An ore sorting self-adaptive adjustment method and a sorting machine

Abstract

The present disclosure relates to the technical field of sorting equipment, and particularly relates to a mineral sorting self-adaptive adjusting method and a sorting machine. The mineral sorting self-adaptive adjusting method comprises the following steps: obtaining characteristic values of a plurality of sorted minerals, and taking the characteristic values as reference characteristic values; for each reference characteristic value, determining an estimated grade value of the corresponding mineral according to a grade estimation model; determining a current predicted grade mean value according to a plurality of estimated grade values; determining a current grade deviation value according to the current predicted grade mean value and a target grade mean value; determining a current adjustment amount according to the current grade deviation value based on a PID control method; and adjusting sorting control parameters of the sorting machine according to the current adjustment amount. According to the present disclosure, the sorting control parameters such as the characteristic threshold value adjusted based on the current adjustment amount can be smoothly, accurately and quickly adjusted, so that the grade fluctuation of the mineral can be quickly responded to and the grade fluctuation can be effectively inhibited.

Description

Technical Field

[0001] This application mainly relates to the field of sorting equipment technology, and in particular to an adaptive adjustment method and sorting machine for ore sorting. Background Technology

[0002] In the field of ore sorting, sorting machines typically have identification devices that analyze ore parameters and determine the ore type, such as concentrate and tailings, based on the analysis results. Thus, the sorting unit in the sorting machine can separate the ore according to its type.

[0003] However, in actual production, the grades of target minerals in a large number of ores are different, which causes the concentrate and tailings grades of the ore batches sorted by the separator to fluctuate greatly at different time periods, making it difficult to effectively stabilize the grades. Summary of the Invention

[0004] To overcome the problems existing in related technologies, an exemplary embodiment of this disclosure provides, in a first aspect, an adaptive adjustment method for ore sorting, applied to a sorting machine. The sorting machine includes a data acquisition device and a classification device. The data acquisition device is configured to acquire feature values ​​of the ore, and the classification device is configured to determine the classification result of the ore based on the feature values ​​and feature thresholds. The classification result includes tailings and concentrate. The adaptive adjustment method for ore sorting includes: acquiring multiple feature values ​​of sorted ore and using the feature values ​​as reference feature values; for each reference feature value, determining the estimated grade value of the corresponding ore based on a grade estimation model; determining the current predicted grade average based on the multiple estimated grade values; determining the current grade deviation value based on the current predicted grade average and the target grade average; determining the current adjustment amount based on the current grade deviation value using a PID control method; and adjusting the sorting control parameters of the sorting machine based on the current adjustment amount, wherein the sorting control parameters include feature thresholds.

[0005] In some embodiments, the current adjustment amount is determined based on a proportional term, an integral term, and a derivative term. The proportional term includes the proportional gain and the current grade deviation value, the integral term includes the integral gain and the integral of the current grade deviation value, and the derivative term includes the derivative gain and the derivative of the current grade deviation value. Based on the PID control method, before determining the current adjustment amount according to the current grade deviation value, the following steps are taken: if the error zero-crossing frequency corresponding to the current grade deviation value within a preset time period is greater than a preset frequency threshold, then the proportional gain is decreased or the derivative gain is increased; and / or if the absolute value of the current grade deviation value is not less than a preset dead zone threshold for a duration greater than a preset deviation duration, then the proportional gain is increased or the integral gain is decreased; and / or if the absolute value of the current grade deviation value is greater than a preset threshold, then the integral gain is set to 0.

[0006] In some embodiments, adjusting the sorting control parameters of the sorter according to the current adjustment amount includes: acquiring feedforward characteristic values ​​of multiple ores to be sorted; determining a feedforward adjustment amount based on the feedforward characteristic values; and adjusting the sorting control parameters according to the current adjustment amount and the feedforward adjustment amount.

[0007] In some embodiments, the feedforward feature value includes a feature value; determining the feedforward adjustment amount based on the feedforward feature value includes: determining the average feedforward grade based on a plurality of feedforward feature values; and determining the feedforward adjustment amount based on the average feedforward grade and the average target grade.

[0008] In some embodiments, the feedforward feature values ​​include one or more of the following: average transmission gray values ​​corresponding to multiple ores before sorting, surface color distribution histogram of each ore before sorting, and surface texture roughness of each ore before sorting; determining the feedforward adjustment amount based on the feedforward feature values ​​includes: determining the predicted grade trend change amount corresponding to multiple ores to be sorted according to the feedforward feature values; and determining the feedforward adjustment amount according to the predicted grade trend change amount.

[0009] In some embodiments, adjusting the sorting control parameters of the sorting machine according to the current adjustment amount includes: obtaining the average historical true grade corresponding to multiple historical sorting windows, wherein each historical sorting window includes multiple sorted ores; determining the average advanced predicted grade corresponding to the current sorting window based on the multiple average historical true grades; determining the advanced adjustment amount based on the average advanced predicted grade and the average target grade; and adjusting the sorting control parameters of the sorting machine based on the current adjustment amount and the advanced adjustment amount.

[0010] In some embodiments, the grade estimation model includes feature values ​​and true grade values ​​corresponding to multiple ore samples. For each benchmark feature value, the estimated grade value of the corresponding ore is determined according to the grade estimation model, including: for each benchmark feature value, based on linear interpolation, the estimated average grade of the corresponding ore is determined according to the two feature values ​​adjacent to the benchmark feature value in the grade estimation model and the corresponding true grade value.

[0011] In some embodiments, determining the current predicted average grade based on multiple estimated grade values ​​includes: for each sorted ore, determining the estimated weight corresponding to the ore based on the difference between the ore sorting time and the sampling time, wherein the estimated weight increases as the difference decreases; and determining the current predicted average grade based on the estimated weights and estimated grade values ​​corresponding to multiple ores.

[0012] In some embodiments, the sorting machine further includes a sorting device that sorts the ore according to the classification result. The sorting control parameters also include a sorting frequency. The sorting control parameters of the sorting machine are adjusted according to the current adjustment amount, including: determining the adjusted feature threshold according to the current adjustment amount; and determining the adjusted sorting frequency according to the adjusted feature threshold.

[0013] Secondly, this disclosure also provides a sorting machine, comprising: a data acquisition device configured to acquire characteristic values ​​of ore; a classification device configured to determine the classification result of the ore based on the characteristic values ​​and a characteristic threshold, wherein the classification result includes tailings and concentrate; and an adjustment device for adjusting the sorting control parameters of the sorting machine according to the ore sorting adaptive adjustment method of the first aspect, wherein the sorting control parameters include the characteristic threshold.

[0014] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure.

[0015] According to the ore sorting adaptive adjustment method provided in this disclosure, the overall grade of the most recently sorted batch of ore, i.e., the current predicted average grade, can be effectively determined based on the characteristic values ​​of multiple sorted ores corresponding to the most recent sampling window. Based on the current grade deviation value, which reflects the deviation between the current predicted average grade and the target average grade, a corresponding current adjustment amount is obtained using a PID control method. This allows the characteristic threshold to be adjusted according to the current adjustment amount, enabling the sorter to bring the overall grade of subsequent batches of sorted ore closer to the target average grade. Because of the PID control method, sorting control parameters such as the characteristic threshold adjusted based on the current adjustment amount can be smoothly, accurately, and quickly adjusted. This allows for rapid response to and effective suppression of ore grade fluctuations, thus improving the sorter's response speed and enabling timely updates to the control of the next batch of ore to be sorted. Furthermore, since the characteristic values ​​of the sorted ore can be obtained in real time, the grade fluctuations of the ore can be detected in real time based on the characteristic values, and the sorting control parameters can be adjusted. This avoids excessive manual intervention by operators and can greatly improve the degree of automation and real-time adjustability, thereby further improving the overall sorting efficiency of the sorting machine. Attached Figure Description

[0016] The accompanying drawings are included to provide a further understanding of this application; they are incorporated into and constitute a part of this application. The drawings illustrate embodiments of this application and, together with this specification, serve to explain the principles of this application. In the drawings: Figure 1 This is a flowchart illustrating an adaptive adjustment method for ore sorting according to an exemplary disclosed invention; Figure 2This is a flowchart illustrating an adaptive adjustment method for ore sorting according to an exemplary disclosed invention; Figure 3 This is a flowchart illustrating an adaptive adjustment method for ore sorting according to an exemplary disclosed invention; Figure 4 This is a flowchart illustrating an adaptive adjustment method for ore sorting according to an exemplary disclosed invention; Figure 5 This is a flowchart illustrating an adaptive adjustment method for ore sorting according to an exemplary disclosed invention; Figure 6 This is a flowchart illustrating an adaptive adjustment method for ore sorting according to an exemplary disclosed invention; Figure 7 This is a flowchart illustrating an adaptive adjustment method for ore sorting according to an exemplary disclosed invention; Figure 8 This is a flowchart illustrating an adaptive adjustment method for ore sorting according to an exemplary disclosed invention; Figure 9 This is a flowchart illustrating an adaptive adjustment method for ore sorting according to an exemplary embodiment disclosed in a publication. Detailed Implementation

[0017] The following describes specific embodiments of this disclosure. It should be noted that, in order to maintain brevity, this specification cannot provide a detailed description of all features of the actual embodiments. It should be understood that, in the actual implementation of any embodiment, just as in any engineering or design project, various specific decisions are often made to achieve the developer's specific goals and to meet system-related or business-related constraints, and this can change from one embodiment to another. Furthermore, it is understood that although the efforts made in this development process may be complex and lengthy, for those skilled in the art related to the content of this disclosure, changes in design, manufacturing, or production based on the technical content disclosed herein are merely conventional technical means and should not be construed as insufficient content of this disclosure.

[0018] Unless otherwise defined, the technical or scientific terms used in the claims and description shall have the ordinary meaning understood by one of ordinary skill in the art to which this disclosure pertains. The terms “first,” “second,” and similar terms used in the specification and claims of this patent application do not indicate any order, quantity, or importance, but are merely used to distinguish different components. The terms “an” or “a” and similar terms do not indicate a quantity limitation, but rather indicate the presence of at least one. The terms “comprising” or “including” and similar terms mean that the element or object preceding “comprising” or “including” encompasses the element or object listed following “comprising” or “including” and its equivalents, and do not exclude other elements or objects. The terms “connected” or “linked” and similar terms are not limited to physical or mechanical connections, nor are they limited to direct or indirect connections.

[0019] The target minerals are unevenly distributed in mines, leading to fluctuations in the grade of the mined ore. Furthermore, the presence of unusable or nonexistent tailings in the mine further contributes to significant fluctuations in the grade of the ore fed into the separator. When a fixed separation threshold is used in conjunction with the characteristic values ​​of the ore to be separated, the grades of both the separated concentrate and tailings fluctuate wildly, failing to stabilize within the expected target range. In some related technologies, operators analyze the separated concentrate to determine the true concentrate grade and then manually adjust the separation parameters—i.e., fix the separation threshold—based on this true concentrate grade to reduce grade fluctuations. However, while this method typically improves the suppression of grade fluctuations with increased operator intervention, it obviously requires substantial manpower to effectively control these fluctuations. In addition, this method requires a certain amount of operation time, which may cause production to be interrupted intermittently, thus reducing the overall sorting efficiency of the sorting machine. Alternatively, in continuous production, the adjustment of the fixed sorting threshold may be delayed, resulting in poor sorting effect and difficulty in effectively suppressing grade fluctuations.

[0020] To solve the above technical problems, such as Figure 1 As shown, this disclosure provides an adaptive adjustment method for ore sorting, which can be applied to a sorting machine. The sorting machine includes a data acquisition device and a classification device. The data acquisition device is configured to acquire feature values ​​of the ore, and the classification device is configured to determine the classification result of the ore based on the feature values ​​and a feature threshold. The classification result includes tailings and concentrate. Since the ore corresponding to tailings and the ore corresponding to concentrate have different and non-overlapping feature value ranges, by setting a feature threshold as a comparison benchmark, the ore can be determined as concentrate or tailings by comparing the feature threshold with the feature value corresponding to the ore to be sorted. The adaptive adjustment method for ore sorting may include steps S110 to S160.

[0021] Step S110: Obtain feature values ​​of multiple sorted ores and use these feature values ​​as baseline feature values. Feature values ​​may include equivalent atomic numbers or KK values. Feature values ​​can be calculated based on data obtained after dual-energy X-ray transmission (DE-XRT) processing of the ores. KK values ​​may include low-energy X-ray transmission signals and high-energy X-ray transmission signals. By setting sampling windows, the continuously processed ore stream in the sorting machine can be effectively divided. That is, the sorting machine continuously sorts multiple ores corresponding to each sampling window, thereby obtaining multiple sorted ores corresponding to each sampling window. The sampling window can be a time window, thus allowing all ores sorted by the sorting machine within a preset time period to be considered as multiple sorted ores in step S110. For example, the preset time period can be 3 minutes, 5 minutes, 10 minutes, etc. In other embodiments, the sampling window can be a quantity window, thus allowing a preset number of ores continuously sorted by the sorting machine to be considered as multiple sorted ores in step S110. In other words, step S110 can be further defined as: acquiring the feature values ​​of multiple sorted ores corresponding to the most recent sampling window, and using these feature values ​​as baseline feature values. In step S110, the feature values ​​of multiple sorted ores can be acquired by storing the feature values ​​of the ores using a data acquisition device. In other embodiments, the corresponding feature values ​​of the sorted ores can be reacquired using a device other than the data acquisition device. Multiple sorted ores can be ore that has been sorted into concentrate by the separator, thereby enabling the determination of the concentrate grade corresponding to the current sorting by the separator through subsequent steps. In other embodiments, multiple sorted ores can be ore that has been sorted into tailings by the separator, thereby determining the tailings grade corresponding to the current sorting by the separator. For a two-stage separator, acquiring only the concentrate grade or tailings grade is sufficient to determine the adjustment method of the sorting control parameters in subsequent steps. In other embodiments, for separators with three or more stages of sorting, the adjustment method of the sorting control parameters corresponding to the grade to be adjusted can be determined by acquiring the sorted ore corresponding to the grade to be adjusted. For separators with three or more categories, the tailings and concentrates are adjacent sorting grades in the corresponding sorting results, meaning there is a sorting grade that is better than the concentrate or worse than the tailings.

[0022] Step S120: For each benchmark feature value, determine the corresponding estimated grade value of the ore according to the grade estimation model. The grade estimation model can determine the corresponding estimated grade value based on the input benchmark feature value using a preset function formula. The preset function formula reflects a fitting curve that fits multiple existing feature values ​​and existing grade values, enabling the grade estimation model to determine the corresponding estimated grade value based on the benchmark feature value using the fitting curve. In other embodiments, the grade estimation model can be an existing neural network model trained based on multiple existing feature values ​​and existing grade values, thereby improving the prediction accuracy of the estimated grade value.

[0023] Step S130: Determine the current predicted average grade based on multiple estimated grade values. The current predicted average grade can be the average of multiple estimated grade values, thus effectively reflecting the average grade of multiple sorted ores in the most recent sampling window.

[0024] Step S140: Determine the current grade deviation value based on the current predicted average grade and the target average grade. The target average grade can be a preset value, and the preset value is determined according to the sorting requirements of the sorting machine. The current grade deviation value reflects the deviation of the current predicted average grade from the target average grade, thereby quantifying and refining the sorting deviation of the sorting machine.

[0025] Step S150: Based on the PID control method, determine the current adjustment amount according to the current grade deviation value. That is, the current adjustment amount can be calculated using the PID (Proportional Integral Derivative) algorithm based on the current grade deviation value.

[0026] Step S160: Adjust the sorting control parameters of the sorting machine according to the current adjustment amount. These parameters include a feature threshold. In other words, the adjusted feature threshold is calculated by combining the original feature threshold with the current adjustment amount. The adjusted feature threshold can be the sum of the original feature threshold and the current adjustment amount.

[0027] According to the adaptive adjustment method for ore sorting provided in this embodiment, the overall grade of the most recently sorted batch of ore, i.e., the current predicted average grade, can be effectively determined based on the characteristic values ​​of multiple sorted ores corresponding to the most recent sampling window. Based on this, the current grade deviation value, reflecting the deviation between the current predicted average grade and the target average grade, is used in conjunction with a PID control method to obtain the corresponding current adjustment amount. This allows the characteristic threshold to be adjusted based on the current adjustment amount, enabling the sorting machine to bring the overall grade of subsequent batches of sorted ore closer to the target average grade. It should be noted that, due to the enrichment and dispersion of veins during mining, the grade of the collected raw ore and the ore to be sorted exhibits relatively continuous fluctuations across batches. Therefore, the current predicted average grade corresponding to the most recently sorted batch of ore can be effectively used as prediction data for the overall grade of the next batch of ore to be sorted, thereby achieving early suppression of grade fluctuations. Furthermore, by incorporating PID control, the sorting control parameters, such as the characteristic threshold, adjusted based on the current adjustment amount can be smoothly, accurately, and quickly adjusted. This allows for rapid response to and effective suppression of ore grade fluctuations, thus improving the sorting machine's response speed and enabling timely updates to the sorting control settings for the next batch of ore to be sorted. In addition, since the characteristic values ​​of the sorted ore can be acquired in real time, grade fluctuations can be detected and sorting control parameters adjusted accordingly. This avoids excessive manual intervention by operators, significantly improving automation and real-time adjustability, thereby further enhancing the overall sorting efficiency of the sorting machine. It should be noted that the PID control method generates corresponding current adjustment amounts for different levels of deviation, achieving more refined control of sorting control parameters such as the characteristic threshold, resulting in a more effective suppression of grade fluctuations.

[0028] In some embodiments, the current adjustment amount is determined based on a proportional term, an integral term, and a derivative term. The proportional term includes the proportional gain and the current grade deviation value, the integral term includes the integral gain and the integral of the current grade deviation value, and the derivative term includes the derivative gain and the derivative of the current grade deviation value. For example, the expression for calculating the current adjustment amount is: ΔP(t) = Kp e(t)+Ki e(t)dt+Kd de(t) / dt, where e(t) is the current grade deviation at time t, Kp is the proportional gain, Ki is the integral gain, and Kd is the differential gain. e(t)dt is the integral of the current grade deviation value at time t, and de(t) / dt is the derivative of the current grade deviation value at time t. Furthermore, the PID control method can be an incremental PID control method, thus ensuring that the current adjustment is an incremental value, thereby avoiding error accumulation. For example... Figure 2 As shown, step S150, before determining the current adjustment amount based on the current grade deviation value using the PID control method, may include steps S151 to S153.

[0029] Step S151: If the error zero-crossing frequency corresponding to the current grade deviation value within a preset time period is greater than a preset frequency threshold, then the proportional gain is decreased or the derivative gain is increased. The preset time period can correspond to multiple sampling windows, and the error zero-crossing frequency can be the frequency at which the current grade deviation value is 0 within the preset time period. Therefore, when the error zero-crossing frequency is greater than the preset frequency threshold, high-frequency oscillations exist in the sorted grade. Therefore, by decreasing the proportional gain or increasing the derivative gain, the high-frequency oscillations of the sorted grade can be effectively suppressed.

[0030] Step S152: If the absolute value of the current grade deviation is not less than the preset dead zone threshold and the duration is greater than the preset deviation duration, then increase the proportional gain or decrease the integral gain. When the absolute value of the current grade deviation is not less than the preset dead zone threshold and the duration is greater than the preset deviation duration, the sorting machine has the problem of slow response speed and static error. Increasing the proportional gain or decreasing the integral gain can effectively suppress the above problems.

[0031] Step S153: If the absolute value of the current grade deviation is greater than a preset threshold, the integral gain is set to 0. When the absolute value of the current grade deviation is greater than the preset threshold, the sorting grade of the sorting machine has a large deviation. Therefore, by setting the integral gain to 0, the integral term can be canceled, so that the current adjustment amount does not contain accumulated errors from the past period. Consequently, based on this current adjustment amount, the adjusted feature threshold can better reflect the influence of the current grade deviation, thus improving the rapid response of the sorting machine. It should be noted that in some embodiments, step S150 may include one or more of steps S151 to S153.

[0032] According to the adaptive adjustment method for ore sorting provided in this embodiment, in cases where some ores are mixed during mining and transportation, or where the agglomeration characteristics of the ore itself cause drastic fluctuations in the grade of some batches of ore, the method can identify these special cases based on the error zero-crossing frequency and the corresponding preset frequency threshold, and promptly adjust the relevant gain in the PID control method to effectively address these special cases and suppress the fluctuations in sorting grade caused by such situations. Regarding situations where the sorting machine's response speed is too slow or there is a steady-state error due to its own hardware limitations, the method can effectively identify these special cases based on the duration for which the absolute value of the current grade deviation is not less than a preset dead zone threshold and a preset deviation duration. By adjusting the corresponding gain parameters, the method can effectively alleviate or resolve these special cases and suppress their negative impact on the sorting grade. Furthermore, in cases where the deviation between the current predicted grade average and the target grade average is too large (i.e., the absolute value of the current grade deviation is greater than the preset threshold), canceling the integral term allows the current grade deviation based on the PID control method to disregard the accumulated error over a period of time, thereby making the current grade deviation more aggressive. This enables the sorting machine based on the adjusted characteristic threshold to have a faster response and suppress drastic grade deviations.

[0033] In some embodiments, such as Figure 3 As shown, step S160, adjusting the sorting control parameters of the sorting machine according to the current adjustment amount, may include steps S161 to S163.

[0034] Step S161: Obtain feedforward feature values ​​for multiple ores to be sorted. Multiple ores to be sorted can correspond to the most recent sampling window; that is, multiple ores to be sorted are the most recent batch of ores to be sorted.

[0035] Step S162: Determine the feedforward adjustment amount based on the feedforward eigenvalues. The method for determining the feedforward adjustment amount based on the feedforward eigenvalues ​​can refer to the method for determining the current adjustment amount using the reference eigenvalues.

[0036] Step S163: Adjust the sorting control parameters based on the current adjustment amount and the feedforward adjustment amount. The adjusted feature threshold can be the sum of the current adjustment amount, the feedforward adjustment amount, and the feature threshold before adjustment.

[0037] According to the adaptive adjustment method for ore sorting provided in this embodiment, the current adjustment amount is determined based on the characteristic values ​​of the sorted ore, thereby performing a lag correction on the grade sorting deviation of the current sorter and realizing the effective utilization of historical sorted data. The feedforward adjustment amount is determined based on the feedforward characteristic values ​​of the ore before sorting, thereby correcting for possible grade fluctuations in the current batch of ore to be sorted in advance and realizing the effective utilization of existing data on the ore to be sorted. In other words, the feedforward adjustment amount can respond to sudden grade changes in the ore to be sorted, and the current adjustment amount can be used to eliminate steady-state errors. Therefore, the adjusted characteristic threshold obtained by combining the current adjustment amount and the feedforward adjustment amount can simultaneously correspond to grade changes and steady-state errors, enabling the sorter to more effectively suppress grade fluctuations. Furthermore, since the feedforward adjustment is applied to the ore to be sorted, it avoids relying solely on the current adjustment for the sorted ore. This shortens the response time to sudden changes in ore grade and further solves the problem of fluctuations in product quality caused by the lag in the current adjustment. This meets the need for continuous production of the sorting machine without stopping or to significantly reduce the sorting cycle time, thus ensuring overall sorting efficiency.

[0038] In some embodiments, the feedforward eigenvalues ​​include eigenvalues. For example... Figure 4 As shown, step S162, which determines the feedforward adjustment amount based on the feedforward feature value, may include steps S1621 to S1622.

[0039] Step S1621: Determine the average feedforward grade based on multiple feedforward feature values. Referring to the method of obtaining the current predicted average grade using the benchmark feature values ​​described above, the average feedforward grade corresponding to multiple ores to be sorted before sorting can be determined based on multiple feedforward feature values.

[0040] Step S1622: Determine the feedforward adjustment amount based on the average feedforward grade and the average target grade. According to the offline mapping table, combined with the average feedforward grade and the average target grade, the predicted feedforward grade value can be determined. The mapping table can include multiple mapping groups, each including existing corresponding average feedforward grades, characteristic thresholds, and concentrate grades. Therefore, by looking up the table or using linear interpolation, the corresponding concentrate grade, i.e., the predicted feedforward grade value, can be determined based on the current average feedforward grade and the current characteristic threshold. The difference between the average target grade and the predicted feedforward grade value is taken as the feedforward grade deviation. The feedforward adjustment amount is determined based on the feedforward grade deviation. It is understandable that the predicted grade of each ore in the ore stream may deviate from the actual grade after the separator performs separation on the ore stream due to factors such as the separator's hardware parameters and software control parameters; that is, the predicted grade and the actual grade are not completely consistent. Therefore, setting a mapping table can effectively express the above deviation, thereby achieving accurate calculation of the feedforward adjustment amount. For example, for a batch of ore, if the average feedforward grade is determined to be 28% and the current feature threshold is 30, then by further querying the mapping table, the feedforward predicted grade value corresponding to the aforementioned average feedforward grade and feature threshold can be determined to be 27.5%. Therefore, for a target average grade of 25%, the feedforward grade deviation can be determined to be 25% - 27.5% = -2.5%. Then, according to the feedforward grade deviation and the preset feedforward control rules, for example, for every negative 1% deviation, the feature threshold is increased by 0.5, so the current adjustment is -2.5. (-0.5) = 1.25. Furthermore, in the subsequent step S163, the sum of the feedforward adjustment and the feature threshold before adjustment is 30 + 1.25 = 31.25.

[0041] The adaptive adjustment method for ore sorting provided in this embodiment can further utilize characteristic values ​​as feedforward characteristic values, thereby eliminating the need for additional data acquisition equipment in the sorting machine and reducing the hardware cost of the sorting machine. Based on this, an effective feedforward adjustment amount can be generated, so that the most recently sorted batch of ore, when sorted according to the adjusted characteristic threshold, can further approach the target grade average, thereby effectively suppressing grade fluctuations.

[0042] In some embodiments, the feedforward feature values ​​include one or more of the following: average transmission grayscale values ​​corresponding to multiple pre-sorted ores, a surface color distribution histogram for each pre-sorted ore, and surface texture roughness for each pre-sorted ore. The average transmission grayscale value is the mean of the transmission grayscale values ​​corresponding to each pre-sorted ore, which can be obtained using a low-resolution X-ray transmission detector. The surface color distribution histogram and surface texture roughness can both be obtained from image data of the ores using an industrial RGB color camera. Figure 5As shown, step S162, which determines the feedforward adjustment amount based on the feedforward feature value, may include steps S1623 to S1624.

[0043] Step S1623: Based on the feedforward feature values, determine the predicted grade trend changes for multiple ores to be sorted. Inputting the feature data into a simplified prediction model yields the corresponding predicted grade trend changes. The simplified prediction model can be a neural network model trained on existing data, where the existing data includes existing feature data and corresponding actual grade trend changes for multiple ore samples. In other embodiments, the simplified prediction model may include existing data, thereby determining the predicted grade trend changes based on linear interpolation or similar methods.

[0044] Step S1624: Determine the feedforward adjustment amount based on the predicted grade trend change. For example, when the proportion of gangue color in the surface color distribution histogram obtained from further processing of image data acquired by an industrial RGB color camera increases, the predicted grade trend change corresponds to a decrease in the raw ore grade. Based on this, the feedforward adjustment amount is used to raise the feature threshold to improve the concentrate's acceptance criteria.

[0045] The adaptive adjustment method for ore sorting provided in this embodiment takes into account the low precision of data acquisition hardware in ore sorting. Instead of specifically acquiring feature values ​​for individual ores, it acquires feature data (feedforward feature values) of the ore stream formed by multiple ores before sorting, thereby estimating the average grade trend of this batch of ores—whether the ore grade is decreasing or increasing. In other words, predicting grade trend changes is effectively applicable to scenarios with low-performance data acquisition hardware in the sorting machine and can solve the problem of response lag in adjusting the feature threshold of the current adjustment amount. Therefore, while reducing the hardware cost of the sorting machine, it can further effectively suppress grade fluctuations by shortening the response time using the basic adjustment amount, thus achieving stability in the sorted grade.

[0046] In some embodiments, the grade estimation model includes characteristic values ​​and true grade values ​​corresponding to multiple ore samples. For example... Figure 6 As shown, step S120, for each benchmark characteristic value, determines the estimated grade value of the corresponding ore according to the grade estimation model, which may include step S121.

[0047] Step S121: For each benchmark feature value, based on linear interpolation, determine the estimated average grade of the corresponding ore according to the two feature values ​​adjacent to the benchmark feature value in the grade estimation model and the corresponding true grade values. In linear interpolation, the calculation expression for the estimated grade value corresponding to the benchmark feature value is: G est =G i +(G i+1 -Gi )×(x-KK i ) / (KK i+1 -KK i ), where x is the baseline eigenvalue, G est KK is the estimated grade value corresponding to the baseline characteristic value x. i and KK i+1 G represents the eigenvalue in the grade estimation model that is closest to, greater than or equal to, and less than or equal to, the benchmark eigenvalue x. i For the eigenvalue KK i The corresponding true grade value, G i+1 For KK i+1 The corresponding true grade values. The true grade values ​​corresponding to multiple ore samples can be obtained using existing detection methods, such as chemical analysis. The true grade values ​​corresponding to multiple ore samples can cover the grade range of the ore sorted by the separator; for example, for phosphate rock, the grade range of phosphorus pentoxide includes 5% to 35%. By sorting the characteristic values ​​of the above multiple ore samples according to their numerical values, adjacent characteristic values ​​in the above linear interpolation are determined.

[0048] According to the adaptive adjustment method for ore sorting provided in this embodiment, since the benchmark characteristic value and the true grade value are correlated (either positively or negatively), linear interpolation can not only obtain an estimated grade value close to the true grade value, but also reduce the calculation time. Furthermore, by using a grade estimation model that includes characteristic values ​​and true grade values ​​corresponding to multiple ore samples, the estimated grade value corresponding to each benchmark characteristic value can be effectively determined. Therefore, while reducing the total data processing time of the adaptive adjustment method for ore sorting, it provides an accurate data basis, namely the estimated grade value, thereby shortening the lag time for subsequent adjustments to the sorting control parameters, further stabilizing the sorting grade benchmark of the sorting machine, and reducing the impact of ore quality fluctuations on the grade of the sorted ore.

[0049] In some embodiments, such as Figure 7 As shown, step S130, determining the current predicted average grade based on multiple estimated grade values, may include steps S131 to S132.

[0050] Step S131: For each sorted ore, determine the estimated weight corresponding to the ore based on the difference between the ore sorting time and the sampling time, wherein the estimated weight increases as the difference decreases. The ore sorting time can be the time when the classification device determines the classification result corresponding to the ore. In some other embodiments, the ore sorting time can be the starting time when the ore is located in the corresponding containing space after being sorted by the sorting machine.

[0051] Step S132: Determine the current predicted average grade based on the estimated weights and estimated grade values ​​corresponding to multiple ores. The current predicted average grade can be the sum of the products of each estimated weight and its corresponding estimated grade value; that is, the current predicted average grade can be obtained by weighted averaging of the estimated grade values.

[0052] According to the adaptive adjustment method for ore sorting provided in this embodiment, by assigning a greater estimation weight to ores sorted closer to the current time, the influence of the estimated grade value of sorted ores closer to the current time on the current predicted average grade can be improved. Specifically, the sorting machine is often in a continuous working state, and multiple ores to be sorted input into the sorting machine at the same time often come from the same mining location or adjacent mining locations. Ores sorted within adjacent time periods often have similar grades, i.e., the grade fluctuation is small. Therefore, the above settings fully consider the characteristics of ore mining and target mineral growth, so that when the true grade value of the mineral to be sorted cannot be obtained, the sorting grade of the batch of ore closest to the current time can be accurately calculated and quantified using the data of the sorted ores. In addition, the estimated grade value corresponding to the ore sorting time far from the current time is retained, which affects the current predicted average grade. Therefore, the current predicted average grade can reflect the grade trend of the current ore more quickly, while filtering out the occasional sharp fluctuations or random errors of individual ore pieces, thereby improving the accuracy and reliability of the current predicted average grade.

[0053] In some embodiments, such as Figure 8 As shown, step S160, adjusting the sorting control parameters of the sorting machine according to the current adjustment amount, may include steps S164 to S167.

[0054] Step S164: Obtain the average historical true grade corresponding to multiple historical sorting windows, where each historical sorting window includes multiple sorted ores. The historical sorting windows can be time windows or quantity windows, similar to the sampling windows mentioned earlier, and will not be elaborated further here. Furthermore, the window size of the historical sorting windows can be the same as or different from the window size of the sampling windows. The multiple ores corresponding to multiple historical sorting windows can be multiple sorted ores from before the ore to be sorted at the current moment. The method for obtaining the average historical true grade can refer to the method for obtaining the average current predicted grade mentioned earlier. That is, determine the historical grade value of the corresponding ore based on the characteristic values ​​of each ore in each historical sorting window, and determine the average historical true grade corresponding to that historical sorting window based on all historical grade values. Further, for each historical sorting window, determine the corresponding weight based on the time interval between the sorting time of each corresponding ore and the current moment, and calculate the average grade of a single window by weighted averaging based on each historical grade value and its corresponding weight. The weights can be set according to the size of the corresponding time interval; that is, the smaller the time interval, the greater the weight, which can increase the influence of ore closer to the current moment on the historical true grade average. In other embodiments, the historical true grade average can be the average of all historical grade values.

[0055] Step S165: Determine the predicted grade average corresponding to the current sorting window based on multiple historical average grades. This can be done using least-squares linear fitting or exponential smoothing. When using least-squares linear fitting, a univariate linear regression can be used to fit a straight line based on multiple historical average grades, thereby determining the grade average of the next historical sorting window (i.e., the current batch of ore to be sorted) using the current multiple historical sorting windows. The expression for the univariate linear regression fitting line is y = a + b. t, where, a=(Σy_t- bΣt) / N, b=[N Σ(t y_t)-Σt Σy_t] / [N Σt²-(Σt)²], where t is the sequence number of the historical sorting window and N is the total number of historical sorting windows. Then t=N+1 corresponds to the historical sorting window corresponding to the current batch of ore to be sorted, i.e., the advanced predicted average grade y= a+b (N+1), where y_t is the historical true grade mean corresponding to the t-th historical sorting window. Exponential smoothing is an existing method and will not be elaborated here.

[0056] Step S166: Determine the advance adjustment amount based on the average predicted grade and the average target grade. The advance adjustment amount can be calculated by combining the difference between the average target grade and the average predicted grade with the PID control method. For details, please refer to the calculation method of the current adjustment amount in the previous text, which will not be repeated here.

[0057] Step S167: Adjust the sorting control parameters of the sorting machine based on the current adjustment amount and the advance adjustment amount. The adjusted feature threshold can be the sum of the feature threshold before adjustment, the current adjustment amount, and the advance adjustment amount.

[0058] According to the adaptive adjustment method for ore sorting provided in this embodiment, the total adjustment amount of the feature threshold is determined based on the possible average grade of multiple ores to be sorted in the current batch and the average grade of multiple ores sorted in the most recent batch. This allows for advance compensation for trend changes by utilizing the adjustment amount corresponding to the advanced predicted average grade, reducing the impact of adjustment lag caused by relying solely on the adjustment amount corresponding to the current predicted average grade. Specifically, based on the feature threshold adjustment according to the deviation between the current predicted average grade and the target average grade of multiple sorted ores corresponding to the sampling window closest to the current time, the method further predicts the average grade corresponding to the current window (i.e., the advanced predicted average grade) based on multiple historical sorting windows closest to the current time, and further adjusts the feature threshold based on the deviation between the advanced predicted average grade and the target average grade. Therefore, the adaptive adjustment method for ore sorting not only considers the possibility of sudden changes in ore grade but also the characteristic of a certain degree of grade continuity between adjacent batches of ore input into the sorting machine, thus achieving more accurate and reliable grade fluctuation suppression.

[0059] In some embodiments, the sorting machine further includes a sorting device that sorts the ore according to the classification results, and the sorting control parameters also include the sorting frequency. Figure 9 As shown, step S160, adjusting the sorting control parameters of the sorting machine according to the current adjustment amount, may include steps S168 to S169.

[0060] Step S168: Determine the adjusted feature threshold based on the current adjustment amount. The feature threshold can be the boundary value between the feature values ​​corresponding to tailings and concentrate. Therefore, by adjusting the feature threshold, the strictness of the sorting device's separation of concentrate can be changed to control the concentrate grade and tailings grade. For example, increasing the feature threshold can reduce the subsequent concentrate grade.

[0061] Step S169: Determine the adjusted sorting frequency based on the adjusted feature threshold. The sorting frequency can be the frequency at which the sorting device performs sorting operations on the ore corresponding to the concentrate, or the frequency at which the sorting device performs sorting operations on the ore corresponding to the tailings. Since the sorting device does not operate on the ore corresponding to the concentrate or tailings, the corresponding type of ore can be naturally transported to the corresponding receiving area. Therefore, after adjusting the feature threshold, the frequency of the sorting operation of the sorting device needs to be changed accordingly to meet the adjusted sorting cycle time. For example, when the sorting device performs sorting operations on the ore corresponding to the tailings, and the feature threshold is increased (i.e., the intake standard of the concentrate is increased) to improve the subsequent concentrate grade, more ore to be sorted is judged as tailings and needs to be sorted by the sorting device. Therefore, it is necessary to increase the sorting frequency so that more tailings-corresponding ore can be accurately sorted, avoiding the problem that the sorting device cannot process the tailings-corresponding ore in time and causes it to fall into the receiving area corresponding to the concentrate, thus leading to a decrease in the concentrate grade. Furthermore, the sorting frequency can be adjusted based on the execution time of the sorting operation and the interval between sorting operations. For example, the sorting frequency can be increased by shortening the execution time of the sorting operation and the interval between sorting operations.

[0062] According to the adaptive adjustment method for ore sorting provided in this embodiment, the characteristic threshold can be adjusted by the current adjustment amount to make the grade of the sorted ore close to the target average grade. Furthermore, the impact of adjusting the characteristic threshold on corresponding hardware devices in the sorting machine, such as the sorting unit, is considered. Therefore, the sorting frequency is further adjusted based on the adjusted characteristic threshold, so that the adjusted sorting control parameters enable the various hardware devices in the sorting machine to work collaboratively, reducing the negative impact of adjusting only the characteristic threshold on the overall sorting efficiency and effect of the sorting machine. In other words, while adjusting the subsequent average grade, the overall sorting efficiency and accuracy of the sorting machine are ensured.

[0063] Based on the same inventive concept, this disclosure also provides a sorting machine, which may include: a data acquisition device, a sorting device, and an adjustment device.

[0064] A data acquisition device is configured to acquire characteristic values ​​of an ore. The data acquisition device may include a dual-energy X-ray transducer to irradiate the ore with X-rays and acquire corresponding data to generate characteristic values ​​of the ore, such as KK values ​​or equivalent atomic numbers.

[0065] A classification device is configured to determine the classification result of ore based on characteristic values ​​and characteristic thresholds, wherein the classification result includes tailings and concentrate. The classification device can communicate with a data acquisition device to receive characteristic values ​​and determine the classification result based on the characteristic values ​​and characteristic thresholds.

[0066] An adjustment device is used to adjust the sorting control parameters of the sorting machine according to the ore sorting adaptive adjustment method provided in any of the foregoing embodiments, wherein the sorting control parameters include characteristic thresholds. The adjustment device can be communicatively connected to the classification device, thereby transmitting the adjusted sorting control parameters to the classification device so that the classification device can promptly determine the classification result of subsequent ores based on the adjusted characteristic thresholds, thereby improving the response speed of grade fluctuation suppression.

[0067] This application uses specific terms to describe embodiments of the application. Terms such as "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic associated with at least one embodiment of the application. Therefore, it should be emphasized and noted that references to "an embodiment," "one embodiment," or "an alternative embodiment" in different locations throughout this specification do not necessarily refer to the same embodiment. Furthermore, certain features, structures, or characteristics in one or more embodiments of the application can be appropriately combined.

[0068] In the context of this application, unless the context clearly indicates otherwise, the words "a," "an," "an," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0069] Similarly, it should be noted that, in order to simplify the description of the present application and thus aid in the understanding of one or more embodiments, the foregoing description of the embodiments of the present application sometimes combines multiple features into a single embodiment, drawing, or description thereof. However, this disclosure method does not imply that the subject matter of the present application requires more features than those mentioned in the claims. In fact, the embodiments contain fewer features than all the features of the single embodiments disclosed above.

[0070] The basic concepts have been described above. Obviously, for those skilled in the art, the above disclosure is merely illustrative and does not constitute a limitation of this application. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this application. Such modifications, improvements, and corrections are suggested in this application, and therefore remain within the spirit and scope of the embodiments of this application.

Claims

1. An adaptive adjustment method for ore sorting, characterized in that, An adaptive adjustment method for ore sorting is applied to a sorting machine, which includes a data acquisition device and a classification device. The data acquisition device is configured to acquire feature values ​​of the ore, and the classification device is configured to determine the classification result of the ore based on the feature values ​​and a feature threshold. The classification result includes tailings and concentrate. The characteristic values ​​of the sorted ores are obtained, and the characteristic values ​​are used as the reference characteristic values; For each of the aforementioned benchmark characteristic values, the estimated grade value of the corresponding ore is determined according to the grade estimation model; Based on the multiple estimated grade values, determine the current predicted average grade; Based on the current predicted average grade and the target average grade, determine the current grade deviation value; Based on the PID control method, the current adjustment amount is determined according to the current grade deviation value; Based on the current adjustment amount, the sorting control parameters of the sorting machine are adjusted, wherein the sorting control parameters include the feature threshold.

2. The adaptive adjustment method for ore sorting as described in claim 1, characterized in that, The current adjustment amount is determined based on a proportional term, an integral term, and a derivative term. The proportional term includes the proportional gain and the current grade deviation value. The integral term includes the integral gain and the integral of the current grade deviation value. The derivative term includes the derivative gain and the derivative of the current grade deviation value. Before determining the current adjustment amount based on the current grade deviation value using the PID control method, the process includes: If the error zero-crossing frequency corresponding to the current grade deviation value within a preset time period is greater than a preset frequency threshold, then the proportional gain is reduced or the differential gain is increased. And / or, If the absolute value of the current grade deviation is not less than the preset dead zone threshold and the duration is greater than the preset deviation duration, then increase the proportional gain or decrease the integral gain; and / or, If the absolute value of the current grade deviation is greater than a preset threshold, the integral gain is set to 0.

3. The adaptive adjustment method for ore sorting as described in claim 1, characterized in that, The step of adjusting the sorting control parameters of the sorting machine according to the current adjustment amount includes: Obtain feedforward feature values ​​of multiple ores to be sorted; Based on the aforementioned feedforward characteristic values, the feedforward adjustment amount is determined; The sorting control parameters are adjusted based on the current adjustment amount and the feedforward adjustment amount.

4. The adaptive adjustment method for ore sorting as described in claim 3, characterized in that, The feedforward eigenvalues ​​include the eigenvalues; The step of determining the feedforward adjustment amount based on the feedforward feature value includes: The average feedforward grade is determined based on multiple feedforward characteristic values. The feedforward adjustment amount is determined based on the average feedforward grade and the average target grade.

5. The adaptive adjustment method for ore sorting as described in claim 3, characterized in that, The feedforward feature values ​​include one or more of the following: average transmission grayscale values ​​corresponding to multiple ores before sorting, surface color distribution histograms of each ore before sorting, and surface texture roughness of each ore before sorting; determining the feedforward adjustment amount based on the feedforward feature values ​​includes: Based on the feedforward feature value, determine the predicted grade trend change for multiple ores to be sorted; The feedforward adjustment amount is determined based on the predicted grade trend change.

6. The adaptive adjustment method for ore sorting as described in claim 1, characterized in that, The step of adjusting the sorting control parameters of the sorting machine according to the current adjustment amount includes: Obtain the average historical true grade corresponding to multiple historical sorting windows, wherein each historical sorting window includes multiple sorted ores; The mean of the advanced predicted grade corresponding to the current sorting window is determined based on multiple historical true grade averages. The advance adjustment amount is determined based on the average predicted grade and the average target grade. The sorting control parameters of the sorting machine are adjusted based on the current adjustment amount and the advance adjustment amount.

7. The adaptive adjustment method for ore sorting as described in claim 1, characterized in that, The grade estimation model includes the characteristic values ​​and true grade values ​​corresponding to multiple ore samples. For each benchmark characteristic value, determining the estimated grade value of the corresponding ore based on the grade estimation model includes: For each of the benchmark feature values, based on linear interpolation, the estimated average grade of the corresponding ore is determined according to the two feature values ​​adjacent to the benchmark feature value in the grade estimation model and the corresponding true grade value.

8. The adaptive adjustment method for ore sorting as described in claim 1, characterized in that, The step of determining the current predicted average grade based on multiple estimated grade values ​​includes: For each sorted ore, the estimated weight corresponding to the ore is determined based on the difference between the sorting time and the sampling time, wherein the estimated weight increases as the difference decreases; The current predicted average grade is determined based on the estimated weights and estimated grade values ​​corresponding to the multiple ores.

9. The adaptive adjustment method for ore sorting as described in claim 1, characterized in that, The sorting machine further includes a sorting device, which sorts the ore according to the classification result. The sorting control parameters also include a sorting frequency. Adjusting the sorting control parameters of the sorting machine according to the current adjustment amount includes: Based on the current adjustment amount, determine the adjusted feature threshold; The adjusted sorting frequency is determined based on the adjusted feature threshold.

10. A sorting machine, characterized in that, include: A data acquisition device is configured to acquire characteristic values ​​of the ore; A classification device is configured to determine a classification result of the ore based on the feature value and the feature threshold, wherein the classification result includes tailings and concentrate; An adjustment device is used to adjust the sorting control parameters of the sorting machine according to the ore sorting adaptive adjustment method as described in any one of claims 1 to 9, wherein the sorting control parameters include the characteristic threshold.

Images