Intelligent Control Method and Device for the Quality of Circulating Water in Cooling Towers Based on Machine Vision

Abstract

The present invention discloses an intelligent control method and device for the circulating water quality of a cooling tower based on machine vision. The device and method perform dual recognition of foam and circulating water quality based on machine vision according to the water surface image of the sump and issue corresponding control instructions, which can help eliminate foam in a timely manner and adjust the water quality, eliminate the influence of foam and water quality deterioration, and can ensure the stable operation of the circulating water in the cooling tower in real time, efficiently, automatically and intelligently.

Description

Technical Field

[0001] The present invention relates to an intelligent control method and device for the water quality of circulating water in a cooling tower based on machine vision. Background Art

[0002] Circulating cooling water is the main heat exchange medium in the production process of the chemical industry, accounting for more than 90% of the total industrial water consumption. The low-temperature circulating water is driven by a circulating pump into the production heating equipment, and through the heat transfer system inside the equipment, the heat energy is transferred to the circulating water, so that the temperature of the circulating water rises. The high-temperature circulating water is sent to the cooling equipment, and through a fan or the like, heat exchange is carried out with the atmosphere, and the heat energy is transferred to the atmosphere again, and then it is cooled down and continues to enter the circulation system. The cooling tower is one of the widely used cooling equipment, mainly using the evaporation heat dissipation of water to achieve the purpose of cooling.

[0003] During the operation of the cooling tower, due to processes such as evaporation, wind blowing, and sewage discharge, there is inevitably a loss of circulating water volume, resulting in water quality deterioration, and the dissolved solids and suspended solids in the water gradually accumulate, causing corrosion and scaling of the circulating water system and affecting the normal operation of the circulating water system. In addition, due to specific water quality or process conditions in some circulating water systems, the circulating water in the cooling tower is very easy to generate foam, which not only affects the normal operation of the liquid level switch in the cooling tower sump, but also seriously affects the heat dissipation effect of the cooling tower. And the method of using defoaming agents to defoam is easy to pollute the circulating water and increase the makeup water volume.

[0004] Therefore, how to be able to timely identify and adjust the water quality of the circulating water in the cooling tower is an urgent problem to be solved at present. Summary of the Invention

[0005] The purpose of the present invention is to provide an intelligent control method and device for the water quality of circulating water in a cooling tower based on machine vision to solve the problem that the water quality of the circulating water in the cooling tower cannot be timely controlled at present.

[0006] To solve the above technical problems, the present invention provides an intelligent control method for the water quality of circulating water in a cooling tower based on machine vision, including the following steps:

[0007] S1: Collect the water surface sample image of the sump;

[0008] S2: Manually judge the water quality situation and foam situation of the sump according to the water surface sample image, and perform a tagging operation on the water surface sample image according to the judgment results of the water quality situation and foam situation;

[0009] S3: Preprocess the water surface sample image, then construct an image data set for supervised learning according to the preprocessed water surface sample image, and merge the image data set with the label corresponding to the water surface sample image and convert the label using one-hot encoding;

[0010] S4: Construct a foam judgment model and a water quality judgment model based on the image dataset and the labels transformed by one-hot encoding.

[0011] S5: Collect the real-time water surface image of the collecting sump, preprocess the collected real-time water surface image, and then input the preprocessed real-time water surface image into the foam judgment model and the water quality judgment model respectively for real-time judgment. When it is judged that the foam or / and water quality in the current real-time water surface image does not meet the standard, a water quality adjustment instruction or / and a foam cleaning instruction is issued.

[0012] According to an embodiment of the present application, while collecting the water surface sample image of the collecting sump, collect the cooling tower circulating water quality analysis record and the make-up water valve / pump opening and closing record corresponding to the water surface sample image; when manually judging the water quality situation and the foam situation of the collecting sump according to the water surface sample image, it is necessary to combine the cooling tower circulating water quality analysis record and the make-up water valve / pump opening and closing record corresponding to the water surface sample image to judge the water quality situation and the foam situation.

[0013] According to an embodiment of the present application, the operation of manually tagging the water surface sample image according to the water quality judgment result includes tagging the image according to whether there is foam and the water quality on the water surface sample image according to the manual judgment result.

[0014] According to an embodiment of the present application, the preprocessing of the water surface sample image specifically includes: performing central cropping and grayscale processing on the water surface sample image; when performing central cropping on the water surface sample image, the area of the cropped image should not be less than half of the original image.

[0015] According to an embodiment of the present application, the image dataset is the grayscale matrix of the water surface sample image, expressed as:

[0016] X = {[X1],[X2],[X3]……[Xn]}

[0017] The label transformed by one-hot encoding is marked as:

[0018] Y = {[y1, y_1],[y2,y_2],[y3,y_3]……[yn,y_n]}

[0019] Among them, X1 represents the grayscale matrix 1, and y1 and y_1 respectively represent the one-hot encoding of the foam and water quality corresponding to X1.

[0020] According to an embodiment of the present application, the method for constructing a foam judgment model includes: after the image data set is processed by fast Fourier transform, determining whether there is foam with reference to the average value of the spectral amplitude; the judgment threshold is set according to the filtering results of all image data sets. If the average value of the amplitude is greater than the threshold, it is determined that there is foam, otherwise there is no foam.

[0021] According to an embodiment of the present application, the method for constructing a water quality judgment model includes: adopting the method of building a convolutional neural network, using a 2D convolutional kernel to extract the image features of the image data set, setting the convolutional kernel structure and quantity, activation layer and fully connected layer, and performing supervised training through one-hot encoding of the water quality situation to obtain a water quality judgment model.

[0022] The present invention also provides an intelligent control device for the circulating water quality of a cooling tower based on machine vision, including a data acquisition module and an intelligent control instruction generation module connected by communication; the data acquisition module is used to acquire the water surface sample image of the sump and the real-time water surface image of the sump; one or more computer programs are provided in the intelligent control instruction generation module, and when the one or more computer programs are executed, the above steps S2-S5 are implemented.

[0023] According to an embodiment of the present application, the device further includes a foam cleaning unit and a clean water replenishing unit communicatively connected to the intelligent control instruction generation module; the foam cleaning unit is used to receive a foam cleaning instruction and clean the foam on the water surface according to the foam cleaning instruction; the clean water replenishing unit is used to receive a water quality adjustment instruction and replenish a certain amount of clean water into the condensing tower according to the water quality adjustment instruction.

[0024] According to an embodiment of the present application, the data acquisition module includes a high-definition industrial camera installed in the cooling tower.

[0025] The beneficial effects of the present invention are as follows: by double-identifying foam and circulating water quality based on machine vision according to the water surface image of the sump and issuing corresponding control instructions, it is beneficial to timely eliminate foam and adjust water quality, eliminate the influence of foam and water quality deterioration, and can ensure the stable operation of the circulating water in the cooling tower in real time, efficiently, automatically and intelligently. BRIEF DESCRIPTION OF THE DRAWINGS

[0026] The drawings described herein are used to provide a further understanding of the present application and constitute a part of the present application. The same reference numerals are used to represent the same or similar parts in these drawings. The schematic embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation of the present application. In the drawings:

[0027] Figure 1 It is a schematic diagram of an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0028] An intelligent control method for the circulating water quality of a cooling tower based on machine vision, comprising the following steps:

[0029] S1: Collect the water surface sample image of the collecting pool;

[0030] S2: Manually judge the water quality condition and foam condition of the collecting pool according to the water surface sample image, and perform a tagging operation on the water surface sample image according to the judgment results of the water quality condition and foam condition;

[0031] S3: Preprocess the water surface sample image, then construct an image data set for supervised learning according to the preprocessed water surface sample image, and merge the image data set with the label corresponding to the water surface sample image and convert the label using one-hot encoding;

[0032] S4: Construct a foam judgment model and a water quality judgment model according to the image data set and the label converted using one-hot encoding;

[0033] S5: Collect the real-time water surface image of the collecting pool, preprocess the collected real-time water surface image, and then input the preprocessed real-time water surface image into the foam judgment model and the water quality judgment model respectively for real-time judgment. When it is judged that the foam or / and water quality in the current real-time water surface image does not meet the standard, a water quality adjustment instruction or / and a foam cleaning instruction is issued.

[0034] This water quality intelligent control method performs dual recognition of foam and circulating water quality based on machine vision according to the water surface image of the collecting pool and issues corresponding control instructions, and controls the corresponding water quality adjustment equipment through the control instructions, which is beneficial to timely eliminate foam and adjust the water quality, eliminates the influence of foam and water quality deterioration, and can ensure the stable operation of the circulating water in the cooling tower in real time, efficiently, automatically and intelligently.

[0035] According to an embodiment of the present application, while collecting the water surface sample image of the collecting pool, collect the analysis record of the circulating water quality of the cooling tower corresponding to the water surface sample image and the opening and closing record of the makeup water valve / pump; when manually judging the water quality condition and foam condition of the collecting pool according to the water surface sample image, it is necessary to combine the analysis record of the circulating water quality of the cooling tower corresponding to the water surface sample image and the opening and closing record of the makeup water valve / pump to judge the water quality condition and foam condition. By combining the analysis record of the circulating water quality of the cooling tower corresponding to the water surface sample image and the opening and closing record of the makeup water valve / pump to judge the water quality condition and foam condition, the accuracy of the judgment result can be improved, thereby ensuring the accuracy of the tagging of the water surface sample image, which is beneficial to improving the accuracy of the judgment results of the foam judgment model and the water quality judgment model.

[0036] According to an embodiment of the present application, the operation of manually tagging the water surface sample image based on the water quality judgment result includes tagging the image according to the manual judgment result on whether there is foam on the water surface sample image and the quality of the water. The tags can be set to include "yes" and "no" for foam, and "good" and "bad" for water quality.

[0037] According to an embodiment of the present application, the preprocessing of the water surface sample image specifically includes: performing central cropping and grayscale processing on the water surface sample image; when performing central cropping on the water surface sample image, the area of the cropped image should not be less than one-half of the original image, and then grayscale the cropped water surface sample image using an appropriate gray level.

[0038] According to an embodiment of the present application, the image data set is the grayscale matrix of the water surface sample image, expressed as:

[0039] X = {[X1],[X2],[X3]……[Xn]}

[0040] The tags after one-hot encoding transformation are marked as:

[0041] Y = {[y1, y_1],[y2,y_2],[y3,y_3]……[yn,y_n]}

[0042] Wherein, X1 represents the grayscale matrix 1, and y1 and y_1 respectively represent the one-hot encodings corresponding to the foam and water quality conditions of X1.

[0043] According to an embodiment of the present application, the method for constructing the foam judgment model includes: after the image data set is processed by fast Fourier transform, determining whether there is foam based on the average value of the spectral amplitudes; the judgment threshold is set according to the filtering results of all image data sets. If the average value of the amplitudes is greater than the threshold, it is determined that there is foam, otherwise there is no foam.

[0044] According to an embodiment of the present application, the method for constructing the water quality judgment model includes: adopting the method of building a convolutional neural network, using a 2D convolutional kernel to extract the image features of the image data set, setting the convolutional kernel structure and number, activation layer and fully connected layer, and performing supervised training through the one-hot encoding of the water quality conditions to obtain a water quality judgment model that outputs "good" and "bad".

[0045] The present invention also provides an intelligent control device for the circulating water quality of a cooling tower based on machine vision, including a data acquisition module and an intelligent control instruction generation module connected by communication (such as Figure 1the industrial control computer shown in [reference]; the data acquisition module is used to acquire the water surface sample image and the real-time water surface image of the collecting pool; the intelligent control instruction generation module is provided with one or more computer programs, and when the one or more computer programs are executed, the above steps S2-S5 are implemented.

[0046] According to an embodiment of the present application, the device further includes a foam cleaning unit and a clean water supplement unit communicatively connected to the intelligent control instruction generation module; the foam cleaning unit is used to receive a foam cleaning instruction and clean the foam on the water surface according to the foam cleaning instruction; the clean water supplement unit is used to receive a water quality adjustment instruction and supplement a fixed amount of clean water into the cooling tower according to the water quality adjustment instruction. The foam cleaning unit is a foam scraper, and the foam scraper is a thin plate made of anti-corrosion material with a thickness of about 10 cm (such as Figure 1 the scraper blade shown in [reference]), longitudinally installed above the water surface near the collecting pool, and each time the scraping is triggered, the plate surface moves back and forth on the water surface once, and the foam is scraped into the opening groove at the end by the scraping plate and enters the sewage pipe. The clean water supplement unit is a clean water tank, a water supply pipeline connected to the clean water tank, and a water supply pump / valve installed on the water supply pipeline. The water supply pump / valve receives an opening and closing instruction (water quality adjustment instruction) from the intelligent control instruction generation module, and each time the water supply is triggered, the water pump / valve is opened to supplement a fixed amount of clean water into the cooling tower.

[0047] According to an embodiment of the present application, the data acquisition module includes a high-definition industrial camera installed in the cooling tower (such as Figure 1 the camera shown in [reference]). Using the high-definition industrial camera as the data acquisition module, a certain image shooting frequency is set, and each newly taken image is used as the real-time water surface image of the collecting pool. The camera needs to have anti-corrosion and anti-fog functions and be installed in the cooling tower to clearly capture the position of the water surface of the collecting pool. The installation position changes slightly depending on the size of the cooling tower, and is about 30-100 cm away from the water surface.

[0048] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not restrictive. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered by the scope of the claims of the present invention.

Claims

1. An intelligent control method for the circulating water quality of a cooling tower based on machine vision, characterized in that, It includes the following steps: S1: Collect the water surface sample image of the sump; S2: Manually judge the water quality condition and foam condition of the sump according to the water surface sample image, and perform a tagging operation on the water surface sample image according to the judgment results of the water quality condition and foam condition; S3: Preprocess the water surface sample image, then construct an image dataset for supervised learning based on the preprocessed water surface sample image, and merge the image dataset with the label corresponding to the water surface sample image and convert the label using one-hot encoding; S4: Construct a foam judgment model and a water quality judgment model according to the image dataset and the label converted using one-hot encoding; S5: Collect the real-time water surface image of the sump, preprocess the collected real-time water surface image, and then input the preprocessed real-time water surface image into the foam judgment model and the water quality judgment model respectively for real-time judgment. When it is judged that the foam or / and water quality in the current real-time water surface image does not meet the standard, a water quality adjustment instruction or / and a foam cleaning instruction is issued; The method for constructing the foam judgment model includes: after the image dataset is processed by fast Fourier transform, judge whether there is foam with the average value of the spectral amplitude as a reference; the judgment threshold is set according to the filtering results of all image datasets. If the average value of the amplitude is greater than the threshold, it is judged that there is foam, otherwise there is no foam; The method for constructing the water quality judgment model includes: adopting the method of building a convolutional neural network, using a 2D convolutional kernel to extract the image features of the image dataset, setting the convolutional kernel structure and quantity, activation layer and fully connected layer, and performing supervised training through the one-hot encoding of the water quality condition to obtain the water quality judgment model.

2. The intelligent control method for the circulating water quality of a cooling tower based on machine vision according to claim 1, wherein When collecting the water surface sample image of the sump, collect the analysis record of the cooling tower circulating water quality and the opening and closing record of the makeup water valve / pump corresponding to the water surface sample image; when manually judging the water quality condition and foam condition of the sump according to the water surface sample image, it is necessary to combine the analysis record of the cooling tower circulating water quality and the opening and closing record of the makeup water valve / pump corresponding to the water surface sample image to judge the water quality condition and foam condition.

3. The intelligent control method for the quality of circulating water in a cooling tower based on machine vision according to claim 1 or 2, characterized in that, Manually performing a tagging operation on the water surface sample image according to the water quality judgment result includes tagging the image according to whether there is foam and the water quality on the water surface sample image according to the manual judgment result.

4. The intelligent control method for the quality of circulating water in a cooling tower based on machine vision according to claim 1, characterized in that, Specifically, preprocessing the water surface sample image includes: performing central cropping and grayscale processing on the water surface sample image; when performing central cropping on the water surface sample image, the area of the cropped image should not be less than half of the original image.

5. The intelligent control method for the quality of circulating water in a cooling tower based on machine vision according to claim 1 or 4, characterized in that, The image dataset is the grayscale matrix of the water surface sample image, expressed as: X = {[X1],[X2],[X3]……[Xn]} The label after being converted by the one-hot encoding is marked as: Y = {[y1, y_1],[y2,y_2],[y3,y_3]……[yn,y_n]} Among them, X1 represents the grayscale matrix 1, and y1 and y_1 respectively represent the one-hot encodings of the foam and water quality conditions corresponding to X1.

6. An intelligent control device for the circulating water quality of a cooling tower based on machine vision, characterized in that, It includes: A data acquisition module for collecting water surface sample images of the collecting sump and real-time water surface images of the collecting sump; An intelligent control instruction generation module, in which one or more computer programs are provided. When the one or more computer programs are executed, the steps S2-S5 described in any one of claims 1-5 are implemented.

7. The intelligent control device for the circulating water quality of a cooling tower based on machine vision according to claim 6, characterized in that, The device further includes a foam cleaning unit and a clean water supplement unit communicatively connected to the intelligent control instruction generation module; the foam cleaning unit is used to receive a foam cleaning instruction and clean the foam on the water surface according to the foam cleaning instruction; the clean water supplement unit is used to receive a water quality adjustment instruction and supplement a fixed amount of clean water into the cooling tower according to the water quality adjustment instruction.

8. The intelligent control device for the quality of circulating water in a cooling tower based on machine vision according to claim 6 or 7, characterized in that, The data acquisition module includes a high-definition industrial camera installed in the cooling tower.

Images