calculated based on adjacent spatial partitions, and the adjacency relationship of spatial partitions is determined by the reactor structure, usually adjacent along the height direction or along the circumferential direction. The gradient value can be the absolute value of the difference between the average dissolved oxygen values ​​of adjacent partitions, or the sign can be retained to indicate the gradient direction. To avoid the gradient being amplified by fluctuations in a single partition, the average dissolved oxygen value is lightly smoothed before the gradient value is calculated, with a smoothing length of two to three frames. The dissolved oxygen distribution index is composed of the average dissolved oxygen value and the dissolved oxygen concentration gradient value. The average dissolved oxygen value is used to describe the oxygen supply level of the partition, and the dissolved oxygen concentration gradient value is used to describe the degree of spatial non-uniformity. In the subsequent matching deviation calculation, the two serve as the information sources for "horizontal alignment" and "difference alignment" respectively, thereby avoiding the masking of local hypoxia problems by relying solely on a single average value.

[0062] Obtaining the solid concentration profile includes: acquiring measurement data from suspended solids concentration sensors arranged along the height of the reactor; and resampling the measurement data from the suspended solids concentration sensors according to the height coordinate to obtain the solid concentration profile.

[0063] In one embodiment, the newly added limiting point in this paragraph is to provide a direct measurement and resampling path for the solid phase concentration profile. This limiting point moves solid phase stratification from a conceptual description to an engineering-deployable sensor configuration and data processing method. Specifically, multiple suspended solids concentration sensors are arranged along the reactor height. The sensor arrangement height is aligned with or covers the center height of the spatial partitions to ensure subsequent partitioning correspondence.

[0064] Suspended solids concentration data are obtained at each height point during sampling. To reduce drift caused by sludge buildup on the sensor surface, the suspended solids concentration sensor is equipped with an automatic scraper or air-washing structure and a fixed cleaning cycle is set; the cleaning cycle can be staggered with the aeration switching cycle to avoid simultaneous disturbance. The collected suspended solids concentration measurement data is first corrected for zero-point drift. Drift correction is completed by taking the average value of the stable segment shortly after cleaning and comparing it with the calibration value. The correction amount is updated slowly to avoid overcompensation due to a single fluctuation. Subsequently, a solids concentration profile is formed.

[0065] The solid concentration profile uses height coordinates as the independent variable and suspended solids concentration as the dependent variable. Since sensor height points are typically discrete, resampling is required. Resampling uses interpolation to convert discrete height points into a sequence with a fixed height step size. Piecewise linear interpolation can be used, which is easy to implement and does not introduce excessive smoothing; when the height point spacing is large, piecewise cubic interpolation can be used, but the interpolated output must be limited to the range of adjacent measurements to avoid unreasonable overshoot.

[0066] The resampled solids concentration profiles are then converged along the spatial partition boundaries to obtain representative solids concentration values ​​for each spatial partition. These representative values ​​can be the average of the resampled points within the partition, or the median to enhance robustness. Through this approach, the solids concentration profile can stably reflect solids stratification along the reactor height and is aligned with the dissolved oxygen partition index on the same spatial partition scale, providing a consistent data basis for subsequent matching bias calculations.

[0067] Obtaining the solid phase concentration profile includes: acquiring differential pressure measurement data from differential pressure sensors arranged along the height of the reactor; determining the liquid column density distribution based on the differential pressure measurement data; and converting the liquid column density distribution with a preset conversion relationship to obtain the solid phase concentration profile, wherein the preset conversion relationship is obtained through experimental calibration.

[0068] In one embodiment, the newly added limiting point in this paragraph is to provide a path for obtaining the solid phase concentration profile based on differential pressure measurement. This limiting point provides an alternative implementation for scenarios where it is not possible to directly deploy suspended solids concentration sensors. Specifically, multiple sets of differential pressure sensor taps are arranged along the height of the reactor.

[0069] Pressure taps are paired to form differential pressure measurement channels across multiple height ranges. Differential pressure measurement data for each height range are obtained through sampling. To reduce the impact of pressure tap blockage, the pressure tapping pipeline is periodically backflushed, and a self-checking rule for the differential pressure channels is set; when the differential pressure change is significantly inconsistent with adjacent channels, the channel is marked as an abnormal channel and temporarily excluded from calculation. The liquid column density distribution is determined based on the differential pressure measurement data.

[0070] The liquid column density distribution is calculated using the static pressure difference relationship. Given the length of the height interval, the pressure difference is converted into the average liquid density of the corresponding interval. Since temperature affects the liquid density and the sensor zero point, the water temperature is recorded simultaneously during sampling, or a temperature compensation coefficient is used to correct the density results. After obtaining the liquid column density distribution, a preset conversion relationship is used to convert it into a solid phase concentration profile.

[0071] The preset conversion relationship is obtained through experimental calibration. During the calibration phase, differential pressure data and suspended solids concentration data measured in the laboratory are collected under different solid concentration conditions to establish the correspondence between density and solid concentration. The correspondence can be established by using a piecewise linear lookup table, with the lookup points covering common operating concentration ranges; alternatively, a low-order regression curve can be used, but the curve output must have upper and lower limits set to prevent unreasonable extrapolation beyond the calibration range.

[0072] During operation, solid phase concentration profiles are output according to height intervals and can be converged according to spatial partition boundaries to obtain representative solid phase concentration values ​​for each spatial partition. Through this alternative path, solid phase stratification can be continuously estimated without adding complex equipment, while maintaining a spatial scale consistent with dissolved oxygen partition indicators, facilitating subsequent matching deviation calculations and control decisions.

[0073] The partitioned mapping and normalization process includes: mapping the solid phase concentration profile according to spatial partitions, and normalizing the mapped solid phase concentration profiles to obtain a solid phase concentration distribution vector; normalizing the average dissolved oxygen value to obtain a dissolved oxygen distribution vector; determining the matching deviation index based on the sum of the absolute values ​​of the differences between the dissolved oxygen distribution vector and the solid phase concentration distribution vector in each spatial partition; and generating partitioned aeration control instructions when the matching deviation index exceeds the matching deviation threshold, including increasing the aeration rate of the spatial partition with a lower average dissolved oxygen value and decreasing the aeration rate of the spatial partition with a higher average dissolved oxygen value, wherein the matching deviation threshold is determined by historical operating data.

[0074] In one embodiment, the newly added limiting point in this paragraph is to provide the calculation method for partition correspondence, normalization processing, and matching deviation index, and to clarify the source and purpose of the matching deviation threshold. This limiting point transforms the "mismatch degree" from a qualitative judgment to a reproducible calculation result and limits the subsequent control triggering conditions. In specific implementation, partition correspondence processing is completed first. Partition correspondence processing uses the spatial partition number as an index to convert the solid phase concentration profile into a solid phase concentration sequence that corresponds one-to-one with the spatial partition, and converts the average dissolved oxygen value in the dissolved oxygen distribution index into the same spatial partition sequence. To ensure comparability, normalization processing is performed on both types of sequences. Normalization processing adopts interval normalization, and the upper and lower limits of the interval are determined by historical stable operating data or by the sensor range and operating experience range. After normalization, the dissolved oxygen distribution vector and the solid phase concentration distribution vector are obtained.

[0075] The matching deviation index is calculated using the difference between the two vectors in each spatial partition. The matching deviation index can be calculated using the following formula:

[0076]

[0077] in, To match the deviation index, The number of spatial partitions, For the first Normalized average dissolved oxygen for each spatial partition For the first The normalized solid concentration representative value for each spatial partition. This calculation method is intuitive, feasible, and easy for engineering parameter tuning. The matching deviation threshold is determined by historical operating data.

[0078] When determining the optimal operating period, a stable operating timeframe is selected, and the corresponding matching deviation index sequence is calculated. The high quantile is used as a threshold to accommodate normal fluctuations. During operation, when the matching deviation index exceeds the matching deviation threshold, the zone aeration redistribution logic is triggered. This redistribution logic is based on the average dissolved oxygen (DOO) value, increasing the aeration rate of zones with lower DOO values ​​and decreasing the aeration rate of zones with higher DOO values. The adjustment range uses a tiered step size, determined by the minimum controllable flow rate of the aeration actuator. Upper and lower limits for zone aeration rates are also set to avoid excessive adjustments that could cause drastic fluctuations in DOO. Through this process, the matching deviation index not only reflects the deviation between oxygen supply distribution and solid-phase stratification but also forms clear triggering conditions and executable adjustment actions, providing a stable intermediate value for subsequent closed-loop control.

[0079] Based on the oxygen demand index, dissolved oxygen distribution index, and matching deviation index, a zoned aeration control command is generated and output to the aeration actuator to adjust the zoned aeration volume. When the oxygen demand index and dissolved oxygen distribution index change in the same direction, the oxygen demand stability range is updated.

[0080] In one embodiment, the system generates zone aeration control commands at a fixed control cycle and outputs these commands to the aeration actuators to adjust the aeration rate of each spatial zone. The control cycle is consistent with the sampling cycle of the multi-point dissolved oxygen measurement data, typically ranging from thirty seconds to two minutes.

[0081] When generating zoned aeration control commands, the oxygen demand index is used as the basis for determining the total oxygen supply demand, the dissolved oxygen distribution index is used as the basis for determining the spatial distribution status, and the matching deviation index is used as the basis for determining whether zoned redistribution is triggered. First, the direction of total aeration adjustment is determined based on the relationship between the oxygen demand index and the stable oxygen demand range: when the oxygen demand index is higher than the upper limit of the stable oxygen demand range, the total aeration volume is increased; when the oxygen demand index is lower than the lower limit of the stable oxygen demand range, the total aeration volume is decreased; when the oxygen demand index is within the stable oxygen demand range, the total aeration volume remains unchanged or undergoes only minor adjustments.

[0082] The total aeration volume is adjusted using a tiered step size, determined by the minimum controllable flow rate of the aeration actuator. An upper and lower limit for the total aeration volume are set to avoid exceeding equipment capacity or falling below the minimum aeration volume required to maintain mixing. Subsequently, the initial allocation of aeration volume to each zone is determined based on dissolved oxygen distribution indicators. The initial allocation uses the average dissolved oxygen value of each spatial zone as the core input: spatial zones with lower average dissolved oxygen values ​​are allocated a higher aeration ratio, while those with higher average dissolved oxygen values ​​are allocated a lower aeration ratio. Simultaneously, the degree of spatial non-uniformity is assessed using dissolved oxygen concentration gradient values. When the dissolved oxygen concentration gradient value between spatial zones increases, the sensitivity of zone adjustment is increased, causing aeration to more quickly shift towards areas with low dissolved oxygen.

[0083] After the initial allocation is completed, it is determined whether zonal reallocation is necessary. The trigger condition for zonal reallocation is given by the matching deviation index. When the matching deviation index is greater than the matching deviation threshold, it is considered that there is a significant mismatch between dissolved oxygen distribution and solid phase stratification, and zonal reallocation is triggered. When the matching deviation index is not greater than the matching deviation threshold, the initial allocation is maintained. When zonal reallocation is executed, the principle of "prioritizing compensation for low average dissolved oxygen values" is followed, increasing the aeration rate of spatial zones with lower average dissolved oxygen values ​​and decreasing the aeration rate of spatial zones with higher average dissolved oxygen values. To avoid overcompensation for a single spatial zone, the variation range of aeration rate for each spatial zone within a single cycle is limited to a preset upper limit, and the variation range of all spatial zones is normalized to keep the total aeration rate consistent with the aforementioned total aeration adjustment.

[0084] After the zone aeration control command is generated, it is output to the aeration actuator. The aeration actuator may include a blower and zone aeration valve groups. The zone aeration control command is used to set the blower output level or output flow rate, and to set the opening degree or distribution ratio of each zone aeration valve group. To ensure the verifiability of the control action, the system records the oxygen demand index, dissolved oxygen distribution index, and matching deviation index within the control cycle after each zone aeration control command is output, and records the set value and execution feedback value of the zone aeration volume. When the deviation between the execution feedback value and the set value exceeds the allowable range, it is marked as an execution anomaly, and the zone aeration volume setting of the previous cycle is maintained to avoid further amplification and adjustment when the actuator is in an abnormal state.

[0085] In this embodiment, the updating of the oxygen demand stable range is decoupled from the control action. The update is triggered only when the consistency condition is met, to avoid changing the boundary during unstable phases. Specifically, the consistency condition is defined as the oxygen demand index and the dissolved oxygen distribution index changing in the same direction over multiple consecutive control cycles. The determination of the consistent direction of change is based on the sign of the difference before and after the control cycle: when the oxygen demand index rises and the dissolved oxygen distribution index rises overall, or when the oxygen demand index falls and the dissolved oxygen distribution index falls overall, the direction of change is considered consistent; the overall change of the dissolved oxygen distribution index is represented by the change in the average or median of the dissolved oxygen average values ​​of each spatial partition.

[0086] The number of consecutive control cycles is determined by a preset number of consecutive samplings, typically three to ten. When the consistency condition is met, the system is considered to be in a stable response range, and an oxygen demand stable range update is initiated. During the update, an oxygen demand index sequence within a recent stable data window is selected, and the lower and upper limits of the oxygen demand index are recalculated, replacing the original boundaries with a slow update. The slow update uses a weighted update, merging the new and old boundaries according to a preset ratio to avoid frequent switching of subsequent judgments due to abrupt boundary changes. If the consistency condition is not met, the oxygen demand stable range is not updated, and only the existing boundaries continue to be used for oxygen demand index judgment. Through the above control and update mechanism, total aeration adjustment, zone aeration allocation, mismatch-triggered redistribution, and stable range update form a closed loop: the control output changes the dissolved oxygen distribution, the matching status of the dissolved oxygen distribution and solid phase stratification is fed back to the zone redistribution decision of the next cycle through the matching deviation index, and the oxygen demand stable range is self-updated during the stable response phase to adapt to long-term operational drift, thereby ensuring the continuous availability of the control strategy in the engineering field.

[0087] like Figure 2 As shown, an intelligent control system for an aerobic granular sludge process is provided, which is used to implement an intelligent control method for the aerobic granular sludge process. The system includes:

[0088] The structure extraction module acquires image data of aerobic granular sludge, extracts particle size distribution features, surface roughness features, and boundary integrity features to obtain particle structure state indicators. The module consists of an industrial camera, lens, supplementary lighting source, imaging window or sampling imaging cavity, and edge computing unit. The industrial camera acquires granular sludge image data, the supplementary lighting source provides stable illumination, the imaging window isolates water bodies and ensures a stable field of view, and the edge computing unit performs image preprocessing, particle segmentation, and feature calculation, outputting the particle structure state indicators to the control network.

[0089] The demand determination module acquires multi-point dissolved oxygen measurement data and determines the oxygen demand index based on particle structure state indicators, multi-point dissolved oxygen measurement data, and oxygen demand stability intervals. The module consists of a multi-point dissolved oxygen sensor array, a signal conditioning and acquisition unit, a clock synchronization unit, and a control calculation unit. The multi-point dissolved oxygen sensor array acquires multi-point dissolved oxygen measurement data; the signal conditioning and acquisition unit performs isolation, filtering, and analog-to-digital conversion; the clock synchronization unit ensures multi-channel sampling time alignment; and the control calculation unit calculates the oxygen demand index by combining particle structure state indicators and oxygen demand stability intervals, and outputs the oxygen demand index to the control logic.

[0090] The deviation calculation module is used to divide spatial zones based on multi-point dissolved oxygen measurement data according to the location of the measurement points, determine the dissolved oxygen distribution index, obtain the solid phase concentration profile along the height of the reactor, and perform partition mapping and normalization processing on the dissolved oxygen distribution index and the solid phase concentration profile to determine the matching deviation index. The deviation calculation module consists of a spatial partition mapping storage unit, solid phase concentration measurement hardware, and a data fusion calculation unit. The spatial partition mapping storage unit is used to store the correspondence between the measurement point locations and the spatial partitions; the solid phase concentration measurement hardware is used to obtain the solid phase concentration profile along the height of the reactor, and the solid phase concentration measurement hardware can be a suspended solids concentration sensor array or a differential pressure sensor array; the data fusion calculation unit is used to complete the calculation of dissolved oxygen distribution index, solid phase concentration profile partition mapping, normalization processing, and matching deviation index calculation, and outputs the matching deviation index to the control module.

[0091] The aeration control module generates zoned aeration control commands based on oxygen demand, dissolved oxygen distribution, and matching deviation indicators, and outputs them to the aeration actuators to adjust the zoned aeration volume. It also updates the oxygen demand stability range when the oxygen demand and dissolved oxygen distribution indicators change in the same direction. The aeration control module consists of a programmable logic controller (PLC) or industrial control computer, zoned aeration valve assemblies, a blower frequency converter, and an execution feedback acquisition unit. The PLC or industrial control computer generates zoned aeration control commands based on the oxygen demand, dissolved oxygen distribution, and matching deviation indicators. The zoned aeration valve assemblies execute zoned flow distribution. The blower frequency converter adjusts the total air supply. The execution feedback acquisition unit collects valve position, flow rate, or pressure feedback and triggers the calculation and storage of the oxygen demand stability range update when the oxygen demand and dissolved oxygen distribution indicators change in the same direction.

[0092] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects.

[0093] The above are merely embodiments of the present invention and are not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principle of the present invention should be included within the scope of the claims of the present invention.

Claims

1. An intelligent control method for an aerobic granular sludge process, characterized in that, The method includes: Acquire aerobic granular sludge image data, extract particle size distribution features, surface roughness features, and boundary integrity features to obtain particle structure state indicators; Acquire multi-point dissolved oxygen measurement data, and determine the oxygen demand index based on the particle structure state index, the multi-point dissolved oxygen measurement data and the oxygen demand stability interval. The oxygen demand stability interval is determined based on historical operating data and includes a lower limit value and an upper limit value of the oxygen demand index. Based on the multi-point dissolved oxygen measurement data, spatial partitions are divided according to the location of the measurement points to determine the dissolved oxygen distribution index, obtain the solid phase concentration profile in the direction of reactor height, and perform partition correspondence and normalization processing on the dissolved oxygen distribution index and the solid phase concentration profile to determine the matching deviation index. Based on the oxygen demand index, the dissolved oxygen distribution index, and the matching deviation index, a zone aeration control command is generated and output to the aeration execution device to adjust the zone aeration volume. When the oxygen demand index and the dissolved oxygen distribution index change in the same direction, the oxygen demand stability range is updated. Determining the dissolved oxygen distribution index includes: Calculate the average dissolved oxygen value for each of the aforementioned spatial partitions; The dissolved oxygen concentration gradient value between spatial zones is determined based on the average dissolved oxygen value corresponding to each spatial zone. The dissolved oxygen distribution index includes the average dissolved oxygen value and the dissolved oxygen concentration gradient value; The partition mapping and normalization process includes: The solid concentration profile is processed according to the spatial partition, and the processed solid concentration profile is normalized to obtain the solid concentration distribution vector. The average dissolved oxygen value is normalized to obtain the dissolved oxygen distribution vector; The matching deviation index is determined based on the sum of the absolute values ​​of the differences between the dissolved oxygen distribution vector and the solid phase concentration distribution vector in each of the spatial partitions. Based on the oxygen demand index, dissolved oxygen distribution index and matching deviation index, a partition aeration control command is generated and output to the aeration execution device. When generating zoned aeration control commands, the oxygen demand index is used as the basis for determining the total oxygen supply demand, the dissolved oxygen distribution index is used as the basis for determining the spatial distribution status, and the matching deviation index is used as the basis for determining whether zoned redistribution is triggered. First, the direction of total aeration adjustment is determined based on the relationship between the oxygen demand index and the stable oxygen demand range. The total aeration volume is adjusted using a stepped approach. Then, the initial allocation of aeration volume for each zone is determined based on the dissolved oxygen distribution index. The initial allocation uses the average dissolved oxygen value of each spatial zone as the core input: spatial zones with lower average dissolved oxygen values ​​are allocated a higher aeration ratio, while spatial zones with higher average dissolved oxygen values ​​are allocated a lower aeration ratio. At the same time, the degree of spatial non-uniformity is judged by the dissolved oxygen concentration gradient value. When the dissolved oxygen concentration gradient value between spatial zones increases, the sensitivity of the zone adjustment is improved, so that aeration is tilted towards the low dissolved oxygen area more quickly. After the initial allocation is completed, it is determined whether partition reallocation is required. The triggering condition for partition reallocation is given by the matching deviation index. When the matching deviation index is greater than the matching deviation threshold, it is considered that there is a significant mismatch between dissolved oxygen distribution and solid phase stratification, and partition reallocation is triggered; when the matching deviation index is not greater than the matching deviation threshold, the initial allocation is maintained. The zone aeration control command includes increasing the zone aeration rate of the spatial zone with a lower average dissolved oxygen value and decreasing the zone aeration rate of the spatial zone with a higher average dissolved oxygen value within the upper and lower limits of the zone aeration rate, wherein the matching deviation threshold is determined by historical operating data.

2. The method according to claim 1, characterized in that, Acquiring the aerobic granular sludge image data and extracting the particle size distribution features, surface roughness features, and boundary integrity features includes: The aerobic granular sludge image data was subjected to noise reduction and brightness normalization. The aerobic granular sludge image data, after noise reduction and brightness normalization, is subjected to particle segmentation processing to obtain particle segmentation results. Based on the particle segmentation results, the particle size distribution features, the surface roughness features, and the boundary integrity features are extracted.

3. The method according to claim 1, characterized in that, Determining the oxygen demand index includes: The oxygen transport resistance index is determined based on the particle structure state index. The oxygen consumption intensity index is determined based on the multi-point dissolved oxygen measurement data. The oxygen demand index is determined based on the oxygen transport resistance index and the oxygen consumption intensity index. After obtaining the oxygen demand index, the oxygen demand index is compared with the stable oxygen demand range: when the oxygen demand index is lower than the lower limit of the oxygen demand index, it is determined that there is a trend of insufficient oxygen supply; when the oxygen demand index is higher than the upper limit of the oxygen demand index, it is determined that there is a trend of excessive oxygen supply; when the oxygen demand index is within the stable oxygen demand range, it is determined that the oxygen supply demand is within an acceptable range.

4. The method according to claim 3, characterized in that, The oxygen consumption intensity index is determined by the rate of change of the multi-point dissolved oxygen measurement data over a preset sampling time window. The oxygen transport resistance index is determined based on the mapping relationship between the particle size distribution characteristics, the surface roughness characteristics, and the boundary integrity characteristics, and the mapping relationship is obtained through experimental calibration.

5. The method according to claim 1, characterized in that, Obtaining solid concentration profiles includes: Acquire measurement data from suspended solids concentration sensors arranged along the height of the reactor; The measurement data from the suspended solids concentration sensor are resampled according to the height coordinate to obtain the solid phase concentration profile.

6. The method according to claim 1, characterized in that, Obtaining solid concentration profiles includes: Acquire differential pressure measurement data from differential pressure sensors arranged along the height of the reactor; The liquid column density distribution is determined based on the differential pressure measurement data, and the solid phase concentration profile is obtained by converting the liquid column density distribution with a preset conversion relationship, wherein the preset conversion relationship is obtained by experimental calibration.

7. An intelligent control system for an aerobic granular sludge process, used to implement the intelligent control method for the aerobic granular sludge process according to any one of claims 1-6, characterized in that, The system includes: The structure extraction module is used to acquire aerobic granular sludge image data, extract particle size distribution features, surface roughness features, and boundary integrity features, and obtain particle structure state indicators. The demand determination module is used to acquire multi-point dissolved oxygen measurement data and determine the oxygen demand index based on the particle structure state index, the multi-point dissolved oxygen measurement data and the oxygen demand stability interval. The deviation calculation module is used to divide the space into spatial partitions according to the location of the measurement points based on the multi-point dissolved oxygen measurement data, determine the dissolved oxygen distribution index, obtain the solid phase concentration profile in the direction of reactor height, and perform partition correspondence and normalization processing on the dissolved oxygen distribution index and the solid phase concentration profile to determine the matching deviation index. The aeration control module is used to generate zone aeration control commands based on the oxygen demand index, the dissolved oxygen distribution index, and the matching deviation index, and output them to the aeration execution device to adjust the zone aeration volume, and update the oxygen demand stability range when the oxygen demand index and the dissolved oxygen distribution index change in the same direction.

Images