Sludge dryer uniform material control system and method

Abstract

The present application relates to a kind of sludge dryer material control system and method.The sludge dryer material control system described in the present application, comprising: data acquisition module is equipped with first detection unit and second detection unit, acquires sludge thickness data and sludge image data;Data processing module is used to extract, fusion and dimension reduction to sludge data feature, to output the feature vector after dimension reduction;Control module is used to build material decision model, and output control strategy;Material module is used to execute control strategy and carry out material to sludge.The sludge dryer material control system and control method described in the present application can more effectively fuse thickness feature, texture feature and shape feature together, while reducing the dimension of data, providing better input features for subsequent material decision model, achieving the precise material of sludge in drying machine, making the sludge distribution more uniform, avoiding the problem of poor drying effect caused by uneven distribution.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of sludge treatment, in particular to a sludge dryer uniform material control system and method. BACKGROUND

[0002] With the continuous growth and improvement of sewage treatment capacity in China, sludge treatment and disposal has become a new difficulty and challenge in environmental comprehensive management. Sludge disposal often adopts the sludge disposal process of "drying + separate incineration". In this sludge disposal process, the moisture content of the sludge after drying treatment needs to be reduced to meet the standard before the next step of incineration treatment. Therefore, effective drying of the sludge is crucial. At present, due to the energy saving and environmental protection of the sludge dryer, it is widely used in the drying scene of sludge.

[0003] During the drying process of the sludge in the drying oven of the drying machine, uneven distribution of the material may occur, and the material on the mesh belt needs to be evenly distributed. If the number of uniform materials is not enough, the moisture content of the sludge will be too high, which will directly affect the drying effect of the sludge; and if the number of uniform materials is too frequent, a lot of dust will be generated, which will cause the filter equipment to be quickly blocked and directly reduce the service life of the sludge drying machine. Therefore, effective and accurate uniform material and reduction of unnecessary uniform material times are crucial. SUMMARY

[0004] Therefore, the purpose of the present application is to provide a sludge dryer uniform material control system and control method, which can more effectively integrate the thickness features, texture features and shape features together, while reducing the dimensionality of the data, highlighting the main information in the data, providing better input features for the uniform material decision model, improving the accuracy and efficiency of the model, and realizing accurate uniform material of the sludge on the mesh belt in the drying oven of the drying machine, so that the sludge is more evenly distributed, avoiding the problem of poor drying effect caused by uneven distribution of the material. At the same time, the dust generated by excessive uniform material is effectively reduced, the risk of blockage of the filter system by the dust is reduced, and the service life of the filter equipment is prolonged.

[0005] A sludge dryer uniform material control system, comprising:

[0006] A data acquisition module for acquiring sludge thickness data and sludge image data;

[0007] A data processing module for feature extraction after preprocessing of the sludge thickness data and image data, as well as feature fusion and dimensionality reduction, to output a feature vector after dimensionality reduction;

[0008] A control module for building a uniform material decision model, receiving the feature vector after dimensionality reduction and inputting it into the uniform material decision model, and outputting a control strategy;

[0009] The even material module is arranged in the oven of the sludge dryer and is used for performing an even material control strategy on sludge.

[0010] Further, the data acquisition module comprises:

[0011] The first detection unit is provided with a distance measuring sensor and a temperature sensor, and the sludge thickness data is obtained by performing thickness compensation on the original sludge thickness obtained by the distance measuring sensor and the temperature obtained by the temperature sensor.

[0012] The second detection unit is provided with an industrial camera and is used for shooting sludge images and outputting sludge image data.

[0013] Further, the data processing module comprises:

[0014] The preprocessing unit is used for denoising the sludge thickness data output by the first detection unit, and is used for performing grayscale, normalization and filtering processing on the image data output by the second detection unit.

[0015] The feature extraction unit is used for extracting sludge thickness features and sludge image features.

[0016] The feature fusion unit is used for fusing and reducing dimensions of a feature vector composed of the extracted thickness features and image features.

[0017] Further, the sludge thickness features comprise average thickness, thickness standard deviation and thickness gradient; the sludge image features comprise texture features and shape features, the texture features comprise contrast, correlation, energy and entropy; and the shape features comprise area, perimeter and circularity.

[0018] Further, the control module comprises:

[0019] The sample data acquisition unit is used for acquiring sludge thickness data and sludge image data of the data acquisition module and operation data of the even material module.

[0020] The model construction unit is used for constructing an even material decision model based on a multilayer perception machine.

[0021] The decision output unit is used for inputting the reduced dimension feature vector into the trained even material decision model and outputting a control strategy.

[0022] Further, the even material decision model comprises an input layer and an output layer; the number of input layer nodes is equal to the dimension of the reduced dimension feature vector; and the output nodes of the output layer at least comprise even material decision and even material intensity.

[0023] The present application has the following beneficial effects:

[0024] (1) The thickness data of the laser triangulation sensor and the image data of the industrial camera are comprehensively utilized to give full play to the advantages of different sensors, describe the distribution state of the sludge from multiple angles, improve the accuracy of the control strategy, and make the sludge evenly distributed on the mesh belt through accurate material equalization, thereby improving the uniformity of drying and the drying efficiency of the entire drying machine, and reducing the drying time and energy consumption;

[0025] (2) In view of the problem that the ranging sensor is affected by temperature, the measured thickness data is corrected through a temperature compensation formula to improve the accuracy of the sludge thickness data;

[0026] (3) The image data is subjected to grayscale, normalization and filtering processing to reduce the influence of environmental factors such as light and noise on image feature extraction;

[0027] (4) The sludge thickness features, texture features and shape features are extracted, feature fusion and dimension reduction are performed, the data dimension is reduced while the main information is retained, better input features are provided for the material equalization decision model, and the efficiency and accuracy of the material equalization decision model are improved;

[0028] (5) The material equalization decision model based on a neural network learns the complex relationship between the feature vectors and the control decisions under different working conditions, so as to accurately output the material equalization decision and the material equalization strength according to the real-time collected data, control the material equalization module to perform material equalization, and realize closed-loop control;

[0029] (6) The entire material equalization control system realizes automatic and intelligent material equalization control, reduces manual intervention, improves the intelligent level of the production process, and reduces the possibility of manual errors and operation errors.

[0030] In order to better understand and implement, the present application will be described in detail below with reference to the accompanying drawings. BRIEF DESCRIPTION OF DRAWINGS

[0031] Figure 1 A schematic diagram of a sludge dryer material equalization control system provided for an embodiment of the present application Figure One ;

[0032] Figure 2 A schematic diagram of a sludge dryer material equalization control system provided for an embodiment of the present application Figure Two ;

[0033] Figure 3 A flowchart of a material equalization control method of a sludge dryer provided for an embodiment of the present application

[0034] Figure 4 A step flowchart of sludge feature fusion and dimension reduction provided for an embodiment of the present application.

[0035] In the figure: 10-data acquisition module; 11-first detection unit; 12-second detection unit; 20-data processing module; 21-preprocessing unit; 22-feature extraction unit; 23-feature fusion unit; 30-control module; 31-sample data acquisition unit; 32-model construction unit; 33-decision output unit; 40-uniform material module. DETAILED DESCRIPTION

[0036] The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative labor fall within the protection scope of the present application.

[0037] During the drying process of sludge in the oven of the drying machine, uneven distribution of materials is prone to occur, and the materials on the mesh belt need to be uniformly distributed. If the uniform distribution frequency is not enough, the water content of the sludge will be too high, which directly affects the drying effect of the sludge; and if the uniform distribution frequency is too high, a lot of dust will be generated, which will cause rapid clogging of the filtering equipment and directly reduce the service life of the sludge drying machine. Therefore, effective and accurate uniform distribution and reduction of unnecessary uniform distribution frequency are crucial.

[0038] In the prior art, the thickness of the sludge in the sludge drying machine is measured by a distance measuring sensor, and whether the sludge needs to be uniformly distributed is controlled based on the thickness data of the sludge. However, this uniform distribution control method based on sludge thickness measurement has a single feature dimension, which will result in low accuracy of the model, and further result in inaccurate uniform distribution control. Meanwhile, the uniform distribution strength cannot be controlled, and efficient uniform distribution cannot be achieved.

[0039] Based on this, the embodiments of the present application extract thickness features, texture features and shape features from multiple angles and fuse them together, while reducing the dimension of the data, highlighting the main information in the data, providing better input features for the uniform distribution decision model, improving the accuracy and efficiency of the model, achieving accurate uniform distribution of the sludge on the mesh belt in the oven of the drying machine, and making the sludge distribution more uniform, thereby avoiding the problem of poor drying effect caused by uneven distribution of materials.

[0040] Example 1

[0041] Please refer to Figure 1 and Figure 2The embodiment of the present application provides a sludge dryer uniform material control system, comprising: a data acquisition module 10, a data processing module 20, a control module 30 and a uniform material module 40. The data acquisition module 10 is arranged in the oven of the sludge dryer and is used for collecting sludge data. Specifically, the data acquisition module 10 is provided with a first detection unit 11 and a second detection unit 12, wherein the first detection unit 11 is used for collecting sludge thickness data, and the second detection unit 12 is used for collecting sludge image data. The data processing module 20 is used for extracting features after pre-processing the sludge thickness data and image data, and performing feature fusion and dimension reduction to output a feature vector after dimension reduction. The control module 30 is used for constructing a uniform material decision model, receiving the feature vector after dimension reduction and inputting the feature vector into the uniform material decision model to output a control strategy. The uniform material module 40 is arranged in the oven of the sludge dryer and is used for receiving the control strategy to uniformly distribute the sludge.

[0042] Further, the first detection unit 11 is provided with a distance measuring sensor arranged in the oven of the sludge dryer. It can be understood that the distance measuring sensor is a sensor for detecting distance based on optical principle, which can adopt an infrared distance measuring sensor, an ultrasonic sensor or a laser distance measuring sensor. In the embodiment of the present application, the distance measuring sensor adopts a laser triangulation sensor, which is arranged at different positions above the mesh belt in the oven. In the running process, the plurality of laser triangulation sensors continuously emit laser beams to the sludge surface and receive reflected light. Each laser triangulation sensor measures the distance from the sludge surface according to the change of the reflected light angle, and calculates the original sludge thickness data.

[0043] Further, the first detection unit 11 further comprises a temperature sensor arranged in cooperation with the distance measuring sensor, which is used for testing real-time temperature and, in view of the problem that the laser triangulation sensor is affected by temperature, performing thickness compensation on the original sludge thickness data obtained by the distance measuring sensor and the temperature tested by the temperature sensor to obtain sludge thickness data, correcting the measured original sludge thickness data through temperature compensation, and improving the accuracy of the sludge thickness data.

[0044] Specifically, the sludge thickness data T compensated after temperature compensation is:

[0045] T compensated = T raw ×(1+k×(T env -T ref ));

[0046] Wherein, T raw is the original sludge thickness data; T env is the environmental temperature; the reference temperature T ref is a reference temperature, which is set to 25℃; and k is a temperature compensation coefficient, which is obtained through calibration experiment.

[0047] For example, at a certain time, a distance sensor measures T raw = 25.3 mm, T env = 60°C, then the compensated sludge thickness is:

[0048] T compensated = 25.3 x ((1 + 0.001 x (60 - 25)) = 25.6 mm.

[0049] Further, the second detection unit 12 is provided with an industrial camera for shooting a sludge image and outputting sludge image data. In the embodiment of the present application, a plurality of industrial cameras are arranged on the top of the oven, the lenses of which are vertically downwardly aligned with the mesh belt, and the industrial cameras shoot images at certain time intervals to obtain image data of the sludge on the mesh belt. Specifically, the image data contains information such as the texture and shape of the sludge.

[0050] Further, in some specific embodiments, the data processing module 20 comprises a preprocessing unit 21, a feature extraction unit 22 and a feature fusion unit 23, wherein the preprocessing unit 21 is configured to denoise the sludge thickness data output by the first detection unit 11, and to perform grayscale, normalization and filtering processing on the image data output by the second detection unit 12; the feature extraction unit 22 is configured to extract sludge thickness features and sludge image features; and the feature fusion unit 23 is configured to fuse and reduce the dimension of the feature vector composed of the extracted thickness features and image features.

[0051] Further, the preprocessing unit 21 comprises a thickness filtering subunit for the sludge thickness data, which performs wavelet transform filtering on the temperature-compensated sludge thickness data to remove noise interference. The filtered thickness signal is:

[0052]

[0053] wherein w(a, b) is a wavelet coefficient, is a wavelet basis function, and A and B are scale and displacement parameters of the wavelet transform.

[0054] Further, in the embodiment of the present application, the preprocessing unit 21 further comprises a gray value calculation subunit, a normalization processing subunit and a filtering processing subunit for the sludge image data, which are configured to perform grayscale, normalization and filtering processing on the image data to reduce the influence of environmental factors such as light and noise on image feature extraction.

[0055] Specifically, the gray value calculation subunit is configured to perform grayscale processing on the sludge image data, and uses a weighted average method to convert a color image into a grayscale image. The red, green and blue channel values of a pixel in the image are defined as R, G and B respectively, and the gray value Gray is:

[0056] Gray = 0.299 * R + 0.587 * G + 0.114 * B.

[0057] Specifically, the normalization processing subunit is configured to scale the value ranges of different features to similar intervals, so as to eliminate the dimensional differences between different variables, so that different features have similar numerical ranges, that is, the gray value Gray is mapped to the interval [0, 1], so that the input signals of all samples have a mean value close to 0 or a very small mean square error compared with the mean value; and this is conducive to improving the convergence of the subsequent uniform material decision model.

[0058] The normalized gray value Normalized_Gray is:

[0059]

[0060] wherein, Gray max and Gray min are the maximum and minimum values of the gray values in the image, respectively.

[0061] Specifically, in the embodiment of the present application, the filtering processing subunit removes the noise in the image through median filtering, sets the gray value of each pixel point to the median of the gray values of all pixel points in a certain neighborhood window of the point, makes the surrounding pixel values close to the true values, and thus eliminates isolated noise points, so as to protect the edges of the signal and prevent them from being blurred. The original image pixel value is I(x, y), the filtering window is W, and the pixel value of the filtered image at the coordinates (x, y) is:

[0062] Median_Filerted_Image(x, y) = Median{I(x+i, y+j), (i, j) ∈ W}.

[0063] Further, the feature extraction unit 22 includes a thickness feature extraction subunit and an image feature extraction subunit. In the embodiment of the present application, the thickness features extracted by the thickness feature extraction subunit include the average thickness the thickness standard deviation σ T and the thickness gradient

[0064] Specifically, the average thickness is the average value of the thickness of the sludge calculated based on the temperature-compensated sludge thickness data and is calculated by the following formula:

[0065]

[0066] wherein, N is the number of measurement points.

[0067] Specifically, the thickness standard deviation σ TBased on the average thickness and the sludge thickness data after temperature compensation For measuring the dispersion degree of thickness, the calculation formula is:

[0068]

[0069] Specifically, the thickness gradient is calculated by a differential method to calculate the change rate of the thickness of adjacent measurement points, and is used to capture the local thickness change trend.

[0070] Further, the image feature extraction subunit extracts the texture features and shape features of the sludge. Specifically, in the embodiments of the present application, the gray level co-occurrence matrix is used to extract the texture features of the sludge, which can accurately reflect the roughness and repeated direction of the texture. The gray level co-occurrence matrix reflects the texture features of the image by counting the gray level co-occurrence frequency of the pixel pairs, and then extracts the texture features such as contrast, correlation, energy and entropy. It can be understood that the gray level co-occurrence matrix is only an effective method for extracting the texture features of the sludge, and in other embodiments, methods such as Tamura texture features, autoregressive texture model, wavelet transform, etc. can also be used, and are not limited thereto.

[0071] Specifically, in the embodiments of the present application, the sludge block is identified by an image segmentation algorithm, and the shape features of the sludge such as area, perimeter, circularity, etc. are calculated.

[0072] Further, the feature fusion unit 23 includes a standardization subunit, a feature vector acquisition subunit, and a dimension reduction processing subunit.

[0073] In the embodiments of the present application, the standardization subunit is used to standardize the extracted thickness features, shape features and texture features of the sludge, so that they have the same mean and standard deviation, and obtain a standardized feature matrix.

[0074] Specifically, in the embodiments of the present application, it is embodied that, based on the thickness features such as the average thickness the thickness standard deviation σ T , the thickness gradient the texture features such as contrast C, correlation R, energy E, and entropy H; and the shape features such as area A, perimeter P, and circularity C r are combined into an original feature matrix X, and the original feature matrix X is standardized to make each feature dimension have a mean and a standard deviation.

[0075] Corresponding to the i-th row and j-th column feature element x ij of the original feature matrix X, the mean μ j and the standard deviation σ j are calculated:

[0076]

[0077] where m is the number of samples.

[0078] Further, based on the mean μ j and the standard deviation σ j , the feature elements in the normalized feature matrix X std are calculated.

[0079] In the embodiments of the present application, the feature vector acquisition subunit is used to construct the feature vector. In the embodiments of the present application, the covariance matrix of the normalized feature matrix is calculated, and the eigenvalue decomposition and the calculation of the feature vector are performed to obtain the feature vector.

[0080] Specifically, the element c pq of the covariance matrix C of the normalized feature matrix X std is:

[0081]

[0082] where p, q = 1, 2, ···, n, and n is the total number of features, i.e., the total number of thickness features, texture features, and shape features.

[0083] Then, the eigenvalue decomposition is performed on the covariance matrix C:

[0084] C Vk = λ k υ k (k = 1, 2, ···, n);

[0085] The eigenvalues λ1, λ2, ··· λ n and the corresponding eigenvectors υ1, υ2, ···, υ n are obtained.

[0086] In the embodiments of the present application, the dimension reduction processing subunit adopts principal component analysis (PCA) to perform dimension reduction processing on the feature vector to obtain the dimension-reduced feature vector F PCA , which reduces the data dimension while retaining the main information, improves the efficiency and accuracy of the uniform material decision model.

[0087] Specifically, the feature vectors are sorted according to the sizes of the eigenvalues. The larger the eigenvalue, the greater the data variance in the direction corresponding to the feature vector, which means that this direction contains more information. The feature vectors corresponding to the first k largest eigenvalues are selected as the principal components to form the transformation matrix W, where the selection of k is based on the cumulative variance contribution rate (CVR), and the calculation formula of the cumulative variance contribution rate (CVR) is:

[0088]

[0089] As an example, the k value when the cumulative variance contribution rate reaches 90% is selected.

[0090] Further, the principal component analysis (PCA) is used to calculate the reduced dimension feature vector F PCA :

[0091] F PCA = W T X std

[0092] where W T is the transpose matrix of W.

[0093] The embodiment of the application can more effectively fuse the thickness feature, texture feature and shape feature of the sludge together by fusing and reducing the dimension of the feature vector composed of the extracted thickness feature and image feature, highlight the main information in the data, and provide better input features for the subsequent uniform material decision model, thereby improving the accuracy and efficiency of the model.

[0094] Further, in the embodiment of the application, the control module 30 includes a sample data acquisition unit 31, a model construction unit 32 and a decision output unit 33, wherein the sample data acquisition unit 31 is configured to acquire the sludge thickness data and sludge image data of the data acquisition module 10 and the operation data of the uniform material module 40; the model construction unit 32 uses a multi-layer perception (MLP) as a neural network structure to construct a uniform material decision model; and the decision output unit 33 is configured to input the reduced dimension feature vector into the trained uniform material decision model and output a control strategy. The control strategy at least includes a uniform material decision D and a uniform material intensity S, and the uniform material decision D is represented by 0 or 1, wherein 0 represents no need for uniform material and 1 represents the need for uniform material; the value range of the uniform material intensity S is between [0, 1], reflecting the operation parameters of the uniform material module. As an example, the model output is D = 1 and S = 0.6, indicating that the uniform material operation is needed and the uniform material intensity is moderate.

[0095] Specifically, the sample data acquisition unit 31 collects the sludge thickness data and sludge image data collected by the first detection unit 11 and the second detection unit 12 under a large number of different working conditions, as well as the sludge feeding speed, humidity, mesh belt running speed and the like of the sludge drier under different working conditions. For each group of data, the values of the uniform material decision D and the uniform material intensity S are manually labeled, and these labeled data are divided into a training set, a validation set and a test set.

[0096] Specifically, the model construction unit 32 uses a multi-layer perception (MLP) as a neural network structure. The neural network structure includes an input layer, a plurality of hidden layers and an output layer. The number of nodes of the input layer is equal to the reduced dimension feature vector F PCAThe output layer outputs the control strategy. In the embodiment, the output layer includes two nodes, namely, the uniform material decision D and the uniform material intensity S. It can be understood that in some other embodiments, more nodes can be provided in the output layer, for example, a drying temperature, for controlling the working temperature of the oven of the sludge dryer. It can be understood that the number of layers and the number of nodes of the hidden layer are determined through experiments and cross-validation. As an example, the number of layers of the hidden layer is set to 2-3 layers, and the number of nodes of each layer is between 10-50.

[0097] Based on the constructed uniform material decision model, the training set data is used for training. As an example, in the training process, the cross-entropy loss function is used for the output of the uniform material decision D, and the mean square error loss function is used for the output of the uniform material intensity S. At the same time, an optimization algorithm is used to adjust the weights of the neural network to minimize the value of the loss function. As an example, the optimization algorithm can be a stochastic gradient descent algorithm and its variants such as Adam and Adagrad, which are not described in detail herein. In the training process, the performance of the uniform material decision model is evaluated on the validation set regularly, and the training is stopped when the performance on the validation set no longer improves to prevent overfitting.

[0098] Specifically, the decision output unit 33 inputs the reduced dimension feature vector F PCA into the uniform material decision model, and the uniform material decision model outputs the uniform material decision D and the uniform material intensity S.

[0099] Further, the uniform material module 40 is arranged in the oven of the sludge dryer. The uniform material module 40 adjusts the moving speed and the swing amplitude of the scraper according to the uniform material decision and the uniform material intensity to perform uniform material on the sludge on the mesh belt. As an example, when the uniform material intensity S output by the uniform material decision model is 0.6, the uniform material module 40 adjusts the moving speed of the scraper to 60% of the preset speed of the system and adjusts the swing amplitude to 1.2 times of the preset amplitude of the system. It should be noted that the specific parameter adjustment relationship of the uniform material module 40 is determined according to the characteristics of the equipment and experiments, which is not described in detail herein. After the uniform material is completed, the uniform material control system enters the data acquisition stage again to repeat the above process to continuously monitor and adjust the distribution state of the sludge.

[0100] The sludge dryer uniform material control system provided in the embodiment realizes automatic and intelligent uniform material control, reduces manual intervention, improves the intelligent level of the production process, and reduces the possibility of manual cost and operation errors.

[0101] Example 2

[0102] Please refer to Figure 3 and Figure 4 , based on the above sludge dryer uniform material control system, the embodiment further provides a sludge dryer uniform material control method, including the following steps:

[0103] S1, obtaining sludge thickness data and sludge image data of sludge on the mesh belt;

[0104] S2, preprocessing the sludge thickness data and extracting sludge thickness features; meanwhile, preprocessing the sludge image data and extracting sludge image features;

[0105] S3, combining the sludge thickness features and the sludge image features to obtain an original feature matrix, fusing the original feature matrix to obtain a feature vector, and performing dimension reduction processing on the feature vector;

[0106] S4, constructing a uniform material decision-making model based on the feature vector;

[0107] S5, inputting the dimension-reduced feature vector into the uniform material decision-making model to obtain a control strategy;

[0108] S6, executing the control strategy.

[0109] Further, in step S1, the sludge thickness data is sludge thickness data obtained by temperature compensation on sludge original thickness data tested based on a distance measuring sensor.

[0110] Further, in step S2, based on the sludge thickness data after temperature compensation, preprocessing is performed, and then feature extraction is performed, including the following steps:

[0111] S211, performing wavelet transform filtering on the sludge thickness data after temperature compensation to separate the wavelet transform of the signal from the wavelet transform of the noise to remove the noise, and the filtered thickness signal is:

[0112]

[0113] wherein w(a, b) is a wavelet coefficient, is a wavelet basis function, and A and B are scale and displacement parameters of the wavelet transform.

[0114] S212, based on the preprocessed sludge thickness data, performing average thickness thickness standard deviation σ T and thickness gradient extraction, and the specific steps are as follows:

[0115] First, based on the sludge thickness data after temperature compensation calculate the average thickness

[0116]

[0117] wherein N is the number of measurement points.

[0118] Then, based on the average thickness and the temperature-compensated sludge thickness data Calculate the thickness standard deviation σ T :

[0119]

[0120] Next, the thickness gradient is calculated by the difference method to calculate the change rate of the thickness of adjacent measurement points

[0121] Further, in step S2, the sludge image data is preprocessed and the sludge image features are extracted, including the following steps:

[0122] S221, the image data of the sludge image data is subjected to grayscale processing, normalization processing and filtering processing to reduce the influence of environmental factors such as light and noise on image feature extraction, and the specific steps are:

[0123] First, the weighted average method is used to convert the color image to a grayscale image, and the grayscale value is calculated;

[0124] Then, the grayscale value is mapped to the interval [0, 1] to obtain the normalized grayscale value;

[0125] Next, the noise in the image is removed by median filtering, and the grayscale value of each pixel point is set to the median value of the grayscale values of all pixel points in a certain neighborhood window of the point, so that the surrounding pixel values are close to the true value, thereby eliminating isolated noise points.

[0126] S222, based on the preprocessed sludge image data, the texture features calculated by the gray level co-occurrence matrix, and the shape features calculated by the image segmentation algorithm. Specifically, the extracted texture features include contrast, correlation, energy and entropy; the extracted shape features include area, perimeter and circularity.

[0127] Further, in step S3, based on the sludge thickness features and the sludge image features, the original feature matrix is obtained, and the original feature matrix is fused to obtain the feature vector, and then the feature vector is processed. Dimension reduction includes the following steps:

[0128] S31, the feature data of the original feature matrix is standardized to obtain the standardized feature matrix X std ;

[0129] S32, based on the standardized feature matrix X std , the covariance matrix C is calculated;

[0130] S33, the eigenvalue decomposition of the covariance matrix is carried out to obtain the eigenvalue λ i and the corresponding eigenvector υi ;

[0131] S34, rank the eigenvectors according to the size of the eigenvalues, select the eigenvectors corresponding to the first k largest eigenvalues as principal components, and form a transformation matrix W; the selection of k value is based on the cumulative variance contribution rate CVR, and the k value when the CVR reaches a set proportion is selected, and the calculation formula is:

[0132]

[0133] S35, calculate the feature vector F PCA :

[0134] F PCA =W T X std ;

[0135] Where W T is the transpose matrix of W.

[0136] In the above steps, steps S31-S33 are specific steps for fusing the original feature matrix to obtain eigenvalues and eigenvectors; steps S34-S35 are specific steps for dimensionality reduction processing of the eigenvectors to obtain the dimensionality-reduced eigenvectors F PCA .

[0137] Further, in step S4, the specific steps for constructing the uniform feeding decision model based on the eigenvectors are:

[0138] S41, collect a large amount of sludge thickness data and sludge image data collected under different working conditions, as well as sludge feeding speed, humidity, mesh belt running speed and other conditions of the sludge dryer under different working conditions; for each group of data, manually label the uniform feeding decision D and the uniform feeding strength S value, and divide these labeled data into a training set, a validation set and a test set.

[0139] S42, use a multilayer perception as a neural network structure, and based on the collected data and parameters, establish a uniform feeding decision model.

[0140] S43, based on the constructed uniform feeding decision model, use the training set data to train it; during the training process, periodically evaluate the performance of the uniform feeding decision model on the validation set, and stop training when the performance on the validation set no longer improves.

[0141] Further, input the dimensionality-reduced eigenvectors F PCA to the uniform feeding decision model, and the uniform feeding decision model outputs the uniform feeding decision D and the uniform feeding strength S.

[0142] Further, according to the uniform material decision model outputting the uniform material decision D and the uniform material strength S, the moving speed and the swing amplitude of the scraper and other parameters are adjusted to uniformly distribute the sludge on the mesh belt.

[0143] The uniform material control method of the sludge dryer provided by the embodiment of the application realizes accurate uniform material of the sludge on the mesh belt in the drying oven of the dryer, makes the sludge thickness distribution more uniform, and avoids the problem of poor drying effect caused by uneven distribution of the sludge. At the same time, accurate uniform material control helps to maintain stable operation of the entire drying system, reduces equipment failure and unstable operation caused by uneven distribution of the sludge, and improves the reliability and production continuity of the equipment.

[0144] Compared with the prior art, the application has the following advantages:

[0145] (1) The thickness data of the laser triangulation sensor and the image data of the industrial camera are comprehensively utilized to give full play to the advantages of different sensors, describe the distribution state of the sludge from multiple angles, improve the accuracy of the control strategy, and make the sludge uniformly distributed on the mesh belt through accurate uniform material, thereby improving the uniformity of drying and the drying efficiency of the entire dryer, and reducing the drying time and energy consumption;

[0146] (2) The temperature compensation formula is used to correct the measured thickness data to improve the accuracy of the sludge thickness data in view of the problem that the ranging sensor is affected by temperature;

[0147] (3) The image data is subjected to grayscale, normalization and filtering processing to reduce the influence of environmental factors such as light and noise on image feature extraction;

[0148] (4) The sludge thickness features, texture features and shape features are extracted, feature fusion and dimension reduction are performed, the data dimension is reduced while the main information is retained, better input features are provided for the uniform material decision model, and the efficiency and accuracy of the uniform material decision model are improved;

[0149] (5) The uniform material decision model based on the neural network learns the complex relationship between the feature vectors and the control decisions under different working conditions, thereby accurately outputting the uniform material decision and the uniform material strength according to the real-time collected data, controlling the uniform material module to uniformly distribute the sludge, and realizing closed-loop control;

[0150] (6) The entire uniform material control system realizes automatic and intelligent uniform material control, reduces manual intervention, improves the intelligent level of the production process, and reduces the possibility of human error and labor cost.

[0151] The above embodiments only express several implementation manners of the present application, and the description is relatively specific and detailed, but it should not be understood as a limitation on the scope of the patent. It should be noted that for ordinary skilled persons in the art, several modifications and improvements can be made without departing from the concept of the present application, and the present application also intends to include these modifications and improvements.

Claims

1. A sludge dryer material equalization control system, characterized in that, include: The data acquisition module (10) is used to acquire sludge thickness data and sludge image data; the data acquisition module (10) includes a first detection unit (11) and a second detection unit (12); the first detection unit (11) is equipped with a distance sensor and a temperature sensor, and performs thickness compensation based on the original sludge thickness obtained by the distance sensor and the temperature obtained by the temperature sensor to obtain sludge thickness data; the second detection unit (12) is equipped with an industrial camera, used to capture sludge images and output sludge image data; The data processing module (20) is used to preprocess the sludge thickness data and image data, perform feature extraction, feature fusion and dimensionality reduction, and output the dimensionality-reduced feature vector; the data processing module (20) includes a feature extraction unit (22), which is used to extract sludge thickness features and sludge image features; The sludge thickness features include average thickness, thickness standard deviation, and thickness gradient; the sludge image features include texture features and shape features, wherein the texture features include contrast, correlation, energy, and entropy; and the shape features include area, perimeter, and roundness. The control module (30) is used to construct a material equalization decision model, receive the dimension-reduced feature vector and input it into the material equalization decision model, and output a control strategy. The control module (30) includes: a sample data acquisition unit (31), a model construction unit (32), and a decision output unit (33). The sample data acquisition unit (31) is used to acquire sludge thickness data and sludge image data from the data acquisition module (10), as well as the operation data of the material equalization module (40). The model construction unit (32) is used to construct a material equalization decision model. The decision output unit (33) inputs the dimension-reduced feature vector into the trained material equalization decision model based on the trained material equalization decision model and outputs a control strategy. The material equalization decision model outputs material equalization decision and material equalization intensity. The material equalization module (40) is located inside the drying chamber of the sludge dryer and is used to execute the control strategy, adjust the moving speed and swing amplitude of the scraper according to the material equalization decision and the material equalization intensity, and equalize the sludge.

2. The sludge dryer material distribution control system according to claim 1, characterized in that, The data processing module (20) also includes: The preprocessing unit (21) is used to denoise the sludge thickness data output by the first detection unit (11) and to perform grayscale, normalization and filtering processing on the image data output by the second detection unit (12). The feature fusion unit (23) is used to fuse and reduce the dimensionality of the feature vector composed of the extracted thickness features and image features.

3. The sludge dryer material distribution control system according to claim 2, characterized in that: The material homogenization decision model includes an input layer and an output layer; the number of nodes in the input layer is equal to the dimension of the feature vector after dimensionality reduction; the output nodes of the output layer include at least material homogenization decision and material homogenization intensity.

4. A method for controlling the uniform material distribution in a sludge dryer, characterized in that, Includes the following steps: S1, acquire sludge thickness data and sludge image data on the mesh belt; S2, preprocess the sludge thickness data and extract sludge thickness features; at the same time, preprocess the sludge image data and extract sludge image features. The sludge image features include texture features and shape features; The sludge thickness features include average thickness, thickness standard deviation, and thickness gradient; the sludge image features include texture features and shape features, wherein the texture features include contrast, correlation, energy, and entropy; and the shape features include area, perimeter, and roundness. S3. Based on the combination of sludge thickness features and sludge image features, the original feature matrix is ​​obtained, and the original feature matrix is ​​fused to obtain the feature vector. The feature vector is then subjected to dimensionality reduction processing. S4, Construct a uniform material decision model based on feature vectors; S5, input the dimensionality-reduced feature vector into the material homogenization decision model to obtain the control strategy; the material homogenization decision model outputs the material homogenization decision and the material homogenization intensity; S6 executes the control strategy, adjusting the scraper's moving speed and oscillation amplitude based on the material uniformity decision and the material uniformity intensity.

5. The material uniformity control method for a sludge dryer according to claim 4, characterized in that: In step S2, the sludge thickness characteristic is a temperature-compensated thickness characteristic.

6. The material uniformity control method for a sludge dryer according to claim 5, characterized in that: In step S3, the specific steps for fusing the original feature matrix are as follows: S31, standardize the feature data of the original feature matrix to obtain the standardized feature matrix. ; S32, based on the standardized feature matrix Calculate its covariance matrix C; S33, Perform eigenvalue decomposition on the covariance matrix to obtain the eigenvalues. and the corresponding feature vector .

7. The method for controlling the uniform material distribution in a sludge dryer according to claim 6, characterized in that: In step S3, the specific steps for dimensionality reduction of the feature vector are as follows: The eigenvectors are sorted according to the size of the eigenvalues. The eigenvectors corresponding to the k largest eigenvalues ​​are selected as principal components to form the transformation matrix W. The value of k is selected based on the cumulative variance contribution rate (CVR), and the k value is selected when the CVR reaches a set proportion. Dimensionally reduced feature vectors for, ； in, Let W be the transpose of W.

Images