CN114841453A - Clean coal ash content prediction method in flotation process - Google Patents
Clean coal ash content prediction method in flotation process Download PDFInfo
- Publication number
- CN114841453A CN114841453A CN202210542985.5A CN202210542985A CN114841453A CN 114841453 A CN114841453 A CN 114841453A CN 202210542985 A CN202210542985 A CN 202210542985A CN 114841453 A CN114841453 A CN 114841453A
- Authority
- CN
- China
- Prior art keywords
- flotation
- clean coal
- coal ash
- ash content
- prediction
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000005188 flotation Methods 0.000 title claims abstract description 219
- 239000010883 coal ash Substances 0.000 title claims abstract description 137
- 238000000034 method Methods 0.000 title claims abstract description 63
- 230000008569 process Effects 0.000 title claims abstract description 36
- 239000003245 coal Substances 0.000 claims abstract description 61
- 239000007788 liquid Substances 0.000 claims abstract description 36
- 238000012545 processing Methods 0.000 claims abstract description 25
- 238000010606 normalization Methods 0.000 claims abstract description 18
- 239000002956 ash Substances 0.000 claims description 38
- 238000012549 training Methods 0.000 claims description 19
- 239000006260 foam Substances 0.000 claims description 17
- 238000013528 artificial neural network Methods 0.000 claims description 8
- 238000013527 convolutional neural network Methods 0.000 claims description 6
- 238000003384 imaging method Methods 0.000 claims description 3
- 238000005259 measurement Methods 0.000 claims description 3
- 238000002360 preparation method Methods 0.000 abstract description 6
- 238000004519 manufacturing process Methods 0.000 description 10
- 230000006870 function Effects 0.000 description 5
- 229910052500 inorganic mineral Inorganic materials 0.000 description 5
- 239000011707 mineral Substances 0.000 description 5
- 238000001514 detection method Methods 0.000 description 4
- 238000005070 sampling Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 3
- 239000003814 drug Substances 0.000 description 3
- 239000000463 material Substances 0.000 description 3
- 238000007781 pre-processing Methods 0.000 description 3
- 230000001105 regulatory effect Effects 0.000 description 3
- 238000004458 analytical method Methods 0.000 description 2
- 239000012141 concentrate Substances 0.000 description 2
- 238000005520 cutting process Methods 0.000 description 2
- 238000001914 filtration Methods 0.000 description 2
- 238000013507 mapping Methods 0.000 description 2
- 238000012216 screening Methods 0.000 description 2
- 238000003860 storage Methods 0.000 description 2
- 238000012360 testing method Methods 0.000 description 2
- 230000003321 amplification Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 229910052799 carbon Inorganic materials 0.000 description 1
- 239000003153 chemical reaction reagent Substances 0.000 description 1
- 238000004590 computer program Methods 0.000 description 1
- 238000002790 cross-validation Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 238000003199 nucleic acid amplification method Methods 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000011084 recovery Methods 0.000 description 1
- 238000007790 scraping Methods 0.000 description 1
- 238000006467 substitution reaction Methods 0.000 description 1
- 230000002195 synergetic effect Effects 0.000 description 1
- 230000009466 transformation Effects 0.000 description 1
- 239000002699 waste material Substances 0.000 description 1
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/04—Forecasting or optimisation specially adapted for administrative or management purposes, e.g. linear programming or "cutting stock problem"
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q10/00—Administration; Management
- G06Q10/06—Resources, workflows, human or project management; Enterprise or organisation planning; Enterprise or organisation modelling
- G06Q10/063—Operations research, analysis or management
- G06Q10/0639—Performance analysis of employees; Performance analysis of enterprise or organisation operations
- G06Q10/06393—Score-carding, benchmarking or key performance indicator [KPI] analysis
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q50/00—Information and communication technology [ICT] specially adapted for implementation of business processes of specific business sectors, e.g. utilities or tourism
- G06Q50/02—Agriculture; Fishing; Forestry; Mining
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06V—IMAGE OR VIDEO RECOGNITION OR UNDERSTANDING
- G06V10/00—Arrangements for image or video recognition or understanding
- G06V10/10—Image acquisition
Landscapes
- Business, Economics & Management (AREA)
- Engineering & Computer Science (AREA)
- Human Resources & Organizations (AREA)
- Strategic Management (AREA)
- Economics (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- General Business, Economics & Management (AREA)
- Development Economics (AREA)
- Marketing (AREA)
- Tourism & Hospitality (AREA)
- Entrepreneurship & Innovation (AREA)
- Quality & Reliability (AREA)
- Operations Research (AREA)
- Game Theory and Decision Science (AREA)
- Educational Administration (AREA)
- Agronomy & Crop Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Health & Medical Sciences (AREA)
- Mining & Mineral Resources (AREA)
- Marine Sciences & Fisheries (AREA)
- General Health & Medical Sciences (AREA)
- Primary Health Care (AREA)
- Animal Husbandry (AREA)
- Multimedia (AREA)
- Investigating Or Analysing Materials By Optical Means (AREA)
Abstract
The invention relates to a method for predicting clean coal ash content in a flotation process, belongs to the technical field of prediction of key indexes in the coal preparation industry, and solves the problem that the clean coal ash content in the existing flotation cannot be effectively, accurately and quickly detected. The method comprises the following steps: collecting data in real time, wherein the data comprises a flotation froth image, a tailing image, a flotation feed concentration, a flotation feed flow and a flotation liquid level; inputting the flotation froth image collected in real time into a trained clean coal ash content prediction main model, and processing to obtain a clean coal ash content prediction value; respectively carrying out normalization processing on the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level of the tailing image collected in real time, inputting the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level after the normalization processing into a trained clean coal ash content prediction compensation model, and processing to obtain a clean coal ash content error predicted value; and compensating the predicted value of the clean coal ash content by using the error predicted value of the clean coal ash content to obtain a clean coal ash content prediction result.
Description
Technical Field
The invention relates to the technical field of prediction of key indexes in the coal dressing industry, in particular to a method for predicting clean coal ash content in a flotation process.
Background
China is a large country for coal production and consumption, and with the proposal of a double-carbon target, an efficient clean utilization system needs to be constructed in the coal industry, and coal preparation plants are inevitable in digitalization, automation and intelligent transformation. Flotation is the most widely used method for sorting fine coal, mainly separates fine coal from gangue by using the difference of mineral surface hydrophobicity, and is a very complex physicochemical process influenced by multivariable synergistic effect.
However, the coal preparation plant has always monitored the clean coal ash content through the fast ash test, which requires at least one hour for the flow of the fast ash test, and this has a serious hysteresis for guiding the flotation production. Therefore, the flotation process is mainly regulated and controlled by field workers according to experience, but the method has strong subjectivity, large error and low efficiency, and can cause the waste of flotation reagents and mineral resources.
Therefore, how to effectively, accurately and quickly detect the ash content of the flotation clean coal has important significance for reducing medicament consumption and efficiently recovering resources in a coal preparation plant.
Disclosure of Invention
In view of the foregoing analysis, an embodiment of the present invention is directed to providing a method for predicting clean coal ash in a flotation process, so as to solve a problem that the existing flotation clean coal ash cannot be effectively, accurately and quickly detected.
The invention discloses a method for predicting clean coal ash content in a flotation process, which comprises the following steps:
collecting data in real time, wherein the data comprises a flotation froth image, a tailing image, a flotation feed concentration, a flotation feed flow and a flotation liquid level;
inputting the flotation froth image collected in real time into a trained clean coal ash content prediction main model, and processing to obtain a clean coal ash content prediction value;
respectively carrying out normalization processing on the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level of the tailing image collected in real time, inputting the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level after the normalization processing into a trained clean coal ash content prediction compensation model, and obtaining a clean coal ash content error predicted value after processing;
and compensating the predicted value of the clean coal ash content by using the error predicted value of the clean coal ash content to obtain a final clean coal ash content prediction result.
On the basis of the scheme, the invention also makes the following improvements:
further, training the clean coal ash prediction main model by:
obtaining a first set of samples, each set of first sample data in the first set of samples comprising: flotation froth image, real measurement value of clean coal ash content;
and respectively taking the flotation froth image of the clean coal in each group of first sample data as input and the measured value of the ash content of the clean coal as a label, training the main model for predicting the ash content of the clean coal, determining the structure and the parameters of the main model for predicting the ash content of the clean coal, and obtaining the trained main model for predicting the ash content of the clean coal.
Further, training the clean coal ash prediction compensation model by:
obtaining a second sample set, each set of second sample data in the second sample set comprising: flotation froth image, tailing image, flotation feed concentration, flotation feed flow, flotation liquid level and clean coal ash measured value;
inputting the flotation froth images in each group of second sample data into the trained clean coal ash content prediction main model to obtain a clean coal ash content prediction value;
respectively carrying out normalization processing on the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level of the tailing image in each group of second sample data;
and respectively taking the normalized gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level corresponding to each group of second sample data as input, taking the difference value between the corresponding measured value of the clean coal ash content and the predicted value of the clean coal ash content as a label, training the clean coal ash content prediction compensation model, determining the structure and the parameters of the clean coal ash content prediction compensation model, and obtaining the trained clean coal ash content prediction compensation model.
Further, the clean coal ash prediction main model is a clean coal ash prediction main model based on a convolutional neural network.
Further, the clean coal ash content prediction compensation model adopts a prediction compensation model based on a radial basis function neural network.
Further, the gray characteristic values comprise a gray average value, a variance, smoothness and an energy entropy;
the gray characteristic value is obtained by extracting a gray histogram of the tailing image.
Further, the normalization processing adopts a maximum and minimum normalization method.
Further, the flotation froth image is acquired by a froth image acquisition system fixed above the flotation froth liquid level;
the foam image acquisition system is internally provided with a CCD industrial camera, and LED light sources which form an angle of 45 degrees downwards are fixed on two sides of the industrial camera.
Further, the tailing image is acquired by a tailing image acquisition system;
a CCD industrial camera is arranged in the tailing image acquisition system;
the tailing image acquisition system is built beside a flotation tailing tank, a real-time flotation tailing sample is extracted through a pump and enters an imaging black box, a light source in the black box is fixed, and a CCD camera is used for acquiring tailing images.
Further, detecting the concentration of the flotation feed material in real time at the flotation feed material pipeline by using a concentration meter;
detecting the flotation feeding flow rate at the flotation feeding pipeline in real time by using a flowmeter;
and detecting the flotation liquid level in real time by using a floating ball level meter.
Compared with the prior art, the invention can realize at least one of the following beneficial effects:
the method for predicting the ash content of clean coal in the flotation process has the following advantages:
firstly, a clean coal ash predicted value is determined based on a flotation foam image collected in real time, a clean coal ash error predicted value is determined based on a tailing image, flotation feed concentration, flotation feed flow and flotation liquid level collected in real time, the clean coal ash error predicted value is used for compensating the clean coal ash predicted value to obtain a final clean coal ash prediction result, effective, accurate and rapid detection of flotation clean coal ash can be achieved, and a coal preparation plant can timely regulate and control coal slime flotation according to the prediction result.
Secondly, can realize the real-time online continuous detection of flotation clean coal ash to have certain self-adaptation ability, can correct the predicted value through the situation of production. The flotation working condition can be timely regulated and controlled according to the predicted value, the medicament consumption is reduced, and the utilization rate of mineral resources is improved.
Thirdly, the method utilizes a clean coal ash prediction main model based on a convolution neural network and a clean coal ash prediction compensation model based on a radial basis function neural network to realize online prediction of the flotation clean coal ash through a foam image, a tailing image and other production process parameters in the flotation production process, and a prediction result is used for guiding the regulation and control of the flotation production process.
In the invention, the technical schemes can be combined with each other to realize more preferable combination schemes. Additional features and advantages of the invention will be set forth in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. The objectives and other advantages of the invention will be realized and attained by the structure particularly pointed out in the written description and drawings.
Drawings
The drawings are only for purposes of illustrating particular embodiments and are not to be construed as limiting the invention, wherein like reference numerals are used to designate like parts throughout.
FIG. 1 is a schematic diagram of a process for predicting ash content of clean coal in a flotation process provided by an embodiment of the present invention;
FIG. 2 is a schematic diagram of a clean coal ash prediction main model training process of the flotation process provided by an embodiment of the invention;
fig. 3 is a schematic diagram of a clean coal ash prediction compensation model training process of a flotation process according to an embodiment of the present invention.
Detailed Description
The accompanying drawings, which are incorporated in and constitute a part of this application, illustrate preferred embodiments of the invention and together with the description, serve to explain the principles of the invention and not to limit the scope of the invention.
The invention discloses a method for predicting clean coal ash content in a flotation process, and a flow chart is shown in figure 1 and comprises the following steps:
step S1: collecting data in real time, wherein the data comprises a flotation froth image, a tailing image, a flotation feed concentration, a flotation feed flow and a flotation liquid level;
step S2: inputting the flotation froth image collected in real time into a trained clean coal ash content prediction main model, and processing to obtain a clean coal ash content prediction value; in specific implementation, image preprocessing (such as filtering) can be performed on the flotation froth image, and then the preprocessed flotation froth image is input into the trained clean coal ash content prediction main model, so that clutter in the flotation froth image is removed.
Respectively carrying out normalization processing on the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level of the tailing image collected in real time, inputting the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level after the normalization processing into a trained clean coal ash content prediction compensation model, and obtaining a clean coal ash content error predicted value after processing; in specific implementation, the tailing image may be subjected to image preprocessing (such as filtering), and then gray level feature values are obtained through gray level histogram processing.
Step S3: and compensating the predicted value of the clean coal ash content by using the error predicted value of the clean coal ash content to obtain a final clean coal ash content prediction result.
The flotation froth image contains abundant froth color, froth texture and froth size characteristics, and the characteristics have strong correlation with flotation clean coal, and are specifically represented as follows: if the flotation foam is dark black, the surface of the foam is wrinkled and has more textures, and the number of the flotation foam in unit area is large, the flotation foam represents that the number of clean coal carried by the current flotation foam is large, and the ash content of the clean coal is low; on the contrary, if the flotation froth is bright black, the froth surface is smooth and mostly is a virtual bubble with a large shape, which means that the quantity of clean coal carried by the current froth is small, at the moment, the flotation clean coal is greatly lost, the ash content of the tail coal is reduced, and the ash content of the clean coal is increased. Therefore, the prediction of clean coal ash can be made based on the flotation froth image.
In addition, there are influence of different degrees in the prediction of tailing image, flotation pan feeding concentration, flotation pan feeding flow and flotation liquid level to clean coal ash content, specifically appear:
(1) the gray scale characteristics of tailing images with different ash contents are obviously different: the lower the ash content of the tailings, the darker the corresponding tailing image, otherwise, the brighter the corresponding tailing image; meanwhile, under the condition that the fluctuation of the ash content of the fed material is not large, the ash content of the tail coal and the ash content of the clean coal have extremely strong linear correlation, and the specific formula is as follows:
wherein, Ad Y 、Ad J 、Ad W Respectively representing the ash content of the flotation raw coal, the ash content of the flotation clean coal and the ash content of the flotation tail coal, gamma J And gamma W Respectively representing the yield of flotation clean coal and the yield of flotation tail coal; in the flotation process, the ash content of the raw coal is floated generally with small fluctuation, and the clean coal yield and the tail coal yield can be detected in real time through a belt weigher, so that the following linear relation can be obtained after simplification:
Ad J ＝K×Ad W +B (2)
as shown in the formula (2), the tail coal ash and the clean coal ash have correlation, and the correlation can be evaluated by analyzing the gray characteristic value of the tail coal image;
(2) the concentration and the flow rate of the flotation feed reflect the current treatment capacity of the flotation system in real time, and when the treatment capacity is in a reasonable range, the ash content result of clean coal is relatively low; when the treatment capacity is increased and other conditions are not changed, the loss of the flotation clean coal is caused, and the ash content of the flotation clean coal is increased along with the loss of the flotation clean coal; when the treatment capacity is reduced, gangue minerals are also brought into the concentrate, and the ash content of the flotation clean coal is also increased;
(3) the flotation liquid level reflects the current operation condition of flotation equipment, the concentrate grade is reduced due to high liquid level, the foam scraping amount is reduced due to low liquid level, and the recovery rate of clean coal is reduced;
based on the analysis, the compensation value of the clean coal ash content can be evaluated based on the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level of the tailing image. And then, compensating the predicted value of the clean coal ash content by using the error predicted value of the clean coal ash content to obtain a final clean coal ash content prediction result. Therefore, the accuracy of the prediction result of the clean coal ash in the flotation process is high.
Preferably, the flotation froth image is acquired by a froth image acquisition system fixed above the flotation froth liquid level; the foam image acquisition system is internally provided with a CCD industrial camera, and LED light sources which form an angle of 45 degrees downwards are fixed on two sides of the industrial camera; to ensure that the flotation froth image is not disturbed by external factors. Preferably, the tailing image is acquired by a tailing image acquisition system; a CCD industrial camera is arranged in the tailing image acquisition system; the tailing image acquisition system is built beside a flotation tailing tank, a real-time flotation tailing sample is extracted through a pump and enters an imaging black box, a light source in the black box is fixed, and a CCD camera is used for acquiring tailing images. The system is in a closed space, and the acquired tailing images are not interfered by the outside. In addition, a concentration meter is used for detecting the concentration of the flotation feed in real time at the flotation feed pipeline; detecting the flotation feeding flow rate at the flotation feeding pipeline in real time by using a flowmeter; and detecting the flotation liquid level in real time by using a floating ball level meter.
Before implementing the above scheme, training of a clean coal ash prediction main model and a clean coal ash prediction compensation model is completed, and the specific implementation manner is described as follows:
(1) training the clean coal ash prediction main model by the following method, wherein the training process is as shown in FIG. 2:
step A1: obtaining a first set of samples, each set of first sample data in the first set of samples comprising: flotation froth image, real measurement value of clean coal ash content; specifically, a first sample set is obtained by:
step A11: collecting a large number of flotation froth images and clean coal actual ash content in corresponding time;
step A12: screening the collected flotation froth images to select the flotation froth images with high definition; preprocessing the screened flotation froth image to remove the noise of the flotation froth image;
step A13: cutting the pretreated flotation froth image to obtain a flotation froth image with a proper size;
step A14: taking each cut flotation froth image and the actual ash content of clean coal at corresponding time as a group of first sample data respectively; all the first sample data are aggregated to form a first sample set.
Step A2: and respectively taking the flotation froth image of the clean coal in each group of first sample data as input and the measured value of the ash content of the clean coal as a label, training the main model for predicting the ash content of the clean coal, determining the structure and the parameters of the main model for predicting the ash content of the clean coal, and obtaining the trained main model for predicting the ash content of the clean coal.
In the process of training the clean coal ash prediction main model, a mapping relation between the foam color, the foam texture and the foam size characteristics in the flotation foam image and the measured value of the clean coal ash can be established, and the mapping relation is embodied by the structure and the parameters of the clean coal ash prediction main model. Therefore, in the real-time processing process, the flotation froth image collected in real time can be input into the trained clean coal ash prediction main model, and a clean coal ash prediction value is obtained after processing.
Preferably, the clean coal ash prediction main model is a clean coal ash prediction main model based on a convolutional neural network. In the training process of the clean coal ash prediction main model, the convolutional neural network carries out back propagation on errors obtained through gradient descent, parameters of all layers of the convolutional neural network are updated layer by layer, and the structure and the parameters of the clean coal ash prediction main model are finally determined through multiple rounds of iterative training.
The clean coal ash prediction main model based on the convolutional neural network can preliminarily predict the flotation clean coal ash, but high-precision prediction results cannot be obtained due to the factors of complex flotation working conditions, limited image acquisition process errors, limited information reflected by images and the like. Meanwhile, other influence factors of the flotation clean coal ash are analyzed in the front, so that the clean coal ash prediction compensation model is trained according to the influence factors of the flotation clean coal ash so as to predict the deviation of the clean coal ash prediction value of the clean coal ash prediction main model.
(2) Training the clean coal ash prediction compensation model by the following method, wherein the training process is as shown in FIG. 3:
step B1: obtaining a second sample set, each set of second sample data in the second sample set comprising: flotation froth image, tailing image, flotation feed concentration, flotation feed flow, flotation liquid level and clean coal ash measured value; specifically, the second sample set is obtained by:
step B11: acquiring sampled data for a plurality of time nodes, comprising: flotation froth image, tailing image, flotation feed concentration, flotation feed flow, flotation liquid level and clean coal ash measured value;
step B12: screening sampling data, and only keeping the sampling data with higher definition of both the flotation froth image and the tailing image as alternative sampling data;
step B13: cutting the flotation froth image and the tailing image in the alternative sampling data to obtain a flotation froth image with a proper size;
step B14: taking each group of the cut alternative sample data as a group of second sample data respectively; and summarizing all the second sample data to form a second sample set.
Step B2: inputting the flotation froth images in each group of second sample data into the trained clean coal ash content prediction main model to obtain a clean coal ash content prediction value;
step B3: respectively carrying out normalization processing on the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level of the tailing image in each group of second sample data;
specifically, in this embodiment, the grayscale characteristic values of the tailing image include a grayscale mean value, a variance, smoothness, and an energy entropy; and the gray characteristic value is obtained by carrying out gray histogram extraction on the tailing image. Preferably, a maximum value and minimum value normalization method is selected to respectively perform normalization processing on the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level of the tailing image. After normalization processing, all data are converted into a [0,1] interval, so that all indexes belong to the same magnitude.
Step B4: and respectively taking the normalized gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level corresponding to each group of second sample data as input, taking the difference value between the corresponding measured value of the clean coal ash content and the predicted value of the clean coal ash content as a label, training the clean coal ash content prediction compensation model, determining the structure and the parameters of the clean coal ash content prediction compensation model, and obtaining the trained clean coal ash content prediction compensation model.
Preferably, the clean coal ash prediction compensation model can be a prediction compensation model based on a radial basis function neural network. In the training process of the clean coal ash content prediction compensation model, the structure and parameters of the radial basis function neural network can be determined by adopting cross validation and gradient descent amplification.
In summary, compared with the prior art, the method for predicting the ash content of the clean coal in the flotation process provided by the embodiment has the following advantages:
firstly, a clean coal ash predicted value is determined based on a flotation foam image collected in real time, a clean coal ash error predicted value is determined based on a tailing image, flotation feed concentration, flotation feed flow and flotation liquid level collected in real time, the clean coal ash error predicted value is used for compensating the clean coal ash predicted value to obtain a final clean coal ash prediction result, effective, accurate and rapid detection of flotation clean coal ash can be achieved, and a coal preparation plant can timely regulate and control coal slime flotation according to the prediction result.
Secondly, can realize the real-time online continuous detection of flotation clean coal ash to have certain self-adaptation ability, can correct the predicted value through the situation of production. The flotation working condition can be timely regulated and controlled according to the predicted value, the medicament consumption is reduced, and the utilization rate of mineral resources is improved.
Thirdly, the method utilizes a clean coal ash prediction main model based on a convolution neural network and a clean coal ash prediction compensation model based on a radial basis function neural network to realize on-line prediction of the flotation clean coal ash through a foam image, a tailing image and other production process parameters in the flotation production process, and the prediction result is used for guiding regulation and control of the flotation production process.
Those skilled in the art will appreciate that all or part of the flow of the method implementing the above embodiments may be implemented by a computer program, which is stored in a computer readable storage medium, to instruct related hardware. The computer readable storage medium is a magnetic disk, an optical disk, a read-only memory or a random access memory.
The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention.
Claims (10)
1. A method for predicting clean coal ash content in a flotation process, comprising:
collecting data in real time, wherein the data comprises a flotation froth image, a tailing image, a flotation feed concentration, a flotation feed flow and a flotation liquid level;
inputting the flotation froth image collected in real time into a trained clean coal ash content prediction main model, and processing to obtain a clean coal ash content prediction value;
respectively carrying out normalization processing on the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level of the tailing image collected in real time, inputting the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level after the normalization processing into a trained clean coal ash content prediction compensation model, and obtaining a clean coal ash content error predicted value after processing;
and compensating the clean coal ash prediction value by using the clean coal ash error prediction value to obtain a final clean coal ash prediction result.
2. The method for clean coal ash prediction of flotation process of claim 1, characterized in that the clean coal ash prediction main model is trained by:
obtaining a first set of samples, each set of first sample data in the first set of samples comprising: flotation froth image, real measurement value of clean coal ash content;
and respectively taking the flotation froth image of the clean coal in each group of first sample data as input and the measured value of the ash content of the clean coal as a label, training the main model for predicting the ash content of the clean coal, determining the structure and the parameters of the main model for predicting the ash content of the clean coal, and obtaining the trained main model for predicting the ash content of the clean coal.
3. The method of clean coal ash prediction for flotation process of claim 1, wherein the clean coal ash prediction compensation model is trained by:
obtaining a second sample set, each set of second sample data in the second sample set comprising: flotation froth image, tailing image, flotation feed concentration, flotation feed flow, flotation liquid level and clean coal ash measured value;
inputting the flotation froth images in each group of second sample data into the trained clean coal ash content prediction main model to obtain a clean coal ash content prediction value;
respectively carrying out normalization processing on the gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level of the tailing image in each group of second sample data;
and respectively taking the normalized gray characteristic value, the flotation feed concentration, the flotation feed flow and the flotation liquid level corresponding to each group of second sample data as input, taking the difference value between the corresponding measured value of the clean coal ash content and the predicted value of the clean coal ash content as a label, training the clean coal ash content prediction compensation model, determining the structure and the parameters of the clean coal ash content prediction compensation model, and obtaining the trained clean coal ash content prediction compensation model.
4. The method for predicting clean coal ash in a flotation process according to any one of claims 1 to 3, wherein the clean coal ash prediction main model is selected from clean coal ash prediction main models based on a convolutional neural network.
5. The method for predicting clean coal ash in a flotation process according to any one of claims 1 to 3, wherein the clean coal ash prediction compensation model is a prediction compensation model based on a radial basis function neural network.
6. The method for predicting clean coal ash of a flotation process according to any one of claims 1 to 3, wherein the gray level characteristic values include a gray level average, a variance, smoothness and an energy entropy;
the gray characteristic value is obtained by extracting a gray histogram of the tailing image.
7. The method for predicting ash content of clean coal in flotation according to claim 6, wherein the normalization process adopts a maximum and minimum normalization method.
8. The method for predicting clean coal ash content of a flotation process according to any one of claims 1 to 3, wherein the flotation froth image is acquired by a froth image acquisition system fixed above the flotation froth liquid level;
the foam image acquisition system is internally provided with a CCD industrial camera, and LED light sources which form an angle of 45 degrees downwards are fixed on two sides of the industrial camera.
9. The method for predicting the ash content of clean coal in the flotation process according to any one of claims 1 to 3, wherein the tailings image is acquired by a tailings image acquisition system;
a CCD industrial camera is arranged in the tailing image acquisition system;
the tailing image acquisition system is built beside a flotation tailing tank, a real-time flotation tailing sample is extracted through a pump and enters an imaging black box, a light source in the black box is fixed, and a CCD camera is used for acquiring tailing images.
10. The method for predicting ash content of clean coal of a flotation process according to any one of claims 1 to 3,
detecting the concentration of the flotation feed in real time at the flotation feed pipeline by using a concentration meter;
detecting the flotation feeding flow rate at the flotation feeding pipeline in real time by using a flowmeter;
and detecting the flotation liquid level in real time by using a floating ball level meter.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210542985.5A CN114841453A (en) | 2022-05-18 | 2022-05-18 | Clean coal ash content prediction method in flotation process |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210542985.5A CN114841453A (en) | 2022-05-18 | 2022-05-18 | Clean coal ash content prediction method in flotation process |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114841453A true CN114841453A (en) | 2022-08-02 |
Family
ID=82569458
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210542985.5A Pending CN114841453A (en) | 2022-05-18 | 2022-05-18 | Clean coal ash content prediction method in flotation process |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114841453A (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115049165A (en) * | 2022-08-15 | 2022-09-13 | 北矿机电科技有限责任公司 | Flotation concentrate grade prediction method, device and equipment based on deep learning |
| CN115389376A (en) * | 2022-10-28 | 2022-11-25 | 佛山科学技术学院 | Static coal slime flotation image ash content detection method based on chromatography filter paper sampling |
| CN115861235A (en) * | 2022-12-05 | 2023-03-28 | 中国矿业大学 | Flotation tailing ash content prediction method based on multi-feature data fusion |
-
2022
- 2022-05-18 CN CN202210542985.5A patent/CN114841453A/en active Pending
Cited By (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115049165A (en) * | 2022-08-15 | 2022-09-13 | 北矿机电科技有限责任公司 | Flotation concentrate grade prediction method, device and equipment based on deep learning |
| CN115049165B (en) * | 2022-08-15 | 2022-11-22 | 北矿机电科技有限责任公司 | Flotation concentrate grade prediction method, device and equipment based on deep learning |
| CN115389376A (en) * | 2022-10-28 | 2022-11-25 | 佛山科学技术学院 | Static coal slime flotation image ash content detection method based on chromatography filter paper sampling |
| CN115861235A (en) * | 2022-12-05 | 2023-03-28 | 中国矿业大学 | Flotation tailing ash content prediction method based on multi-feature data fusion |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN114841453A (en) | Clean coal ash content prediction method in flotation process | |
| CN110779937B (en) | Casting product internal defect intelligent detection device | |
| Liming et al. | Automated strawberry grading system based on image processing | |
| CN110852395A (en) | Ore granularity detection method and system based on autonomous learning and deep learning | |
| Cao et al. | Integrated prediction model of bauxite concentrate grade based on distributed machine vision | |
| CN101915767B (en) | Method for systematically recognizing egg cracks | |
| CN111881909A (en) | Coal and gangue identification method and device, electronic equipment and storage medium | |
| CN108647722B (en) | Zinc ore grade soft measurement method based on process size characteristics | |
| CN110119747A (en) | A kind of coal rock detection method based on radioscopic image | |
| CN113554621B (en) | Cutter wear state identification system and method based on wavelet scale map and deep migration learning | |
| Zhang et al. | Convolutional memory network-based flotation performance monitoring | |
| Liu et al. | Performance evaluation of a deep learning based wet coal image classification | |
| CN108931621B (en) | Zinc ore grade soft measurement method based on process texture characteristics | |
| CN115587191A (en) | Coal mine vein tendency reasoning system | |
| CN115452824B (en) | Method for multi-model prediction of ash content of coal flotation tailings on dynamic overflow surface | |
| CN109034172B (en) | Product appearance defect detection method based on fuzzy relaxation constraint multi-core learning | |
| CN110728253A (en) | Texture feature measurement method based on particle roughness | |
| CN115390450A (en) | Coal flotation intelligent dosing method capable of being interpreted in segmented mode | |
| Wang et al. | Identification of abnormal conditions for gold flotation process based on multivariate information fusion and double‐channel convolutional neural network | |
| Budzan et al. | Grain size determination and classification using adaptive image segmentation with shape-context information for indirect mill faults detection | |
| CN110738674B (en) | Texture feature measurement method based on particle density | |
| Davletshin et al. | Detection of framboidal pyrite size distributions using convolutional neural networks | |
| CN113870328A (en) | Liquid foreign matter visual detection method and system | |
| Zhong et al. | Short-long temporal graph convolution network for grade monitoring in a first zinc rougher | |
| CN118070015A (en) | Clean coal ash online prediction method based on Bayes super-parameter optimization Light GBM regression |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination |