An integrated coal preparation plant coarseness detection device, method and control system

Abstract

The present invention discloses an integrated coal preparation plant coarseness detection device, method and control system, comprising a buffer barrel, a feed screen basket arranged above the buffer barrel, a video acquisition device, a laser radar, an alarm signal generator, a signal transmitter, an industrial control all-in-one computer and a controller, wherein a weighing force sensor is arranged at the bottom of the feed screen basket; wherein the detection method comprises collecting data from the laser radar, the weighing force sensor and the video acquisition device, establishing and analyzing a plurality of models, obtaining the relationship between the input and the particle size distribution of the material and the calculation result of the coarseness grade, and comprehensively obtaining the coarseness analysis result; the present invention realizes that a single sampling can make the sample be detected online by three different detection means at the same time, which not only realizes the real-time monitoring and rating of the coarseness condition, but also can accurately detect the particle size composition distribution of the coarseness particles, traces the cause of the coarseness accident, and solves the problem of low accuracy of "coarseness" detection in the prior art.

Description

Technical Field

[0001] The present invention relates to the technical field of intelligent online detection in coal preparation plants, and in particular to an integrated coal preparation plant coarseness detection device, method and control system. Background Art

[0002] According to the production process of the coal preparation plant, filter screen systems with different functions are often set up, including desludging screens, de-mediating screens and dewatering screens. When the screen plate wears or falls off during system operation, the material on the screen enters the lower part of the screen, which will cause the downstream bucket grate to be blocked or even collapse under pressure. At this time, the material that should have remained on the screen but mistakenly enters the lower part of the screen is called "coarse" particles; the "coarse" particles in the coal slime water system will seriously affect the subsequent coarse coal slime sorting, fine coal slime flotation and concentrated filter press system. The operating conditions, in severe cases, may even cause a production accident that requires the coal preparation plant to shut down.

[0003] In the existing technology, there are generally two methods to achieve "coarse particles" monitoring in coal preparation plants. One is an indirect detection method that uses sensors such as photoelectric switches and water immersion detection sensors to detect the presence of overflow water; the other is to use cameras installed in the space above various barrels to directly detect coarse particles. Method 1 has been promoted and applied in the industry in the past two years due to its low technical threshold and low price, but the false alarm rate is high, and because the screen plate is very easy to clog, it requires frequent manual maintenance, and the use effect is poor; Method 2 uses images to directly detect the area of ​​coarse particles, and the detection accuracy is higher than Method 1, but the coal slurry bucket, mixed medium bucket, dilute medium bucket and other types of barrels in the coal preparation plant are usually installed on the first floor of the coal preparation plant. The space above the barrel is narrow, the water vapor is heavy, the installation environment is poor, and the detection results are easily affected by environmental changes. Therefore, Method 2 has only been piloted in individual coal preparation plants. Summary of the invention

[0004] The purpose of the present invention is to overcome the deficiencies of the prior art and to provide an integrated coal preparation plant coarse coal detection device, method and control system to solve the above problems.

[0005] The objective of the present invention is achieved through the following technical solutions:

[0006] An integrated coal preparation plant coarseness detection device comprises a buffer barrel, a feed screen basket placed above the buffer barrel, a video acquisition device, a laser radar, an alarm signal generator, a signal transmitter, an industrial control integrated machine and a controller; wherein:

[0007] A plurality of weighing force sensors arranged in a balanced manner are provided at the bottom of the feed screen basket, and each weighing force sensor is electrically connected to the signal transmitter. The video acquisition device and the laser radar are both arranged above the feed screen basket. The video acquisition device, the laser radar and the signal transmitter upload the collected data to the industrial control integrated machine through a communication device.

[0008] A feed pipe and a water supply pipe are provided above the feed screen basket, and a feed valve and a flushing valve are provided on the feed pipe and the water supply pipe respectively. The bottom of the feed screen basket is opened and closed or flipped by an actuator, and a discharge pipe and a discharge valve arranged on the discharge pipe are provided below the buffer barrel. The feed valve, flushing valve, discharge valve, actuator, alarm signal generator and industrial control integrated machine are all electrically connected to the controller.

[0009] In the above invention, further, at least four weighing force sensors are provided; when there are four weighing force sensors, the four weighing force sensors are symmetrically arranged at the four corners at the bottom of the feeding screen basket, and the signal transmitter converts the voltage output signal of the weighing force sensor into an RS485 / 232 output signal.

[0010] In the above invention content, further, the video acquisition device includes an industrial camera and an industrial light source, the communication device is a POE switch, the industrial camera directly draws power from the POE switch, and the industrial camera uploads the acquired video to the industrial control all-in-one machine through the communication device. The controller is a programmable logic controller, and the industrial control all-in-one machine and the programmable logic controller are connected via Ethernet, and use TCP / IP protocol for data transmission.

[0011] In the above invention content, further, the feed valve is an electric regulating ball valve, the discharge valve is an electric switching gate valve, and the flush valve is an electric switching butterfly valve or an electric switching ball valve; one side of the bottom sieve plate of the feed screen basket is hinged to the side sieve plate, and the actuator includes an electro-hydraulic push rod, the fixed end of the electro-hydraulic push rod is connected to the side wall of the detection device, and the telescopic end of the electro-hydraulic push rod is hinged to the movable end of the bottom sieve plate of the feed screen basket, and the opening and closing of the feed screen basket is realized by the electro-hydraulic push rod.

[0012] The integrated coal preparation plant roughing detection method includes the following steps:

[0013] S1: collecting a material image on the feed screen basket, and initializing the material image to obtain an initial image of the feed to be analyzed;

[0014] S2, performing image processing on the initial image of the input material to be analyzed to obtain an original image, and performing feature extraction on the original image to obtain image features, wherein the image features include one or more of image color, size and texture;

[0015] S3, acquiring a plurality of ranging data based on a laser radar, forming the plurality of ranging data and obtaining three-dimensional coordinate position feature data based on the ranging data group;

[0016] S4, acquiring a plurality of force measurement data based on the weighing force measurement sensor, forming a force measurement data group from the plurality of force measurement data, and obtaining weight measurement data based on the force measurement data group;

[0017] S5. A calculation model is established based on the image features, the three-dimensional coordinate position feature data and the weight measurement data as inputs, wherein the calculation model is used to obtain the relationship between the input and the material particle size distribution and the coarse and coarse "grade" calculation results, and comprehensively obtain the "coarse and coarse" analysis results and the material particle size distribution composition.

[0018] In the above invention, further, S5 includes:

[0019] S51: using the weight measurement data group as a first input module, using the three-dimensional coordinate position feature data as a second input module, and using the image feature as a third input module;

[0020] S52: establishing a first algorithm model, wherein the first algorithm model is used to convert the data of the first input module into a percentage of the number of coarse particles;

[0021] S53: establishing a second algorithm model, wherein the second algorithm model is used to convert the data of the second input module into a particle size composition distribution of coarse particles and a volume percentage of coarse particles;

[0022] S54: establishing a third algorithm model, wherein the third algorithm model is used to convert the data of the third input module into a fuzzy particle size composition distribution and a coarse particle area ratio;

[0023] S55: Establishing a fourth algorithm model, wherein the fourth algorithm model obtains the “coarseness” level of the coarse particles based on the proportion of the number of coarse particles, the proportion of the area of ​​the coarse particles and the proportion of the volume of the coarse particles.

[0024] In the above invention, further, the calculation formula of the first algorithm model is as shown in formula (1): (1) In the formula, R C is the proportion of coarse particles. is the weight measurement data obtained according to the weight measurement data group, W T The preset threshold for the weight measurement data.

[0025] In the above invention, further, S53 includes:

[0026] S531: Obtaining coarse-grained data points containing three-dimensional coordinate position features based on laser radar x ;

[0027] S532: Establish a mixed Gaussian model containing K components, wherein the K components correspond to the K detected coarse particles, and each of the K components i It is represented by a three-dimensional Gaussian distribution whose mean vector μ i and the covariance matrix Σ i Definition, each component has a weight π i ;

[0028] S533: Based on the mixed Gaussian model, the coarse-grained data points are analyzed. x Perform data clustering, the mixed Gaussian model is used to cluster the coarse-grained data points x The probability density function of is shown in formula (2): (2) In the formula, N ( x ∣ μ i , Σ i ) indicates the i The density function of a Gaussian distribution is θ is the set of all parameters of the model, weights π i Indicates i The volume ratio of the component in the mixture model, the sum of all weights is 1; in three-dimensional space, the mean vector μ i and the covariance matrix Σ i The relationship is shown in formula (3): (3) In the formula, |Σ i | represents the matrix Σ i The determinant of ; exp represents the natural exponential function;

[0029] S534: Estimating parameters based on the maximum likelihood function algorithm θ , to cluster the detected independent coarse-grained data points, and thus obtain the number of coarse-grained particles k And the radius R mmi And the volume proportion of coarse particles R is obtained by formula (4): dfv , (4) In the formula, k To calculate the number of coarse particles, R mmi is the radius of the coarse particles actually detected, R tv is the preset threshold value of the coarse particle radius; then, by comparing R mmi and R tvDetermine whether the "roughness" phenomenon occurs.

[0030] In the above invention, further, S54 includes:

[0031] Based on the monitoring image collected by the industrial camera, a binary image is obtained by image preprocessing, wherein in the binary image, the granular coarse material appears as a connected white area, the third algorithm model is established based on the depth-first algorithm, and the connected white area is identified by a connected domain analysis and recognition method to extract the particle size information; wherein, the connected white area is identified by a connected domain analysis and recognition method to extract the particle size information includes:

[0032] Visit each connected region on the binary image by recursively or by stack traversing the nodes of the graph, and use the Boolean array visited to record the visit status of all nodes, where visited[i] is true, indicating that node i has been visited, and node represents the current node;

[0033] Based on the recursive function dfs(node), all unvisited neighbors of the node node are traversed. For each unvisited neighbor, dfs(neighbor) is used. First, dfs(startNode) is called from any unvisited node in the image. The dfs(startNode) indicates the start of traversing the first connected component. After traversing a connected component, the next unvisited node is found, and the recursive function is called again, and so on, until all nodes are visited.

[0034] The pixel area covered by the reaction particles in each detected connected domain is calculated based on the number of white pixels in the connected domain, and the equivalent circle radius of the coarse particles in the connected domain is further calculated. The calculation formula of the equivalent circle radius of the coarse particles is shown in formula (5): (5) In the formula, R n i is the equivalent circle radius of the coarse particles, A i is the area of ​​each monitored connected domain;

[0035] The pixel area A covered by the reactive particles in all N detected connected domains i The area ratio of coarse particles R dfs , R dfs The calculation formula is shown in formula (6): (6) In the formula, N is the number of coarse particles. R ts is the preset threshold of the equivalent circle radius of the coarse particles, and the equivalent circle radius R of the coarse particles is compared. n i and Rts Determine whether the "roughness" phenomenon occurs.

[0036] In the above invention, further, in S55, the calculation formula of the roughness level R is as shown in formula (7): (7) In the formula, α, β and below and Gamma is a normalization parameter, and α、β as well as Gamma The sum is 1.

[0037] An integrated coal preparation plant coarse particle detection control system comprises an industrial control integrated machine, a programmable controller, an alarm signal generator, a feed valve, a flushing valve, a discharge valve, an actuator, a display module, an algorithm module, a relational database and a non-relational database, wherein the feed valve, flushing valve, discharge valve, actuator and alarm signal generator are all connected to the programmable controller, the programmable controller, the algorithm module and the display module are connected to the industrial control integrated machine, the display module is used to display a content change curve and a particle size distribution table of "coarse particles"; the relational database and the non-relational database are both connected to the industrial control integrated machine, the relational database is used to store preset key parameters required for executing the coarse particle detection method and the control method after different levels of "alarm", the non-relational database is used to store the measurement data of the calculation process of the coarse particle detection method and the execution process of the control method after different levels of "alarm", the feed screen basket state and the valve opening, and the industrial control integrated machine uses the collected data of the relational database and the non-relational database to execute the coarse particle detection method through the algorithm module.

[0038] The beneficial effects of the present invention are:

[0039] The present invention realizes one-time sampling in the production process through a video acquisition device, a laser radar and a weighing force sensor, so that the sample can be detected online by three different detection means at the same time. It can not only monitor and rate the "coarseness" condition, but also accurately detect the particle size composition of the "coarseness" particles. According to the particle size composition and the process configuration of the filtration system, the cause of the "coarseness" can be further analyzed in depth, thus solving the problems of low accuracy in the detection of "coarseness" of coal slime water, high system false alarm rate, high failure rate, inability to trace the cause of the "coarseness" accident and large maintenance workload in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Picture 1 It is a system diagram of the integrated roughness detection device of the present invention;

[0041] Picture 2 It is a schematic diagram of the arrangement of the weighing force sensor in the present invention;

[0042] Picture 3 This is a flow chart of the integrated rough detection method of the present invention;

[0043] Picture 4 This is a flow chart of step S5 in the integrated rough detection method of the present invention;

[0044] Picture 5 It is a schematic diagram of the integrated rough running detection control system of the present invention.

[0045] In the figure, 1-feed pipe, 2-feed valve, 3-feed screen basket, 4-actuator, 5-weighing force sensor, 6-signal transmitter, 7-laser radar, 8-flush valve, 9-buffer barrel, 10-discharge pipe, 11-discharge valve, 12-industrial light source, 13-industrial camera, 14-communication device, 15-programmable logic controller, 16-industrial control all-in-one machine, 17-alarm signal generator, 18-water supply pipe. DETAILED DESCRIPTION

[0046] The following describes the embodiments of the present invention by specific examples, and those skilled in the art can easily understand other advantages and effects of the present invention from the contents disclosed in this specification. The present invention can also be implemented or applied through other different specific embodiments, and the details in this specification can also be modified or changed in various ways based on different viewpoints and applications without departing from the spirit of the present invention. It should be noted that the following embodiments and features in the embodiments can be combined with each other without conflict.

[0047] Embodiment 1:

[0048] Please see attached Picture 1 As shown, an integrated coal preparation plant coarse detection device includes a buffer barrel 9, a feed screen basket 3 placed above the buffer barrel 9, a video acquisition device, a laser radar 7, an alarm signal generator 17, a signal transmitter 6, an industrial control integrated machine 16 and a controller, wherein:

[0049] The controller is a programmable logic controller 15, and a plurality of balanced weighing force sensors 5 are provided at the bottom of the feeding screen basket 3, and each weighing force sensor 5 is electrically connected to the signal transmitter 6, and the signal transmitter 6 converts the mv voltage output signal of the weighing force sensor 5 into an RS485 / 232 output signal;

[0050] The video acquisition device and laser radar 7 are both arranged above the feed screen basket 3. The video acquisition device, laser radar 7 and signal transmitter 6 upload the collected data to the industrial control integrated machine 16 through the communication device 14. A feed pipe 1 and a water supply pipe 18 are also arranged above the feed screen basket 3. The feed pipe 1 and the water supply pipe 18 are respectively provided with a feed valve 2 and a flushing valve 8. The bottom of the feed screen basket 3 can be opened, closed or flipped by the actuator 4; wherein, the screen gap of the feed screen basket 3 is 5mm, which can be adapted.

[0051] A discharge pipe 10 and a discharge valve 11 arranged on the discharge pipe 10 are provided below the buffer barrel 9. The feed valve 2, flush valve 8, discharge valve 11, actuator 4, alarm signal generator 17 and industrial control integrated machine 16 are all electrically connected to the controller, wherein the diameter of the feed pipe 1 is DN100mm, and the diameter of the discharge pipe 10 is DN200mm.

[0052] Please see attached Picture 2 As shown, at least four weighing force sensors 5 are provided. When there are four weighing force sensors 5, the four weighing force sensors 5 are symmetrically arranged at the bottom of the feed screen basket 3 at four corners, so that the weight of the feed screen basket 3 is evenly distributed on the four weighing force sensors 5, ensuring the accuracy of the detection data.

[0053] The video acquisition device includes an industrial camera 13 and an industrial light source 12. The communication device 14 is a POE switch. The industrial camera 13 directly draws power from the POE switch. The industrial camera 13 uploads the acquired video to the industrial control all-in-one machine 16 through the communication device 14. The industrial control all-in-one machine 16 is connected to the programmable logic controller 15 via Ethernet and uses the TCP / IP protocol for data transmission.

[0054] In the above embodiment, preferably, the feed valve 2 is an electric regulating ball valve, the discharge valve 11 is an electric switching gate valve, and the flush valve 8 is an electric switching butterfly valve or an electric switching ball valve; one side of the bottom sieve plate of the feed screen basket 3 is hinged to the side sieve plate, and the actuator includes an electro-hydraulic push rod (not shown in the accompanying drawings), the fixed end of the electro-hydraulic push rod is connected to the side wall of the detection device, the telescopic end of the electro-hydraulic push rod is hinged to the movable end of the bottom sieve plate of the feed screen basket 3, and the telescopic end of the electro-hydraulic push rod is provided with a movable hole in the extension direction and is movably connected to the bottom sieve plate of the feed screen basket 3 through the movable hole, and the bottom sieve basket can be opened and closed by the electro-hydraulic push rod to adapt to the movement stroke of the bottom sieve plate of the feed screen basket 3.

[0055] This detection device has remote and local operation functions and provides a standard Modbus TCP signal interface.

[0056] Embodiment 2:

[0057] Please see attached Picture 3 and attached Picture 4 As shown, a method for detecting roughness in an integrated coal preparation plant comprises the following steps:

[0058] S1: collecting a material image on the feed screen basket 3, and performing initialization processing on the material image to obtain an initial image of the feed to be analyzed, wherein the collected image is cropped to remove useless areas to obtain a first preprocessed image;

[0059] S2, performing image processing on the initial image of the input material to be analyzed to obtain an original image, and performing feature extraction on the original image to obtain image features, wherein the image features include one or more of image color, size and texture;

[0060] S3, acquiring multiple distance measurement data based on the laser radar 7, forming the multiple distance measurement data and obtaining three-dimensional coordinate position feature data based on the distance measurement data group; for subsequent input;

[0061] S4, obtaining a plurality of force measurement data groups based on the weighing force measurement sensors 5 at the four corners, forming a force measurement data group from the plurality of force measurement data and obtaining weight measurement data based on the force measurement data group;

[0062] S5. A calculation model is established based on the image features, the three-dimensional coordinate position feature data and the weight measurement data as inputs, wherein the calculation model is used to obtain the relationship between the input and the material particle size distribution and the coarse and coarse "grade" calculation results, and comprehensively obtain the "coarse and coarse" analysis results and the material particle size distribution composition.

[0063] In the above embodiment, preferably, S5 includes:

[0064] S51: using the above-measured weight data group as the first input module, using the three-dimensional coordinate position feature data as the second input module, and using the extracted image color, area, and texture features as the third input module;

[0065] S52: establishing a first algorithm model, wherein the first algorithm model is used to convert the data of the first input module into a percentage of the number of coarse particles;

[0066] S53: establishing a second algorithm model, wherein the second algorithm model is used to convert the data of the second input module into a particle size composition distribution of coarse particles and a volume percentage of coarse particles;

[0067] S54: establishing a third algorithm model, wherein the third algorithm model is used to convert the data of the third input module into a fuzzy particle size composition distribution and a coarse particle area ratio;

[0068] S55: Establishing a fourth algorithm model, the fourth algorithm model obtains the “coarseness” level of the coarse particles through calculation based on the input of the above-obtained percentage of the number of coarse particles, the percentage of the area of ​​coarse particles and the percentage of the volume of coarse particles.

[0069] In step S52, specifically, the calculation formula of the first algorithm model is as shown in formula (1): (1) In the formula, R C is the proportion of coarse particles. is the weight measurement data obtained according to the weight measurement data group,W T The preset threshold value of the weight measurement data is then Preset thresholds for weight measurement data W T For comparison, if , it is considered that the "running rough" phenomenon has occurred and an alarm is triggered when the running roughness occurs; otherwise, the "running rough" phenomenon has not occurred.

[0070] In step S53, it specifically includes:

[0071] S531: Obtaining coarse-grained data points containing three-dimensional coordinate position features based on laser radar 7 x , it needs to be assigned to a certain coarse particle, and the mixed Gaussian model can be used to cluster the data;

[0072] S532: Establish a mixed Gaussian model containing K components, wherein the K components correspond to the K detected coarse particles, and each of the K components i (i=1,2,...,K) is represented by a three-dimensional Gaussian distribution whose mean vector μ i and the covariance matrix Σ i Definition, each component has a weight π i ;

[0073] S533: Based on the mixed Gaussian model, the coarse-grained data points are analyzed. x Data clustering is performed. For a data point detected by LiDAR 7, the probability that it belongs to the i-th component can be calculated by the density function of the Gaussian distribution, and the weighted sum of all components is considered to express the data point x The overall likelihood of the mixed Gaussian model for the coarse-grained data points x The probability density function of is shown in formula (2): (2) In the formula, N ( x ∣ μi , Σ i ) indicates the i The density function of a Gaussian distribution is θ is the set of all parameters of the model, weights π i Indicates i The volume ratio of the component in the mixture model, the sum of all weights is 1; in three-dimensional space, the mean vector μ i and the covariance matrix Σ i The relationship is shown in formula (3): (3) In the formula, |Σi | represents the matrix Σ i The determinant of ; exp represents the natural exponential function;

[0074] S534: Estimating parameters based on the maximum likelihood function algorithm θ , the clustering of coarse-grained data points can be realized, thereby obtaining the number of coarse-grained data points. k And the radius R mmi And the volume proportion of coarse particles R is obtained by formula (4): dfv , (4) In the formula, k To calculate the number of coarse particles, R mmi is the radius of the coarse particles, R tv is the preset threshold; then, by comparing R mmi and R tv Determine whether the "running rough" phenomenon occurs. R mmi ≥R tv When the error is detected, it is considered that a "roughing" phenomenon has occurred and an alarm is triggered.

[0075] Preferably, in step S54, based on the monitoring image collected by the industrial camera 13, a clear binary image is obtained by image preprocessing, wherein in the binary image, the granular coarse material appears as a connected white area, and the third algorithm model is established based on the depth-first algorithm using the depth-first algorithm, and the particle size information is extracted by connecting domain analysis based on the depth-first algorithm. The connecting domain analysis and identification method is used to identify the connected white area to extract the particle size information, and the connecting domain analysis and identification method is used to identify the connected white area to extract the particle size information, including:

[0076] Traverse the nodes of the graph through recursion or stack to visit each connected area on the binary image, and use the Boolean array visited to record the visit status of all nodes, where visited[i] is true, indicating that node i has been visited, and node represents the current node;

[0077] Based on the recursive function dfs(node), all unvisited neighbors of the node node are traversed. For each unvisited neighbor, dfs(neighbor) is used. The recursive function dfs(node) accepts a node as a parameter. First, dfs(startNode) is called from any unvisited node in the image. The dfs(startNode) indicates that the first connected component is started to be traversed. After traversing a connected component, the next unvisited node is found, and the recursive function is called again, and so on, until all nodes are visited.

[0078] The pixel area covered by the reaction particles in each detected connected domain is calculated based on the number of white pixels in the connected domain, and the equivalent circle radius of the coarse particles in the connected domain is further calculated. The calculation formula of the equivalent circle radius of the coarse particles is shown in formula (5): (5) In the formula, R n i is the equivalent circle radius of the coarse particles, A i The area of ​​each monitored connected domain;

[0079] The pixel area A covered by the reactive particles in all N detected connected domains i The area ratio of coarse particles R dfs , R dfs The calculation formula is shown in formula (6): (6) In the formula, N is the number of coarse particles. R ts is the preset threshold of the equivalent circle radius of the coarse particles, and the equivalent circle radius R of the coarse particles is compared. n i and R ts Determine whether the "running rough" phenomenon occurs. When R n i ≥R ts When the error is 0, it is considered that the "roughing" phenomenon occurs and an alarm is triggered.

[0080] In step S55, the calculation formula of the roughness level R is shown in formula (7): (7) In the formula, α、β as well as Gamma is a normalization parameter, and α、β as well as Gamma The sum is 1, and the specific value can be adjusted according to actual conditions.

[0081] Through the above steps, not only can the monitoring and rating of the "coarseness" condition be achieved, but the particle size composition of the "coarseness" particles can also be accurately detected. Based on the particle size composition and the process configuration of the filtration system, the cause of the "coarseness" can be further analyzed in depth.

[0082] Embodiment 3:

[0083] An integrated coal preparation plant coarseness detection and control system, please refer to the attached Picture 5As shown, it includes an industrial control integrated machine 16, a programmable controller, an alarm signal generator 17, a feed valve 2, a flushing valve, a discharge valve 11, an actuator 4, a display module, an algorithm module, and a relational database and a non-relational database. The feed valve 2, the flushing valve, the discharge valve 11, the actuator 4 and the alarm signal generator 17 are all connected to the programmable controller, the programmable controller, the algorithm module and the display module are connected to the industrial control integrated machine 16, and the display module is used to display the content change curve and particle size distribution table of "coarse" particles;

[0084] The relational database and the non-relational database are both connected to the industrial control integrated machine 16, and the relational database is used to store the preset key parameters required for executing the rough detection method and the control method after different levels of alarm;

[0085] The non-relational database is used to store the measurement data of the calculation process of the roughing detection method and the execution process of the control method after different levels of alarm, the state of the feeding screen basket 3 and the valve opening, and the industrial control integrated machine 16 uses the collected data of the relational database and the non-relational database to execute the roughing detection method through the algorithm module. Please refer to the following working process for the specific implementation method.

[0086] Specific working process:

[0087] The feed valve 2, the actuator 4, the flush valve 8, and the discharge valve 11 are all connected to the PLC system through wired connections; at the same time, the PLC system is connected to the industrial control integrated machine 16 through Ethernet, and the weighing force sensor 5, the laser radar 7, and the industrial camera 13 are all connected to the POE switch through wired connections, and the collected data are uploaded to the industrial control integrated machine 16;

[0088] According to the equipment start-up signal, the feed valve 2 is automatically opened to bypass the material, and the screen water or centrifugal liquid of the coal preparation plant production system enters the feed screen basket 3 through the feed pipe 1. During normal production, all fine particles and slurry pass through the feed screen basket 3 and enter the buffer barrel 9 below. A discharge pipe 10 is provided at the bottom of the buffer, and a discharge valve 11 is provided on the discharge pipe 10. According to the equipment start-up signal, the discharge valve 11 is automatically opened;

[0089] When the feed material is "coarse", the coarse particles are retained on the horizontal screen plate at the bottom of the feed screen basket 3. The laser radar 7 and the industrial camera 13 above the feed screen basket 3 both illuminate the horizontal screen plate at the bottom of the feed screen basket 3. The collected three-dimensional coordinate position feature data and images are transmitted back to the industrial control integrated machine 16 by the POE switch. The force measurement data group collected by the four weighing force sensors 5 at the bottom of the feed screen basket 3 is also transmitted back to the industrial control integrated machine 16 by the POE switch.

[0090] Then, the industrial control integrated machine 16 calculates the particle size composition distribution and the “coarseness” grade of the coarse particles based on the image features, the three-dimensional coordinate position features and the force measurement data set according to the preset algorithm and the set threshold in the above method;

[0091] When the R of the coarse particles is greater than a preset value a, a "coarse particles" alarm is generated, the feed valve 2 is automatically closed, and the bottom sieve plate of the feed screen basket 3 is automatically opened through the linear motion of the electro-hydraulic push rod (to discharge the coarse particles on the screen basket). After 10S (configurable time), the screen basket is automatically closed and the feed valve 2 is automatically opened to continue the material collection and detection;

[0092] When the R of the coarse particles is greater than a preset value b, a "serious coarse particles" alarm is generated, the feed valve 2 is automatically closed, and the bottom screen plate of the feed screen basket 3 is automatically opened through the linear motion of the electro-hydraulic push rod to ensure that the feed is discharged smoothly from the device;

[0093] After all the slurry in the buffer barrel 9 is discharged, the water supply valve opens automatically to flush the feed screen basket 3 and the buffer barrel 9 for 30 seconds (the time is configurable) to ensure that there are no residual coarse particles inside the device;

[0094] After the alarm is automatically or manually eliminated, the system automatically resets to the initial state, automatically opens the feed valve 2, takes the material into the device, and enters the next round of detection process until the equipment shutdown signal is received;

[0095] When receiving the equipment shutdown signal, first, the feed valve 2 is automatically closed, the bottom sieve plate of the feed screen basket 3 is automatically opened by the linear motion of the electro-hydraulic push rod, and the discharge valve 11 remains fully open. It runs in this state for 1 minute (configurable); secondly, the flush valve 8 is automatically opened for 30 seconds (configurable time) to discharge the coarse particles in the device; after that, the shutdown process is automatically executed to restore the system to the initial state and keep the equipment powered.

[0096] Through the integrated coarseness detection method and detection device that combines the three principles of machine vision, lidar and static pressure measurement, it is possible to take a sample once in the production process of the coal preparation plant and simultaneously detect the sample online by three different detection methods. It not only realizes the real-time monitoring and rating of the "coarseness" condition, but also can accurately detect the particle size composition distribution of the "coarseness" particles, and can further trace the cause of the coarseness accident, effectively solving the problems of low "coarseness" detection accuracy, high system false alarm rate, high failure rate, inability to trace the cause of the accident and heavy maintenance workload in the existing technology.

[0097] In the description of the present invention, it is necessary to understand that the directions or positional relationships indicated by the terms "up", "down", "left", "right", etc. are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention.

[0098] The above-mentioned embodiments only express the specific implementation of the present invention, and the description thereof is relatively specific and detailed, but it cannot be understood as limiting the scope of the present invention. It should be pointed out that, for ordinary technicians in this field, several variations and improvements can be made without departing from the concept of the present invention, which all belong to the protection scope of the present invention.

Claims

1. An integrated coal preparation plant coarseness detection device, characterized in that: The device comprises a buffer barrel (9), a feed screen basket (3) placed above the buffer barrel (9), a video acquisition device, a laser radar (7), an alarm signal generator (17), a signal transmitter (6), an industrial control integrated machine (16), and a controller; wherein: A plurality of evenly arranged weighing force sensors (5) are provided at the bottom of the feed screen basket (3), and each weighing force sensor (5) is electrically connected to the signal transmitter (6); the video acquisition device and the laser radar (7) are both arranged above the feed screen basket (3); and the video acquisition device, the laser radar (7) and the signal transmitter (6) upload the collected data to the industrial control integrated machine (16) via a communication device (14); A feed pipe (1) and a water supply pipe (18) are provided above the feed screen basket (3); a feed valve (2) and a flushing valve (8) are provided on the feed pipe (1) and the water supply pipe (18), respectively; the bottom of the feed screen basket (3) is opened, closed or turned over by an actuator (4); a discharge pipe (10) and a discharge valve (11) provided on the discharge pipe (10) are provided below the buffer barrel (9); the feed valve (2), the flushing valve (8), the discharge valve (11), the actuator (4), the alarm signal generator (17) and the industrial control integrated machine (16) are all electrically connected to the controller; The feed valve (2), the actuator (4), the flush valve (8), and the discharge valve (11) are all connected to the PLC system via wired connections; at the same time, the PLC system is connected to the industrial control integrated machine (16) via Ethernet, and the weighing force sensor (5), the laser radar (7), and the video acquisition device are all connected to the communication device (14) via wired connections, and the collected data is uploaded to the industrial control integrated machine (16); When the feed material is "coarse", the coarse particles are retained on the horizontal screen plate at the bottom of the feed screen basket (3), the laser radar (7) and the video acquisition device above the feed screen basket (3) both illuminate the horizontal screen plate at the bottom of the feed screen basket (3), and the collected three-dimensional coordinate position feature data and images are transmitted back to the industrial control integrated machine (16) by the communication device (14), and the force measurement data group collected by the weighing force sensor (5) at the bottom of the feed screen basket (3) is also transmitted back to the industrial control integrated machine (16) by the communication device (14); Then, the industrial control integrated machine (16) calculates the particle size composition distribution and "coarseness" grade of the coarse particles based on the image features, the three-dimensional coordinate position features and the force measurement data set.

2. The integrated coal preparation plant roughness detection device according to claim 1 is characterized in that: At least four weighing force sensors (5) are provided; when there are four weighing force sensors (5), the four weighing force sensors (5) are symmetrically arranged at four corners at the bottom of the feed screen basket (3), and the signal transmitter (6) converts the voltage output signal of the weighing force sensor (5) into an RS485 / 232 output signal.

3. The integrated coal preparation plant coarseness detection device according to claim 1 is characterized in that: The video acquisition device comprises an industrial camera (13) and an industrial light source (12); the communication device (14) is a POE switch; the industrial camera (13) directly draws power from the POE switch; the industrial camera (13) uploads the acquired video to an industrial control all-in-one machine (16) via the communication device (14); the controller is a programmable logic controller (15); the industrial control all-in-one machine (16) and the programmable logic controller (15) are connected via Ethernet and use a TCP / IP protocol for data transmission.

4. The integrated coal preparation plant coarseness detection device according to claim 1 is characterized in that: The feed valve (2) is an electric regulating ball valve, the discharge valve (11) is an electric switch gate valve, and the flush valve (8) is an electric switch butterfly valve or an electric switch ball valve; one side of the bottom sieve plate of the feed screen basket (3) is hinged to the side sieve plate, and the actuator includes an electro-hydraulic push rod, the fixed end of the electro-hydraulic push rod is connected to the side wall of the detection device, and the telescopic end of the electro-hydraulic push rod is hinged to the movable end of the bottom sieve plate of the feed screen basket (3), and the feed screen basket (3) is opened and closed by the electro-hydraulic push rod.

5. A method for detecting roughness in an integrated coal preparation plant according to any one of claims 1 to 4, characterized in that: The following steps are involved: S1: collecting a material image on the feed screen basket (3), and performing initialization processing on the material image to obtain an initial image of the feed to be analyzed; S2, performing image processing on the initial image of the input material to be analyzed to obtain an original image, and performing feature extraction on the original image to obtain image features, wherein the image features include one or more of image color, size and texture; S3, acquiring a plurality of distance measurement data based on the laser radar (7), forming the plurality of distance measurement data and obtaining three-dimensional coordinate position feature data based on the distance measurement data group; S4, acquiring a plurality of force measurement data based on the weighing force measurement sensor (5), forming a force measurement data group from the plurality of force measurement data, and obtaining weight measurement data based on the force measurement data group; S5. A calculation model is established based on the image features, the three-dimensional coordinate position feature data and the weight measurement data as inputs, wherein the calculation model is used to obtain the relationship between the input and the material particle size distribution and the coarse and coarse "grade" calculation results, and comprehensively obtain the "coarse and coarse" analysis results and the material particle size distribution composition.

6. The integrated coal preparation plant roughness detection method according to claim 5, characterized in that S5 include: S51: using the weight measurement data group as a first input module, using the three-dimensional coordinate position feature data as a second input module, and using the image feature as a third input module; S52: establishing a first algorithm model, wherein the first algorithm model is used to convert the data of the first input module into a percentage of the number of coarse particles; S53: establishing a second algorithm model, wherein the second algorithm model is used to convert the data of the second input module into a particle size composition distribution of coarse particles and a volume percentage of coarse particles; S54: establishing a third algorithm model, wherein the third algorithm model is used to convert the data of the third input module into a fuzzy particle size composition distribution and a coarse particle area ratio; S55: Establishing a fourth algorithm model, wherein the fourth algorithm model obtains the "coarseness" level of the coarse particles based on the proportion of the number of coarse particles, the proportion of the area of ​​the coarse particles and the proportion of the volume of the coarse particles.

7. The integrated coal preparation plant roughness detection method according to claim 6, characterized in that: The calculation formula of the first algorithm model is shown in formula (1): In the formula, is the proportion of coarse particles. is the weight measurement data obtained according to the weight measurement data group, The preset threshold for the weight measurement data.

8. The integrated coal preparation plant roughness detection method according to claim 6, characterized in that: S53 includes: S531: Obtaining coarse-grained data points containing three-dimensional coordinate position features based on laser radar (7) ; S532: Establishment of The Gaussian mixture model of components, where The components detected A coarse particle, Each of the components It is represented by a three-dimensional Gaussian distribution whose mean vector and the covariance matrix Definition, each component has a weight ; S533: Based on the mixed Gaussian model, the coarse-grained data points are analyzed. Perform data clustering, the mixed Gaussian model is used to cluster the coarse-grained data points The probability density function of is shown in formula (2): In the formula, Indicates The density function of a Gaussian distribution is is the set of all parameters of the model, weights Indicates The volume ratio of the component in the mixture model, the sum of all weights is 1; in three-dimensional space, the mean vector and the covariance matrix The relationship is shown in formula (3): In the formula, Representation Matrix The determinant of ; exp represents the natural exponential function; S534: Estimating parameters based on the maximum likelihood function algorithm , to cluster the detected independent coarse-grained data points, and thus obtain the number of coarse-grained particles And the radius And the volume proportion of coarse particles is obtained by formula (4): , In the formula, To calculate the number of coarse particles, is the radius of the coarse particles actually detected, is the preset threshold value of the coarse particle radius; then, by comparing and Determine whether the "roughness" phenomenon occurs.

9. The integrated coal preparation plant roughness detection method according to claim 6, characterized in that: S54 includes: Based on the monitoring image captured by the industrial camera (13), a binary image is obtained by image preprocessing, wherein in the binary image, the granular coarse material appears as a connected white area, the third algorithm model is established based on the depth-first algorithm, and the connected white area is identified by a connected domain analysis and recognition method to extract particle size information; wherein the connected white area is identified by a connected domain analysis and recognition method to extract particle size information includes: Visit each connected region on the binary image by recursively or by stack traversing the nodes of the graph, and use the Boolean array visited to record the visited status of all nodes, where visited[i] is true, indicating that node i has been visited, and node represents the current node; Based on the recursive function dfs(node), all unvisited neighbors of the node node are traversed, and each unvisited neighbor is represented by dfs(neighbor). First, dfs(startNode) is called from any unvisited node in the graph. The dfs(startNode) indicates the start of traversing the first connected component. After traversing a connected component, the next unvisited node is found, and the recursive function is called again, and so on, until all nodes are visited. The pixel area covered by the reaction particles in each detected connected domain is calculated based on the number of white pixels in the connected domain, and the equivalent circle radius of the coarse particles in the connected domain is further calculated. The calculation formula of the equivalent circle radius of the coarse particles is shown in formula (5): In the formula, is the equivalent circle radius of the coarse particles, is the area of ​​each monitored connected domain; The pixel area covered by the reactive particles in all N detected connected domains Statistical analysis of particle size characteristics was performed to obtain the area ratio of coarse particles. , The calculation formula is shown in formula (6): Where N is the number of coarse particles. is a preset threshold value of the equivalent circle radius of the coarse particles, and the equivalent circle radius of the coarse particles is compared. and Determine whether the "roughness" phenomenon occurs.

10. The integrated coal preparation plant roughness detection method according to claim 6, characterized in that: In S55, the calculation formula of the roughness level R is shown in formula (7): In the formula, α, β and γ is a normalization parameter, and α、β as well as γ The sum is 1.

11. An integrated coal preparation plant roughing detection and control system, characterized in that: The invention comprises an industrial control integrated machine (16), a programmable controller, an alarm signal generator (17), a feed valve (2), a flushing valve, a discharge valve (11), an actuator (4), a display module, an algorithm module, a relational database and a non-relational database, wherein the feed valve (2), the flushing valve, the discharge valve (11), the actuator (4) and the alarm signal generator (17) are all connected to the programmable controller, the programmable controller, the algorithm module and the display module are connected to the industrial control integrated machine (16), and the display module is used to display a content change curve and a particle size distribution table of "coarse" particles; The relational database and the non-relational database are both connected to the industrial control integrated machine (16); the relational database is used to store preset key parameters required for executing the roughing detection method and the control method after different levels of "alarm", and the non-relational database is used to store the measurement data of the calculation process of the roughing detection method and the execution process of the control method after different levels of "alarm", the feeding screen basket state and the valve opening. The industrial control integrated machine (16) uses the collected data of the relational database and the non-relational database to execute the roughing detection method according to any one of claims 5 to 10 through an algorithm module.

Images