Material stock prediction method and device for bin type feeding machine, storage medium and equipment
By obtaining a variety of data of the warehouse feeder, combining fixed logic and LightGBM model, the accurate prediction of the material stock of the warehouse feeder is achieved, and the problem of low accuracy of material stock prediction in the existing technology is solved, and the improvement of the Prediction accuracy and computational efficiency.
Patent Information
- Application Number
- CN202411172584.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-26
- Publication Date
- 2024-12-31
AI Technical Summary
Existing warehouse feeders cannot accurately measure material stocks, resulting in low accuracy in predicting material stocks.
By obtaining the position data, direction data, time data and material data of the feed reciprocating driving and cabinet of the warehouse feeder, combining fixed logic and LightGBM model, correlation analysis and prediction are carried out, material inventory proportion is calculated, and relevant screening The input variable with a characteristic coefficient greater than the preset value is input into the trained LightGBM model to obtain the predicted value of the material stock.
It improves the accuracy of material inventory prediction, reduces calculation overhead, reduces prediction time error, and prevents data from falling into local optimality within the range of inaccurate data.
Smart Images
Figure CN119227865A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to a method, a device, a storage medium and equipment for predicting the material inventory of a silo feeder, and belongs to the technical field of predicting the material inventory of a silo feeder. Background Art
[0002] The bin feeder is the main auxiliary equipment in the tobacco industry's silk making process, including a feeding reciprocating crane 1, a bin cabinet 2, a conveyor belt at the bottom of the cabinet 3, an elevator 4 and a metering tube 5. Figure 2 As shown. The feeding reciprocating crane 1 is the feeding device, and the warehouse cabinet 2 serves as a buffer for the production process. Under the premise of ensuring continuous production operation, the hoist 4 lifts the material to a high place and drops it into the metering tube 5 through the drop hopper, and the metering tube 5 controls the start and stop of the hoist 4 to ensure the stability of the incoming flow of the subsequent equipment belt scale to ensure the accurate measurement of the belt scale, and at the same time, the rotating cleaning brush continuously rotates to clean the lifting belt. Only under the premise of accurate flow can the accuracy of various indicators such as humidity, temperature, fragrance, addition, blending, etc. in the processes of moistening leaves, adding materials, drying silk, mixing, and adding fragrance be guaranteed, so the warehouse feeder is a necessary pre-equipment for the belt scale.
[0003] The bin feeder is mainly divided into two parts: feeding control and discharging control. The feeding control mainly cooperates with the front-end equipment, including the pre-filling of the bin feeder before production and the replenishment of materials to the bin feeder during production; the discharging control can realize the coordinated operation of the bin feeder and the belt scale, realize the uniform and continuous tobacco output function, and stably provide it to the tobacco drying equipment.
[0004] As the materials in the bin feeder are limited by the existing equipment design structure, the internal inventory cannot be accurately measured. It can only be estimated by relying on the four photoelectric tubes on the front and rear sides of the feeding reciprocating crane, thereby controlling the reciprocating crane to distribute the materials. As the interior of the tobacco material bin feeder presents multiple "small hills", there is no detection device and method for accurately detecting the material inventory in the industry. Summary of the invention
[0005] The object of the present invention is to provide a method, device, storage medium and equipment for predicting material inventory of a silo feeder, so as to solve the problem of low accuracy of material inventory prediction in the prior art.
[0006] To achieve the above objectives, the present invention is implemented by adopting the following technical solutions: In a first aspect, the present invention provides a method for predicting material inventory of a silo feeder, comprising: Obtain the position data, direction data, time data and material data of the feeding reciprocating crane and cabinet in the silo feeder; Based on the acquired data, the predicted value of the material inventory ratio of the silo feeder is calculated through the preset fixed logic; Perform correlation analysis on the acquired data and select input variables whose correlation coefficient with the material inventory of the silo feeder is greater than a preset value; The predicted value of the material inventory ratio of the silo feeder and the input variables are input into the trained LightGBM model to obtain the predicted value of the material inventory of the silo feeder.
[0007] Furthermore, the position data, direction data, time data and material data of the feeding reciprocating crane and the cabinet include: The position of the feeding reciprocating crane is X1. X1 represents the position of the feeding reciprocating crane with the end point of the cabinet in the bin feeder close to the elevator as the zero position and the backward direction of the feeding reciprocating crane as the incremental direction of the horizontal axis; The material status detected by the four photoelectric tubes on the feeding reciprocating carriage: front high X2, front low X3, rear high X4 and rear low X5. The material status of 1 means that the material is detected, and the material status of 0 means that the material is not detected; The material status X6 detected by the photoelectric tube at the front end of the cabinet and the material status X7 detected by the photoelectric tube at the rear end of the cabinet; The feeding direction of the reciprocating conveyor belt is X8. X8 = 1 means forward feeding, X8 = 1 means backward feeding, and X8 = 2 means stop; The conveyor belt advances at the bottom of the cabinet X9; Total length of cabinet L; The number of steps P that the feeding reciprocating crane takes to fill the entire bin feeder from front to back without retreating; The feeding reciprocating crane steps to the new unloading position and continues unloading until the time T when the front high or rear high photoelectric tube detects a high position signal.
[0008] Furthermore, the calculation of the predicted value of the material inventory ratio of the silo feeder by the preset fixed logic includes: The predicted value of the material inventory ratio of the silo feeder is calculated by the formula F= F1+ F2- F3, where F is the predicted value of the material inventory ratio of the silo feeder, F1 is the inventory ratio of the front half of the silo feeder, F2 is the inventory ratio of the back half of the silo feeder, and F3 is the inventory ratio of the material output from the conveyor belt at the bottom of the cabinet.
[0009] Furthermore, the inventory ratio of the first half of the silo feeder is calculated by the following method: If X1<100 and X8=1, then F1=0%; If condition 1 is satisfied within the historical time and condition 2 is satisfied at the current time, then F1 = 50.0%, where condition 1 is L / 2 - 400 < X1 < L / 2 and X2 = 1, and condition 2 is X1 > L / 2 and X8 = 0; If X1 > L / 2 and X8 = 0 within the historical time, then the F1 at the current time remains unchanged and is the same as the value calculated within the historical time; If 400 < X1 < L / 2, X2 ≠ 0, and X2 ≠ 1 within the historical time, then the value of F1 is calculated by the formula F1 = F1 + t / PT, where t represents the time that the bin feeder has been continuously feeding at the new feeding level in the front half of the cabinet; If 400 < X1 < L / 2 and X2 = 1 within the historical time, then the value of F1 is calculated by the formula F1 = F1 + 1 / P.
[0010] Further, the proportion of the stock in the second half of the bin feeder is calculated by the following method: If condition 3 is satisfied within the historical time and condition 4 is satisfied at the current time, then F2 = 50.0%, where condition 3 is L / 2 + 400 < X1 < L and X4 = 1, and condition 4 is that the value of X1 remains unchanged and X8 = 2; If X1 < L / 2 and X8 = 1 within the historical time, then the F2 at the current time remains unchanged and is the same as the value calculated within the historical time; If L / 2 - 400 < X1 < L, X4 ≠ 0, and X4 ≠ 1 within the historical time, then the value of F2 is calculated by the formula F2 = F2 + t' / PT, where t' represents the time that the bin feeder has been continuously feeding at the new feeding level in the second half of the cabinet; If L / 2 - 400 < X1 < L and X4 = 1 within the historical time, then the value of F2 is calculated by the formula F2 = F2 + 1 / P.
[0011] Further, the proportion of the stock output by the conveyor belt at the bottom of the cabinet is calculated by the following formula: F3 = X9 / L.
[0012] Further, the correlation analysis is performed on the obtained data, and the input variables with a correlation coefficient greater than the preset value with the stock of the bin feeder are screened out, including: The correlation coefficients between the obtained data and the stock of the bin feeder are calculated by the Pearson formula, and the data with a correlation coefficient greater than 0.2 with the stock of the bin feeder are screened out, and the screened data are used as input variables.
[0013] In a second aspect, the present invention provides a device for predicting the stock of a bin feeder, including: The data acquisition module is configured to: acquire position data, direction data, time data and material data of the feeding reciprocating crane and the cabinet in the silo feeder; The silo feeder material inventory ratio prediction module is configured to: calculate the silo feeder material inventory ratio prediction value through a preset fixed logic based on the acquired data; The correlation analysis module is configured to: perform correlation analysis on each acquired data, and screen out input variables whose correlation coefficient with the material inventory of the silo feeder is greater than a preset value; The silo feeder material inventory prediction module is configured to: input the silo feeder material inventory ratio prediction value and the input variable into the trained LightGBM model to obtain the silo feeder material inventory prediction value.
[0014] In a third aspect, the present invention provides a computer-readable storage medium having a computer program / instruction stored thereon, which, when executed by a processor, implements the steps of predicting the material inventory of a silo feeder as described in any one of the first aspects.
[0015] In a fourth aspect, the present invention provides a computer device, comprising: Memory, for storing computer programs / instructions; A processor is used to execute the computer program / instructions to implement the steps of the method for predicting material inventory of a silo feeder as described in any one of the first aspects.
[0016] Compared with the prior art, the beneficial effects achieved by the present invention are: The present invention provides a method, device, storage medium and equipment for predicting the material inventory of a silo feeder. Based on the principle of a general silo feeder and the operating rules of upstream and downstream equipment, a fixed logic is designed in combination with a variety of predictable factors to fit the inventory ratio of the silo feeder. A large estimate is made based on the logic of the equipment mechanism, so that the overall algorithm starts searching from relatively accurate data, which greatly reduces the computational overhead of data search, while also improving the effect of model convergence and preventing data from falling into local optimality within the range of other imprecise data. The LightGBM algorithm based on machine learning is used to correct the influence of factors that are unpredictable or have large estimation deviations on the predicted silo feeder inventory, thereby reducing the prediction time error and improving the prediction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] Figure 1 It is a flow chart of a method for predicting material inventory of a silo feeder corresponding to Embodiment 1 of the present invention; Figure 2 It is a structural schematic diagram of a bin-type feeder provided in an embodiment of the present invention; Figure 3It is a flow chart of a method for predicting material inventory of a silo feeder corresponding to Embodiment 2 of the present invention; Figure 4 is a schematic diagram of a prediction algorithm G1 based on fixed logic provided in Example 2 of the present invention; Figure 5 is a schematic diagram of a detection device on a bin-type feeder provided in Example 2 of the present invention; In the figure: 1. Feed reciprocating crane; 2. Warehouse cabinet; 3. Conveyor belt at the bottom of the cabinet; 4. Elevator; 5. Measuring tube; 6. Transmitter end of infrared distance measuring device; 7. Receiving end of infrared distance measuring device; 8. Detection grating; 9. Proximity switch. DETAILED DESCRIPTION
[0018] The present invention will be further described below in conjunction with the accompanying drawings. The following embodiments are only used to more clearly illustrate the technical solutions of the present invention and are not intended to limit the protection scope of the present invention.
[0019] Example 1.
[0020] like Figure 1 As shown, the present invention provides a method for predicting material inventory of a silo feeder, comprising: Obtain the position data, direction data, time data and material data of the feeding reciprocating crane and cabinet in the silo feeder; Based on the acquired data, the predicted value of the material inventory ratio of the silo feeder is calculated through the preset fixed logic; Perform correlation analysis on the acquired data and select input variables whose correlation coefficient with the material inventory of the silo feeder is greater than a preset value; The predicted value of the material inventory ratio of the silo feeder and the input variables are input into the trained LightGBM model to obtain the predicted value of the material inventory of the silo feeder.
[0021] The method provided by the present invention is based on the principle of a universal silo feeder and the operating rules of upstream and downstream equipment, combines a variety of predictable factors, and designs a fixed logic to fit the inventory ratio of the silo feeder, and makes a large estimate based on the logic of the equipment mechanism, so that the overall algorithm starts searching from relatively accurate data, which greatly reduces the computational overhead of data search, while also improving the effect of model convergence and preventing data from falling into local optimality within the range of other imprecise data; the LightGBM algorithm based on machine learning is used to correct the influence of factors that are unpredictable or have large estimation deviations on the prediction of silo feeder inventory, thereby reducing the prediction time error and improving the prediction accuracy.
[0022] Example 2.
[0023] like Figure 3As shown, the present invention provides a method for predicting material inventory of a silo feeder, comprising: Step S101, establish a prediction algorithm G1 based on fixed logic, such as Figure 4 shown.
[0024] The transmitting end 6 of the infrared distance measuring device is installed on the top of the cabinet of the bin-type feeding machine, and the receiving end 7 of the infrared distance measuring device is installed on the feeding reciprocating carriage. Figure 5 As shown. After mechanical design and installation, the infrared distance measuring device can directly measure the position of the feed reciprocating vehicle on the top of the bin feeder cabinet, which can be used as an important factor to estimate the inventory of the bin feeder. The end point of the bin cabinet close to the elevator is taken as the zero position, and the backward direction of the feed reciprocating vehicle is taken as the incremental direction of the horizontal axis. The position of the feed reciprocating vehicle is defined as X1 (unit: mm). X1 is a continuous variable.
[0025] The four photoelectric tubes of front high X2, front low X3, rear high X4, and rear low X5 on the front and rear sides of the feeding reciprocating crane are used as input factors X2, X3, X4, and X5 respectively. The high-position photoelectric tube detects that there is material, which means that the material at the front or rear end of the crane is at a high position. The low-position photoelectric tube detects that there is material, which means that the material at the front or rear end of the crane is at a low position. X2 to X5 are discrete quantities, 1 represents the detection of material, and 0 represents the non-detection of material. If the front high X2 is 1 and the front low X3 is 0, it means a detection error. In the fitting algorithm, X2 and X3 are assigned NaN, which means that the high and low status of the front material is unknown; similarly, if the rear high X4 is 1 and the rear low X5 is 0, it also means a detection error. In the fitting algorithm, X4 and X5 are assigned NaN, which means that the high and low status of the rear material is unknown.
[0026] The feeding direction of the feeding reciprocating conveyor belt is defined as X8. When the reciprocating conveyor belt moves in the direction of lifting and discharging, it represents forward feeding; when the reciprocating conveyor belt moves in the opposite direction of lifting and discharging, it represents backward feeding. X8 is a discrete value, 1 represents forward feeding, 0 represents backward feeding, and 2 represents stop.
[0027] N metal detection components 8 (detection gratings 8) are additionally installed at equal distances on the rotating component of the conveyor belt at the bottom of the cabinet, and proximity switches 9 are installed near the metal detection components. Figure 5 As shown. By counting the number of times the proximity switch is on and off, p, the rotation angle of the bottom conveyor belt rotating part is calculated, and finally the forward distance of the bottom conveyor belt of the cabinet is calculated. The calculation method is as follows: X9=Dπp / N; Among them, X9 is the forward distance of the conveyor belt at the bottom of the cabinet, in mm, i.e. millimeter. D is the diameter of the rotating part of the bottom conveyor belt, which is a fixed value; p is the number of times the counting proximity switch is turned on and off; N is the number of metal detection components installed at equal distances on the rotating part of the bottom conveyor belt, which is a fixed value.
[0028] Define the total length of the bin feeder cabinet as L (unit: mm, i.e., millimeter).
[0029] Define the number of steps P for the feeding reciprocating trolley of the bin feeder to move forward from front to back without retreating to fill the entire bin feeder. P is obtained from the feeding material flow rate, and the calculation formula is: P = K1 / S 进料 ; Where, P is the number of steps for the feeding reciprocating trolley of the bin feeder to move forward from front to back without retreating to fill the entire bin feeder, and S 进料 is the feeding material flow rate, which is inversely proportional to the number of steps. The coefficient K1 is obtained by taking the average of multiple experiments and is an empirical parameter.
[0030] Define the time T (unit: second) for the feeding reciprocating trolley of the bin feeder to step to a new blanking position and continue feeding until the high-level signal is detected by the front high or rear high photoelectric tube. T is obtained from the feeding material flow rate, and the calculation formula is: T = K2 / S 进料 ; Where, T is the duration for the feeding reciprocating trolley of the bin feeder to step to a new blanking position and continue feeding until the high-level signal is detected by the front high or rear high photoelectric tube, and S 进料 is the feeding material flow rate, which is inversely proportional to T. The coefficient K2 is obtained by taking the average of multiple experiments and is an empirical parameter.
[0031] The predicted value of the inventory ratio of the bin feeder based on fixed logic is defined as F, and its calculation formula is defined as: F = F1 + F2 - F3; Where, F1 is the ratio of the inventory in the front half of the bin feeder, which is used to calculate the ratio of the inventory in the front half of the bin feeder when the feeding reciprocating trolley conveyor belt moves forward; F2 is the ratio of the inventory in the rear half of the bin feeder, which is used to calculate the ratio of the inventory in the rear half of the bin feeder when the feeding reciprocating trolley conveyor belt moves backward; F3 is the ratio of the inventory of the material output by the conveyor belt at the bottom of the cabinet.
[0032] The calculation method of F1 is as follows: (1) If a new production batch starts, the feeding reciprocating trolley starts from the initial position (within 100 mm from the zero point) and the reciprocating trolley conveyor belt feeds forward, that is, X1 < 100 and X8 = 1, then F1 = 0%.
[0033] (2) If the condition 1 is satisfied in the previous period and the condition 2 is satisfied at the current time, then F1 = 50.0%; Condition 1 is that when the feeding reciprocating trolley of the bin feeder is searching for a position in the front half of the cabinet in the previous period, the front high is always 1, that is, when L / 2 - 400 < X1 < L / 2 is satisfied in the previous period, and X2 = 1; Condition 2 is that the reciprocating feeder of the bin feeder starts to move backward to the second half of the cabinet at the current time, and the conveyor belt of the reciprocating feeder feeds backward, that is, when the current time satisfies X1 > L / 2 and X8 = 0.
[0034] (3) If the reciprocating feeder of the bin feeder has always been within the range of the second half of the cabinet in the previous period of time, and the conveyor belt of the reciprocating feeder has always fed backward, that is, in the previous period of time, when X1 > L / 2 and X8 = 0, then F1 remains unchanged.
[0035] (4) If the reciprocating feeder of the bin feeder has always moved forward to the first half of the cabinet in the previous period of time, and the conveyor belt of the reciprocating feeder has never fed backward, that is, in the previous period of time, when 400 < X1 < L / 2 and X2 ≠ 0, then ; where t is the time that the bin feeder has continuously fed at the new feeding position in the first half of the cabinet body; P is the number of steps for the reciprocating feeder of the bin feeder to fill the entire bin feeder from front to back without retreating; T is the duration from when the reciprocating feeder of the bin feeder steps to the new empty feeding position and continuously feeds until the high-level signal is detected by the front high or rear high photoelectric tube.
[0036] The meaning of the above formula is that when the bin feeder is feeding at the new feeding position in the first half of the cabinet body, if the front high does not detect a signal, it is considered that the material is fed at the integral value rate of t / T, and the proportion of each step full is 1 / P. Therefore, F1 = F1 + t / PT; When the bin feeder is feeding at the new feeding position in the first half of the cabinet body, if the front high detects a signal, it means that the material in this step is full, and the proportion of this step full is 1 / P. Therefore, F1 = F1 + 1 / P.
[0037] The calculation method of F2 is as follows: (1) If the condition 3 is satisfied in the previous period of time and the condition 4 is satisfied at the current time, then F2 = 50.0%; Condition 3 is that when the reciprocating feeder of the bin feeder is searching for a position in the second half of the cabinet in the previous period of time, the rear high is always 1, that is, in the previous period of time, when L / 2 + 400 < X1 < L and X4 = 1; Condition 4 is that the reciprocating feeder of the bin feeder stops moving at the current time, and the conveyor belt of the reciprocating feeder stops feeding, that is, at the current time, X1 remains unchanged and X8 = 2.
[0038] (2) If the reciprocating feeder of the bin feeder has always been within the range of the first half of the cabinet in the previous period of time, and the conveyor belt of the reciprocating feeder has always fed forward, that is, in the previous period of time, when X1 < L / 2 and X8 = 1, then F2 remains unchanged.
[0039] (3) If the feeding reciprocating trolley of the bin feeder has always moved towards the rear half of the cabinet during the previous period, and the conveyor belt of the reciprocating trolley has never fed forward, that is, when L / 2 - 400 < X1 < L is satisfied during the previous period and X4 ≠ 0, then ; Among them, t’ represents the time that the bin feeder has continuously fed at the new feeding position in the rear half of the cabinet body, P is the number of steps for the feeding reciprocating trolley of the bin feeder to fill the entire bin feeder from front to back without retreating; T is the duration from when the feeding reciprocating trolley of the bin feeder steps to the new empty feeding position and continuously feeds until the high-level signal is detected by the front high or rear high photoelectric tube.
[0040] The meaning of the above formula is that when the bin feeder is feeding at the new feeding position in the rear half of the cabinet body, if the rear high signal is not detected, it is considered that the material is fed at the rate integral value of t / T, and the occupancy ratio of each step position full is 1 / P. Therefore, F2 = F2 + t’ / PT; When the bin feeder is feeding at the new feeding position in the rear half of the cabinet body, if the rear high signal is detected, it means that the material in this step position is full, and the occupancy ratio of this step position full is 1 / P. Therefore, F2 = F2 + 1 / P.
[0041] The calculation method of F3 is as follows: Estimate the proportion of the discharged material according to the forward distance X9 of the conveyor belt at the bottom of the cabinet body. The formula is: F3 = X9 / L; Among them, X9 is the forward distance of the conveyor belt at the bottom of the cabinet body, and L is the total length of the cabinet body.
[0042] Step S102, establish a LightGBM model (bin feeder inventory prediction model G2).
[0043] Step S102-1, define the definitions and calculation formulas of all X inputs and Y output factors of the algorithm.
[0044] In G2, X1 is still defined as the position of the feeding reciprocating trolley, which is consistent with the definition in G1.
[0045] In G2, X2 to X5 are still defined as the detection states of the four photoelectric tubes of front high, front low, rear high, and rear low on both sides of the feeding reciprocating trolley, which are consistent with the definition in G1.
[0046] The detection state of the opposed photoelectric tube at the front end of the bin feeder cabinet body for the material is defined as X6, and the detection state of the opposed photoelectric tube at the rear end of the bin feeder cabinet body for the material is defined as X7. Among them, X7 and X8 are discrete quantities, 1 represents that the material is detected, and 0 represents that the material is not detected.
[0047] In G2, X8 is still defined as the feeding direction of the conveyor belt of the feeding reciprocating trolley, which is consistent with the definition in G1.
[0048] X9 in G2 is still defined as the forward distance of the conveyor belt at the bottom of the cabinet, which is consistent with the definition in G1.
[0049] Add variable X in G2 10 , defined as the output bin feeder inventory share prediction value F of the G1 fixed algorithm, that is: X 10 =F.
[0050] M groups of detection gratings 8 are temporarily installed at equal intervals on the sides of the bin feeder; wherein a detection grating light source transmitter is installed on one side of the bin feeder, and a detection grating light source receiver is installed on the other side, such as Figure 5 As shown. The M group of detection gratings 8 are temporarily installed and used for machine learning algorithm training models. After the model is fitted, the M group of detection gratings 8 can be removed and then installed on other silo feeders for retraining. This training method is relatively cost-effective, and can also improve detection accuracy. It can also enable each silo feeder to establish an independent model to avoid the decrease in accuracy caused by the replication effect of the model. The method of marking the material inventory of the silo feeder by using the M group of detection gratings is as follows: ; Among them, Y1 is the label value of the material inventory of the silo feeder, which is a continuous variable and is used to train the model; i represents the i-th group of detection gratings; H represents the material height ratio of each group of detection gratings (ranging from 0% to 100%); H i It represents the material height proportion of the i-th group of detection gratings, and therefore represents the cumulative value of the material height proportion of the M-group detection gratings. Its ratio to the M value is approximately equal to the material inventory of the silo feeder.
[0051] Step S102-2, export historical data in data format. Data collected at fixed time intervals during tobacco shred production are used as data rows, and X1 to X9 and Y1 values are exported at fixed time intervals. All historical data are marked with time t as the data row, and historical data are exported.
[0052] Step S102-3, perform correlation analysis on input variables X1 to X9. Use Pearson's formula to perform correlation analysis, draw a correlation matrix, and remove variables with a correlation coefficient of less than 0.2 with the Y1 value from input variables X1 to X9. Finally, remove variables with a correlation coefficient of greater than 0.2 with the Y1 value from input variables X1 to X9 and X1. 10 The variables are used as input variables of the LightGBM model.
[0053] Step S102-4, training the LightGBM model based on the historical data X and Y1 values.
[0054] Step S103, prediction step.
[0055] Get real-time X1 to X9 in silo feeder; Based on the acquired data, the predicted value of the material inventory ratio of the silo feeder is calculated through a preset fixed logic (G1 constructed in step S101); Conduct correlation analysis on the acquired data and select input variables with a correlation coefficient greater than 0.2 with the material inventory of the silo feeder; The predicted value of the material inventory ratio of the silo feeder and the input variables are input into the trained LightGBM model (G2 obtained in step S102) to obtain the predicted value Y of the material inventory of the silo feeder.
[0056] The final Y is continuous normalized data, representing the ratio of the material to the entire silo feeder (ranging from 0%-100%).
[0057] Example 3.
[0058] The present invention provides a device for predicting the material inventory of a bin-type feeder, comprising: The data acquisition module is configured to: acquire position data, direction data, time data and material data of the feeding reciprocating crane and the cabinet in the silo feeder; The silo feeder material inventory ratio prediction module is configured to: calculate the silo feeder material inventory ratio prediction value through a preset fixed logic based on the acquired data; The correlation analysis module is configured to: perform correlation analysis on each acquired data, and screen out input variables whose correlation coefficient with the material inventory of the silo feeder is greater than a preset value; The silo feeder material inventory prediction module is configured to: input the silo feeder material inventory ratio prediction value and the input variable into the trained LightGBM model to obtain the silo feeder material inventory prediction value.
[0059] Example 4.
[0060] The present invention provides a computer-readable storage medium having a computer program / instruction stored thereon. When the computer program / instruction is executed by a processor, the steps of predicting the material inventory of a silo feeder provided in Example 1 are implemented: Obtain the position data, direction data, time data and material data of the feeding reciprocating crane and cabinet in the silo feeder; Based on the acquired data, the predicted value of the material inventory ratio of the silo feeder is calculated through the preset fixed logic; Perform correlation analysis on the acquired data and select input variables whose correlation coefficient with the material inventory of the silo feeder is greater than a preset value; The predicted value of the material inventory ratio of the silo feeder and the input variables are input into the trained LightGBM model to obtain the predicted value of the material inventory of the silo feeder.
[0061] Example 5.
[0062] The present invention provides a computer device, comprising: Memory, for storing computer programs / instructions; A processor is used to execute the computer program / instructions to implement the steps of the method for predicting the material inventory of a silo feeder provided in Example 1: Obtain the position data, direction data, time data and material data of the feeding reciprocating crane and cabinet in the silo feeder; Based on the acquired data, the predicted value of the material inventory ratio of the silo feeder is calculated through the preset fixed logic; Perform correlation analysis on the acquired data and select input variables whose correlation coefficient with the material inventory of the silo feeder is greater than a preset value; The predicted value of the material inventory ratio of the silo feeder and the input variables are input into the trained LightGBM model to obtain the predicted value of the material inventory of the silo feeder.
[0063] Those skilled in the art will appreciate that the embodiments of the present application may be provided as methods, systems, or computer program products. Therefore, the present application may adopt the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may adopt the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes.
[0064] The present application is described with reference to the flowcharts and / or block diagrams of the methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each process and / or box in the flowchart and / or block diagram, as well as the combination of the processes and / or boxes in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0065] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0066] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0067] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the technical principles of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.
Claims
1. A method for predicting material inventory of a silo feeder, characterized in that: Including: Obtaining the position data, direction data, time data, and material data of the feeding reciprocating trolley and the cabinet body in the bin feeder; Based on the obtained various data, calculating the predicted value of the material inventory ratio of the bin feeder through a preset fixed logic; Performing a correlation analysis on the obtained various data, and screening out the input variables whose correlation coefficients with the material inventory of the bin feeder are greater than the preset value; Inputting the predicted value of the material inventory ratio of the bin feeder and the input variables into the trained LightGBM model to obtain the predicted value of the material inventory of the bin feeder.
2. The method for predicting material inventory of a silo feeder according to claim 1, characterized in that: The position data, direction data, time data, and material data of the feeding reciprocating trolley and the cabinet body include: The position X1 of the feeding reciprocating trolley, where X1 represents the position of the feeding reciprocating trolley with the end point of the cabinet body in the bin feeder close to the elevator as the zero position, the backward traveling direction of the feeding reciprocating trolley as the incremental direction of the horizontal axis; The material states detected by four photoelectric tubes on the feeding reciprocating trolley: front high X2, front low X3, rear high X4, and rear low X5, where the material state of 1 indicates that material is detected, and the material state of 0 indicates that no material is detected; The material state X6 detected by the opposed photoelectric tube at the front end of the cabinet body and the material state X7 detected by the opposed photoelectric tube at the rear end of the cabinet body; The feeding direction X8 of the conveyor belt of the feeding reciprocating trolley, where X8 = 1 indicates forward feeding, X8 = -1 indicates backward feeding, and X8 = 2 indicates stop; The forward distance X9 of the conveyor belt at the bottom of the cabinet body; The total length L of the cabinet body; The stepping times P for the feeding reciprocating trolley to fill the entire bin feeder from front to back without retreating; The time T for the feeding reciprocating trolley to step into a new blanking position and continuously feed until a high-level signal is detected by the front high or rear high photoelectric tube.
3. The method for predicting material inventory of a silo feeder according to claim 2, characterized in that: The calculating the predicted value of the material inventory ratio of the bin feeder through a preset fixed logic includes: Calculating the predicted value of the material inventory ratio of the bin feeder through the formula F = F1 + F2 - F3, where F is the predicted value of the material inventory ratio of the bin feeder, F1 is the ratio of the inventory in the first half of the bin feeder, F2 is the ratio of the inventory in the second half of the bin feeder, and F3 is the ratio of the output material inventory of the conveyor belt at the bottom of the cabinet body.
4. The method for predicting material inventory of a silo feeder according to claim 3, characterized in that: The ratio of the inventory in the first half of the bin feeder is calculated by the following method: If X1 < 100 and X8 = 1, then F1 = 0%; If the condition 1 is satisfied within the historical time and the condition 2 is satisfied at the current time, then F1 = 50.0%, where the condition 1 is L / 2 - 400 < X1 < L / 2 and X2 = 1, and the condition 2 is X1 > L / 2 and X8 = 0; If X1 > L / 2 and X8 = 0 within the historical time, then the F1 at the current time remains unchanged and is the same as the value calculated within the historical time; If 400 < X1 < L / 2 and X2 ≠ 0 and X2 ≠ 1 within the historical time, then the value of F1 is calculated through the formula F1 = F1 + t / PT, where t represents the time for the bin feeder to continuously feed at the new blanking position in the first half of the cabinet body; If 400 < X1 < L / 2 and X2 = 1 within the historical time, then the value of F1 is calculated through the formula F1 = F1 + 1 / P.
5. The method for predicting material inventory of a silo feeder according to claim 3, characterized in that: The ratio of the inventory in the second half of the bin feeder is calculated by the following method: If condition 3 is satisfied within the historical time and condition 4 is satisfied at the current time, then F2 = 50.0%, where condition 3 is L / 2 + 400 < X1 < L and X4 = 1, and condition 4 is that the value of X1 remains unchanged and X8 = 2; If X1 < L / 2 and X8 = 1 within the historical time, then F2 at the current time remains unchanged and is the same as the value calculated within the historical time; If L / 2 - 400 < X1 < L, X4 ≠ 0, and X4 ≠ 1 within the historical time, then the value of F2 is calculated by the formula F2 = F2 + t’ / PT, where t’ represents the time that the bin feeder has been continuously feeding at the new lower material level in the second half of the cabinet; If L / 2 - 400 < X1 < L and X4 = 1 within the historical time, then the value of F2 is calculated by the formula F2 = F2 + 1 / P.
6. The method for predicting material inventory of a silo feeder according to claim 3, characterized in that: The proportion of the material output by the conveyor belt at the bottom of the cabinet is calculated by the following formula: F3 = X9 / L.
7. The method for predicting material inventory of a silo feeder according to claim 1, characterized in that: Performing correlation analysis on the obtained various data, and screening out input variables whose correlation coefficients with the material inventory of the bin feeder are greater than the preset value, including: Calculating the correlation coefficients between the obtained various data and the material inventory of the bin feeder through the Pearson formula, screening out the data whose correlation coefficients with the material inventory of the bin feeder are greater than 0.2, and using the screened data as input variables.
8. A device for predicting material inventory of a silo feeder, characterized in that: Including: A data acquisition module, configured to: acquire the position data, direction data, time data, and material data of the feeding reciprocating trolley and the cabinet in the bin feeder; A bin feeder material inventory proportion prediction module, configured to: based on the obtained various data, calculate the predicted value of the bin feeder material inventory proportion through a preset fixed logic; A correlation analysis module, configured to: perform correlation analysis on the obtained various data, and screen out input variables whose correlation coefficients with the material inventory of the bin feeder are greater than the preset value; A bin feeder material inventory prediction module, configured to: input the predicted value of the bin feeder material inventory proportion and the input variables into the trained LightGBM model to obtain the predicted value of the bin feeder material inventory.
9. A computer-readable storage medium having a computer program / instruction stored thereon, characterized in that: When the computer program / instructions are executed by the processor, the steps of predicting the material inventory of the bin feeder described in any one of claims 1 to 7 are implemented.
10. A computer device, characterized in that: Including: A memory for storing the computer program / instructions; A processor for executing the computer program / instructions to implement the steps of the method for predicting the material inventory of the bin feeder described in any one of claims 1 to 7.