Partitioned stepping backwashing system and method for rotary filter screen

Abstract

The invention discloses a partitioned stepping backwashing system and method for a rotary filter screen. The partitioned stepping backwashing system comprises a state monitoring module, a blockage characteristic quantification module, a state judgment module, a washing mode selection module and a multi-drive washing module, the state monitoring module is used for collecting vibration spectrum data and visual image data of each partition uniformly divided along the circumference of the rotary filter screen body and pressure difference data of the rotary filter screen body in real time; the blockage feature quantification module can extract a vibration disorder value in the vibration spectrum data and a contrast ratio and a correlation value in the visual image data, and then combines a pressure difference value Y; and the state judgment module can map the multi-dimensional feature array to a high-dimensional space through a classification model. Precise positioning and targeted washing of all the subareas of the rotary filter screen body can be achieved, and the problem that due to excessive washing of a clean area, the area is seriously blocked and cannot be cleaned in time is effectively solved.

Description

Technical Field

[0001] This invention relates to the field of water treatment technology, and specifically to a zoned step-backwashing system and method for a rotating filter screen. Background Technology

[0002] A rotary filter screen for power plant circulating water is a device used to filter impurities in a power plant's circulating water system. It is typically installed at the inlet or outlet of the circulating water system and is driven by a motor to rotate continuously. As water flows through, suspended solids, algae, silt, and other impurities are intercepted by the rotary filter screen, and the filtered water continues to circulate within the system.

[0003] A rotary filter body generally consists of a frame, the rotary filter body itself, a transmission device, and a backwashing system. After long-term operation, the surface of the rotary filter body will become clogged with impurities, leading to a decrease in filtration efficiency. Therefore, it is necessary to backwash the rotary filter body regularly using the backwashing system.

[0004] However, most current rotary filter backwashing systems use a method of continuous rotation and fixed nozzle flushing for backwashing. Regardless of whether the rotary filter is actually clogged, it is flushed at preset time intervals, which can easily lead to over-flushing of clean areas while severely clogged areas are not cleaned in a timely and sufficient manner. Furthermore, it is impossible to identify the differences in clogging in local areas of the rotary filter. Summary of the Invention

[0005] Therefore, the present invention provides a partitioned step-backwashing system and method for rotating filters to solve the problems in the prior art.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] In a first aspect, a partitioned step-backwashing system for a rotating filter screen includes a status monitoring module, a clogging characteristic quantification module, a status judgment module, a flushing mode selection module, and a multi-drive flushing module.

[0008] The status monitoring module is used to collect vibration spectrum data, visual image data, and pressure difference data of the rotating filter body in real time for each partition evenly divided along the circumference.

[0009] The blockage feature quantization module can extract vibration disorder values ​​from vibration spectrum data. Contrast in visual image data and related values Then, by combining the pressure difference Y, multiple parameters are integrated into a multidimensional feature array. And output;

[0010] The state determination module can classify multidimensional feature arrays using a classification model. Map to a high-dimensional space, calculate the regional congestion level reference value S, and output the congestion level corresponding to the regional congestion level reference value S;

[0011] The flushing mode selection module can match and query the built-in data table according to the comprehensive blockage level and type, generate a corresponding control instruction set containing complete flushing parameters, and send it to the multi-drive flushing module.

[0012] The multi-drive flushing module can receive control instruction sets, generate corresponding control instructions, and send the control instructions to the corresponding drive units.

[0013] Furthermore, the blockage feature quantification module includes a first processing unit and a second processing unit;

[0014] The first processing unit can decompose the vibration spectrum data into several rotational components based on the intrinsic time-scale decomposition algorithm, and calculate the vibration disorder value of the first three rotational components. And combine them to form the vibration characteristic vector [ , , ], used to reflect the degree of disorder in vibration values;

[0015] The second processing unit can process visual image data based on the gray-level co-occurrence matrix method, construct a gray-level co-occurrence matrix, and then calculate the contrast based on the gray-level co-occurrence matrix. and related values .

[0016] Furthermore, the specific contents of the second processing unit are as follows:

[0017] 1) Convert visual image data into grayscale image format;

[0018] 2) Construct a gray-level co-occurrence matrix from the converted grayscale image;

[0019] 3) Calculate contrast based on gray-level co-occurrence matrix ;

[0020] Contrast The calculation formula is as follows:

[0021]

[0022] in, This represents the pixel value at a position in the gray-level co-occurrence matrix. The number of gray levels;

[0023] 4) Calculate correlation values ​​based on the gray-level co-occurrence matrix ;

[0024] Correlation value The calculation formula is as follows:

[0025]

[0026] in, and Element values and The mean, and These are the element values.

[0027] Furthermore, the specific content of the state determination module is as follows:

[0028] 1) A classification model uses a multidimensional feature array X as input samples, and a mapping function maps the multidimensional feature array X to a high-dimensional feature space. as follows:

[0029]

[0030] in, and These are the preset weight vectors and bias vectors;

[0031] 2) Based on the mapping function The sign determines the multidimensional feature array The initial category;

[0032] 3) After obtaining the initial category results, calculate the multidimensional feature array. To the optimal hyperplane The geometric distance d is used to calculate the regional congestion level reference value. Regional congestion level reference value The calculation formula is as follows:

[0033] in, For parameter factors;

[0034] 4) Determine the overall congestion level based on the threshold.

[0035] Furthermore, if the mapping function If ≥0, then the multidimensional feature array Classified as blocked; if the mapping function <0, then the multidimensional feature array It was classified as unblocked.

[0036] Furthermore, the rinsing mode selection module includes a reading unit and a matching unit; the reading unit can read the preset rotating filter body partition configuration parameters, and has a built-in data table containing the mapping relationship between each workstation number and its corresponding physical location;

[0037] The matching unit can obtain the blockage level S and blockage type of the partition through the real-time data bus, and then perform a matching query with the data table to generate a control instruction set containing complete flushing parameters.

[0038] Furthermore, the control commands include the valve opening and closing sequence, the pump start and stop frequency, and the valve opening degree.

[0039] Furthermore, the drive unit includes an air pulse drive unit, a water jet drive unit, and an air-water mixed flushing drive unit; the air pulse drive unit can generate air pulses with precisely controllable pulse width and pressure; the water jet drive unit is responsible for regulating the flow rate and pressure of the flushing water; and the air-water mixed flushing drive unit is used to generate intermittent oscillating jets of gas and liquid phases.

[0040] Secondly, a partitioned step backwashing method for a rotating filter screen includes the following steps:

[0041] S1: Collect vibration spectrum data, visual image data, and pressure difference data of the rotating filter body in each zone evenly divided along the circumference.

[0042] S2: Perform intrinsic time-scale decomposition on the collected vibration spectrum data, calculate the vibration disorder values ​​of the first three rotational components to construct the vibration feature vector; at the same time, process the visual image data using the gray-level co-occurrence matrix method to obtain contrast and correlation values;

[0043] S3: Normalize the vibration feature vector, contrast, correlation value and pressure difference and then fuse them into a multi-dimensional feature array;

[0044] S4: Map the multidimensional feature array to a high-dimensional space, calculate the geometric distance and regional blockage level reference value, and determine the blockage level and the corresponding blockage type through the threshold.

[0045] S5: Based on the preset configuration parameters of the rotating filter body and the clogging level of each zone, match and query in the built-in data table to generate a control instruction set containing complete flushing parameters;

[0046] S6: After receiving the control command, the multi-drive flushing module is activated by the timing controller in sequence according to the preset program.

[0047] The present invention has the following advantages: By dividing the working surface of the rotating filter body into multiple rinsing stations and assigning independent numbers to each zone, and combining the real-time acquisition of the clogging level and clogging type, the present invention can achieve precise positioning and targeted rinsing of each zone of the rotating filter body, effectively solving the problem of over-rinsing of clean areas and untimely cleaning of severely clogged areas.

[0048] Meanwhile, based on the multi-dimensional feature array after the fusion of multi-source sensing data, the classification model is used to determine the blockage situation. Then, the flushing mode selection module selects a matching flushing scheme from the predefined strategy library according to different blockage conditions. It can automatically adjust parameters such as flushing medium type, working pressure, and action time, so that the whole system has adaptive capability, improves backflushing efficiency, and reduces resource waste.

[0049] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. Attached Figure Description

[0050] To more intuitively illustrate the prior art and this application, exemplary drawings are provided below. It should be understood that the specific shapes and structures shown in the drawings should not generally be regarded as limiting conditions for implementing this application; for example, based on the technical concept disclosed in this application and the exemplary drawings, those skilled in the art are capable of making conventional adjustments or further optimizations to the addition / reduction / classification of certain units, their specific shapes, positional relationships, connection methods, size ratios, etc.

[0051] Figure 1 This is an implementation architecture diagram of a partitioned step backwashing system for a rotating filter screen according to the present invention.

[0052] Figure 2 This is a flowchart illustrating the implementation of a partitioned step-backwashing method for a rotating filter screen according to the present invention. Detailed Implementation

[0053] The following specific embodiments illustrate the implementation of the present invention. Those skilled in the art can easily understand other advantages and effects of the present invention from the content disclosed in this specification. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. It should be understood that these embodiments are merely for further explanation of the present invention and should not be construed as limiting the scope of protection of the present invention. Technical engineers in the field can make some non-essential improvements and adjustments to the present invention based on the above-described content. 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.

[0054] Please see Figure 1A partitioned step backwashing system for rotating filters includes a status monitoring module 2, a clogging characteristic quantification module, a status judgment module, a flushing mode selection module, and a multi-drive flushing module.

[0055] The status monitoring module 2 is used to collect vibration spectrum data, visual image data and pressure difference data of each partition evenly divided along the circumference of the rotating filter body 1 in real time, so as to provide a real-time and comprehensive status perception data foundation for subsequent analysis of blockage characteristics.

[0056] The condition monitoring module 2 includes a distributed fiber optic vibration sensor array, a multispectral linear array CMOS sensor, and a high-precision differential pressure transmitter group. It is mounted above the rotating filter body 1 via a support frame. The distributed fiber optic vibration sensor array is arranged on the outer frame of the rotating filter body 1, facilitating indirect judgment of the blockage status in each zone by sensing changes in the vibration frequency and amplitude of the rotating filter body 1.

[0057] The spectral linear array CMOS sensor captures image data of the surface of the rotating filter body 1 line by line through a rotating scanning method. It has a switchable narrow band filter wheel integrated in front of it, covering the visible light to near-infrared bands such as 450nm, 550nm, 650nm, and 850nm, and can distinguish the spectral reflectance characteristics of different components of blockages such as algae, mineral deposits, and organic fibers.

[0058] A high-precision differential pressure transmitter is installed at multiple preset specific measurement points on both the inner and outer sides of the rotating filter body 1. It is used to measure the pressure difference between the two sides of the rotating filter body 1, which is a direct parameter characterizing the resistance to water flow.

[0059] Simultaneously, the distributed fiber optic vibration sensor array, multispectral linear CMOS sensor, and high-precision differential pressure transmitter group are installed according to the preset partition configuration parameters of the rotating filter body 1. This divides the entire working surface of the rotating filter body 1 into N rinsing stations and assigns an independent number to each partition, facilitating the acquisition of the corresponding rotating filter body 1 parameters for each partition.

[0060] The blockage feature quantization module can extract vibration disorder values ​​from vibration spectrum data. Contrast in visual image data and related values Then, by combining the pressure difference Y, multiple parameters are integrated into a multidimensional feature array. It also outputs data to process the vibration spectrum data and visual image data, facilitating subsequent auxiliary judgment of blockage conditions.

[0061] The blockage feature quantification module includes a first processing unit and a second processing unit.

[0062] The first processing unit can decompose the vibration spectrum data into several rotational components based on the intrinsic time-scale decomposition algorithm. The rotational components can more clearly present the different characteristic components of the signal, which facilitates subsequent feature extraction.

[0063] After obtaining the rotational components, the vibration disorder values ​​of the first three rotational components are calculated respectively. And combine them to form the vibration characteristic vector [ , , This is used to reflect the degree of disorder in vibration values, and at the same time reflects the inherent characteristics of vibration signals, providing an important basis for judging blockage conditions.

[0064] Vibration disorder value The calculation formula is as follows:

[0065] in, For the first One vibration spectrum value, Where N is the current time point and N is the number of sampling points.

[0066] The aforementioned intrinsic time-scale decomposition algorithms refer to a class of adaptive decomposition methods applicable to non-stationary signals. Their core principle is to construct baseline components by identifying the signal's local extrema, thereby recursively decomposing the original complex signal into a series of physically meaningful rotating components. These methods do not rely on pre-defined basis functions and can adaptively decompose based on the signal's local time-scale characteristics. Such algorithms can also be signal processing methods based on empirical mode decomposition. This is used to decompose the non-stationary, nonlinear original vibration signal collected during the operation of the rotating filter body 1 into several rotating components with quasi-orthogonality and arranged from high to low frequencies. This lays the foundation for subsequent accurate calculation of vibration energy entropy and construction of feature vectors characterizing the blockage state.

[0067] The second processing unit can process visual image data based on the gray-level co-occurrence matrix method, construct a gray-level co-occurrence matrix, and then calculate the contrast based on the gray-level co-occurrence matrix. and related values The specific details are as follows:

[0068] 1) Convert visual image data into grayscale image format

[0069] Using weighted formula The RGB value of each pixel is calculated to obtain a single value, where R1, G, and B represent the original pixel values ​​of the red, green, and blue channels, respectively.

[0070] A single numerical value represents the gray level of a pixel in a grayscale image, thus transforming a visual image that originally had three-channel color information into a grayscale image that only contains gray level information, laying the foundation for subsequent image feature analysis based on the gray-level co-occurrence matrix.

[0071] 2) Construct a gray-level co-occurrence matrix from the converted grayscale image.

[0072] First, set a fixed distance and direction angle, count all pairs of pixels in the image that satisfy the conditions at that distance and direction, then count the frequency of simultaneous occurrence of pixel pairs at different gray levels, and combine and arrange the occurrence frequencies to form a gray-level co-occurrence matrix.

[0073] 3) Calculate contrast based on gray-level co-occurrence matrix ;

[0074] The elements far from the main diagonal in the gray-level co-occurrence matrix reflect the regions with significant gray-level variations in the image. Contrast is calculated by weighted summation of these far-away elements. Contrast Used to reflect the sharpness and texture depth of an image, contrast. The higher the value, the clearer the image texture and the more obvious the grayscale changes. Contrast The calculation formula is as follows:

[0075]

[0076] in, The position in the gray-level co-occurrence matrix ( The pixel value of ) The grayscale level is denoted by .

[0077] 4) Calculate correlation values ​​based on the gray-level co-occurrence matrix

[0078] The similarity between row or column elements in a gray-level co-occurrence matrix reflects the linear dependence of gray levels in a specific direction in an image. The correlation values ​​of the elements in the matrix are calculated along the row and column directions. The size reflects the consistency of image texture across different directions; correlation value The closer the value is to 1, the stronger the linear correlation of the texture in a specific direction. (Correlation value) The calculation formula is as follows:

[0079]

[0080] in, and Element values and The mean, and These are the element values.

[0081] 5) Data fusion

[0082] The vibration feature vector [ , , Contrast Correlation value The pressure difference Y is normalized separately, and then the processed parameters are aggregated into a predefined data structure to generate a multidimensional feature array. , , , This array, as a structured data entity, is fully output to the subsequent classification model through a standard data interface, providing it with complete feature input to perform classification decisions.

[0083] Normalization is based on preset extreme value parameters and uses a processing formula to linearly transform parameters from different physical dimensions to a unified numerical range, thereby forming a multidimensional feature array with consistent dimensions and comparability. The processing formula is as follows:

[0084]

[0085] in, For each parameter, and These are the minimum and maximum values ​​of each parameter obtained statistically from the historical dataset.

[0086] The state determination module can classify multidimensional feature arrays using a classification model. Mapping to a high-dimensional space and calculating a regional congestion level reference value S, the congestion level corresponding to the regional congestion level reference value S is output, thereby ensuring that the classification results match the actual congestion situation; the specific content is as follows:

[0087] 1) The classification model is built based on the Support Vector Machine (SVM) classification method. It can take a multi-dimensional feature array X as input samples and map the multi-dimensional feature array X to a high-dimensional feature space through a mapping function. These correspond to three different physical level measurement indicators: vibration disorder of the rotating filter body 1, surface texture roughness and directionality, and fluid resistance. Mapping function as follows:

[0088]

[0089] in, and These are the preset weight vectors and bias vectors.

[0090] 2) Based on the mapping function The sign determines the multidimensional feature array The initial category. If If ≥0, then the multidimensional feature array Classified as blocked; if <0, then the multidimensional feature array It was classified as unblocked.

[0091] 3) After obtaining the initial category results, calculate the multidimensional feature array. To the optimal hyperplane The geometric distance d, the distance d reflects the multidimensional feature array The confidence level of the classification increases with distance. Then, a reference value for the regional congestion level is calculated based on the distance 'd'. Regional congestion level reference value A larger value indicates a more severe congestion. The formula for calculating the geometric distance d is as follows:

[0092] in, Weight vector The norm of the hyperplane represents the degree of "steepness".

[0093] Regional congestion level reference value The calculation formula is as follows:

[0094] in, For parameter factors.

[0095] 4) Determine the overall congestion level based on the threshold.

[0096] For example: if If the value is less than 0.3, the clogging level is 1, which is the biological slime type; if 0.3 ≤ If the value is less than 0.7, the clogging level is 2, which is mineral scale type; if If the value is ≥0.7, the blockage level is 3, which is fiber entanglement type.

[0097] The flushing mode selection module can match and query the built-in data table based on the overall clogging level and type, generate a corresponding control command set containing complete flushing parameters, and send it to the multi-drive flushing module. This enables the positioning of each zone of the rotating filter body 1, matching appropriate flushing strategies according to the clogging situation, and outputting relevant commands for subsequent execution and driving.

[0098] The rinsing mode selection module includes a reading unit and a matching unit. The reading unit can read the preset configuration parameters of the rotating filter body 1, divide the entire working surface of the rotating filter body 1 into N rinsing stations, and assign an independent number to each station. At the same time, the reading unit has a built-in data table containing the mapping relationship between each station number and its corresponding physical location, which can provide a basis for subsequent accurate positioning of the stations.

[0099] The matching unit can obtain the blockage level S and blockage type of the partition through the real-time data bus, and then perform a matching query with the data table. The data table adopts a three-level index structure: the first level is classified by blockage type, the second level is segmented by blockage level, and the third level stores the specific flushing parameter combinations.

[0100] Upon successful matching, the matching unit generates a control instruction set containing complete flushing parameters. This instruction set explicitly specifies parameters such as flushing medium type, working pressure, duration, and alternation cycle, and is transmitted to the multi-drive flushing module via the industrial control bus using a standard communication protocol. Simultaneously, this module records the log information of this decision for system performance analysis and strategy optimization.

[0101] The multi-drive flushing module can receive control command sets and generate corresponding control commands, including valve opening and closing sequence, pump start and stop frequency, and regulating valve opening degree, and send the control commands to the corresponding drive unit respectively.

[0102] The drive unit comprises an air pulse drive unit, a water jet drive unit, and an air-water mixed flushing drive unit. The air pulse drive unit generates air pulses with precisely controllable pulse width and pressure. The water jet drive unit regulates the flow rate and pressure of the flushing water. The air-water mixed flushing drive unit generates an intermittent oscillating jet of gas and liquid phases.

[0103] The multi-drive flushing module also includes a timing controller. After receiving the flushing command, the timing controller activates the corresponding media drive units sequentially according to a preset program. For the air pulse drive unit, the controller adjusts the opening and closing frequency and duty cycle of the high-speed solenoid valve to achieve the output of a specified air pressure and pulse frequency. For the water jet drive unit, the water flow is controlled by a proportional regulating valve, and the pressure is precisely controlled in conjunction with the frequency conversion regulation of the multi-stage centrifugal pump. The air-water mixing flushing drive unit coordinates the output of the first two drive units according to the command, and generates a two-phase flow with a specific mixing ratio through a Venturi mixer.

[0104] After the set flushing time for each zone is completed, the timing controller executes the shutdown process sequentially. First, the media supply is cut off, and after the pipeline pressure is released, a stepping command is sent to the drive unit. The servo motor drives the rotating filter body 1 to rotate to the next station, while the encoder verifies the position. The entire process follows a cyclical sequence of "positioning-flushing-stepping" to ensure that each zone receives independent and precise targeted flushing.

[0105] Please see Figure 2 A partitioned step backwashing method for rotating filter screens includes the following steps:

[0106] S1: Collect vibration spectrum data, visual image data, and pressure difference data of each partition evenly divided along the circumference of the rotating filter body 1, respectively, to provide comprehensive status information for subsequent analysis.

[0107] S2: Perform intrinsic time-scale decomposition on the collected vibration spectrum data, calculate the vibration disorder values ​​of the first three rotational components to construct the vibration feature vector; at the same time, process the visual image data using the gray-level co-occurrence matrix method to obtain contrast and correlation values;

[0108] S3: Normalize the vibration feature vector, contrast, correlation value and pressure difference and fuse them into a multi-dimensional feature array; this facilitates the transformation of multi-source heterogeneous data into a quantifiable feature array, forming a unified feature representation for subsequent identification.

[0109] S4: Based on the classification model, the multidimensional feature array is mapped to a high-dimensional space to calculate the geometric distance and regional blockage level reference value. The comprehensive blockage level is determined by the threshold, and the blockage level and the corresponding blockage type are determined. Thus, the blockage level and type of the rotating filter body 1 can be accurately evaluated based on the multidimensional feature array, and automatic status recognition can be achieved.

[0110] S5: Based on the preset configuration parameters of the rotating filter body 1 zone and the clogging level of each zone, it matches and queries the built-in data table to generate a control instruction set containing complete flushing parameters, and sends it to the multi-drive flushing module via the industrial control bus; this facilitates matching the optimal flushing strategy according to the clogging situation and generating specific execution parameters to guide subsequent flushing operations.

[0111] S6: After receiving the control command, the multi-drive flushing module is activated by the timing controller in sequence according to the preset program. The corresponding air pulse drive unit, water jet drive unit and air-water mixed flushing drive unit are activated in sequence to complete the flushing operation of the set duration. Then the process is closed in sequence, and the rotating filter body 1 is driven to rotate to the next station. The "positioning-flushing-stepping" cycle is repeated to ensure that each section of the rotating filter body 1 can be effectively and accurately flushed.

[0112] This invention divides the working surface of the rotating filter body 1 into multiple rinsing stations and assigns an independent number to each zone. Combined with the real-time acquisition of the clogging level and clogging type, it can achieve precise positioning and targeted rinsing of each zone of the rotating filter body 1, effectively solving the problem of over-rinsing of clean areas and untimely cleaning of severely clogged areas.

[0113] Based on the multi-dimensional feature array after fusion of multi-source sensing data, the classification model is used to determine the blockage situation. Then, the flushing mode selection module selects a matching flushing scheme from the predefined strategy library according to different blockage conditions. It can automatically adjust parameters such as flushing medium type, working pressure, and action time, so that the whole system has adaptive capability, improves backflushing efficiency, and reduces resource waste.

[0114] Here is a specific application example:

[0115] First, the working surface of the rotating filter body 1 is evenly divided into multiple independent rinsing stations along the circumference, and each zone is assigned a unique number such as (P01, P02, …, P0N) to form N zones.

[0116] The condition monitoring module 2 is installed according to the aforementioned partitions and is used to collect vibration spectrum data, visual image data and pressure difference data of each partition in real time.

[0117] The blockage feature quantization module performs intrinsic time-scale decomposition on the vibration data, extracts the vibration disorder values ​​of the first three rotational components, and forms a vibration feature vector; it calculates the contrast C and correlation value R of the image data through the gray-level co-occurrence matrix, and then normalizes and fuses them with the pressure difference Y to form a multidimensional feature array X.

[0118] The status judgment module is based on a classification model built on support vector machines. It maps the feature array X to a high-dimensional space, calculates the geometric distance d from it to the optimal classification hyperplane, and then converts it into a regional blockage level reference value S through the Sigmoid function. Based on the preset threshold, it determines the specific blockage level and type (such as biological slime type, mineral hard scale type, and fiber entanglement type).

[0119] The flushing mode selection module queries the built-in strategy data table based on the partition number, blockage level and type, and generates a set of control instructions that includes parameters such as flushing medium, pressure and duration.

[0120] After receiving the command, the multi-drive flushing module's timing controller sequentially activates the air pulse, water jet, or air-water mixed flushing drive units to perform targeted flushing on the target area. After completion, it drives the rotating filter body 1 to rotate to the next station.

[0121] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A zoned step backwashing system for a rotating filter screen, characterized in that, It includes a status monitoring module (2), a blockage feature quantification module, a status judgment module, a flushing mode selection module, and a multi-drive flushing module; The status monitoring module (2) is used to collect in real time the vibration spectrum data, visual image data and pressure difference data of each partition evenly divided along the circumference of the rotating filter body (1); The blockage feature quantization module can extract vibration disorder values ​​from vibration spectrum data. Contrast in visual image data and related values Then, by combining the pressure difference Y, multiple parameters are integrated into a multidimensional feature array. And output; The state determination module can classify multidimensional feature arrays using a classification model. Map to a high-dimensional space, calculate the regional congestion level reference value S, and output the congestion level corresponding to the regional congestion level reference value S; The flushing mode selection module can match and query the built-in data table according to the comprehensive blockage level and type, generate a corresponding control instruction set containing complete flushing parameters, and send it to the multi-drive flushing module. The multi-drive flushing module can receive control instruction sets, generate corresponding control instructions, and send the control instructions to the corresponding drive units.

2. The partitioned step backwashing system for a rotating filter screen according to claim 1, characterized in that, The blockage feature quantification module includes a first processing unit and a second processing unit; The first processing unit can decompose the vibration spectrum data into several rotational components based on the intrinsic time-scale decomposition algorithm, and calculate the vibration disorder value of the first three rotational components. And combine them to form the vibration characteristic vector [ , , ], used to reflect the degree of disorder in vibration values; The second processing unit can process visual image data based on the gray-level co-occurrence matrix method, construct a gray-level co-occurrence matrix, and then calculate the contrast based on the gray-level co-occurrence matrix. and related values .

3. A zoned step-backwashing system for a rotating filter screen according to claim 2, characterized in that, The specific contents of the second processing unit are as follows: 1) Convert visual image data into grayscale image format; 2) Construct a gray-level co-occurrence matrix from the converted grayscale image; 3) Calculate contrast based on gray-level co-occurrence matrix ; Contrast The calculation formula is as follows: in, The position in the gray-level co-occurrence matrix ( The pixel value of ) The number of gray levels; 4) Calculate correlation values ​​based on the gray-level co-occurrence matrix ; Correlation value The calculation formula is as follows: in, and Element values and The mean, and These are the element values.

4. A zoned step-backwashing system for a rotating filter screen according to claim 1, characterized in that, The specific content of the status determination module is as follows: 1) A classification model uses a multidimensional feature array X as input samples, and a mapping function maps the multidimensional feature array X to a high-dimensional feature space. as follows: in, and These are the preset weight vectors and bias vectors; 2) Based on the mapping function The sign determines the multidimensional feature array The initial category; 3) After obtaining the initial category results, calculate the multidimensional feature array. To the optimal hyperplane The geometric distance d is used to calculate the regional congestion level reference value. Regional congestion level reference value The calculation formula is as follows: in, For parameter factors; 4) Determine the overall congestion level based on the threshold.

5. A zoned step-backwashing system for a rotating filter screen according to claim 4, characterized in that, If the mapping function If ≥0, then the multidimensional feature array Classified as blocked; if the mapping function <0, then the multidimensional feature array It was classified as unblocked.

6. A zoned step-backwashing system for a rotating filter screen according to claim 1, characterized in that, The rinsing mode selection module includes a reading unit and a matching unit; the reading unit can read the preset rotation filter body (1) partition configuration parameters, and has a built-in data table containing the mapping relationship between each workstation number and its corresponding physical location; The matching unit can obtain the blockage level S and blockage type of the partition through the real-time data bus, and then perform a matching query with the data table to generate a control instruction set containing complete flushing parameters.

7. A zoned step-backwashing system for a rotating filter screen according to claim 1, characterized in that, The control commands include the valve opening and closing sequence, the pump start and stop frequency, and the valve opening degree.

8. A zoned step backwashing system for a rotating filter screen according to claim 1, characterized in that, The drive unit includes a gas pulse drive unit, a water jet drive unit, and a gas-water mixed jet drive unit. The air pulse drive unit can generate air pulses with precise control over both pulse width and pressure; the water jet drive unit is responsible for regulating the flow rate and pressure of the flushing water. The gas-liquid mixing drive unit is used to generate intermittent oscillating jets of gas and liquid phases.

9. A method for partitioned step-backwashing of a rotating filter screen, characterized in that, Includes the following steps: S1: Collect vibration spectrum data, visual image data, and pressure difference data of each partition evenly divided along the circumference of the rotating filter body (1); S2: Perform intrinsic time-scale decomposition on the collected vibration spectrum data, calculate the vibration disorder values ​​of the first three rotational components to construct the vibration feature vector; at the same time, process the visual image data using the gray-level co-occurrence matrix method to obtain contrast and correlation values; S3: Normalize the vibration feature vector, contrast, correlation value and pressure difference and then fuse them into a multi-dimensional feature array; S4: Map the multidimensional feature array to a high-dimensional space, calculate the geometric distance and regional blockage level reference value, and determine the blockage level and the corresponding blockage type through the threshold. S5: Based on the preset rotation filter body (1) partition configuration parameters and the clogging level of each partition, match and query in the built-in data table to generate a control instruction set containing complete flushing parameters; S6: After receiving the control command, the multi-drive flushing module is activated by the timing controller in sequence according to the preset program.

Images