A method, system and medium for evaluating ore separation performance

Abstract

The present invention discloses an ore sorting performance evaluation method, system and medium. The ore sorting performance evaluation method of the present invention includes obtaining collected ore sample data; using the ore sorting algorithm to be evaluated to predict the predicted grade of the ore sample, calculating the recovery rate and the waste rate under different grade thresholds, and constructing a "recovery-waste rate" curve based on the predicted grade; calculating the curve enclosed area of ​​the "recovery-waste rate" curve based on the predicted grade and the area ratio of the curve enclosed area of ​​the preset "recovery-waste rate" curve based on the assay grade as the evaluation result of the ore sorting algorithm to be evaluated. The present invention aims to quantify the actual performance and ideal performance of the ore sorting algorithm by constructing the "recovery-waste rate" curve, and provide a comprehensive ore sorting algorithm evaluation result to guide users to select a suitable ore sorting algorithm.

Description

Technical Field

[0001] The present invention relates to the technical field of ore sorting, and in particular to an ore sorting performance evaluation method, system and medium, which can be used for algorithm evaluation and optimization in the fields of ore pre-disposal, mineral enrichment, etc. Background Art

[0002] With the development of mining, the efficient utilization of ore resources has become increasingly important. Ore sorting technology, as an important part of ore pretreatment, is of great significance to improving ore grade and reducing ore dressing costs. Ore sorting technology based on X-ray transmission imaging (XRT) has been widely used due to its advantages such as high efficiency and non-destructiveness. In the ore sorting process, the performance of the sorting algorithm directly affects the sorting effect. Common sorting algorithms include traditional statistical models and advanced machine learning models. However, the evaluation of these algorithms usually relies on proxy indicators, such as binary classification accuracy, or a single recovery rate and scrap rate, which cannot fully reflect the performance of the algorithm under different conditions. In addition, the lack of knowledge of the upper limit of sorting cannot guide users to choose the optimal sorting strategy based on ore characteristics and production needs. Therefore, there is an urgent need for a method that can comprehensively evaluate the performance of ore sorting, which can not only quantify the overall performance of the ore sorting algorithm, but also help users understand the sorting potential of the target ore and guide the formulation and optimization of the sorting strategy. In the existing ore sorting technology, there are many deficiencies in the way to evaluate the performance of the sorting algorithm. Traditional evaluation methods usually rely on proxy indicators or single indicators, which are difficult to fully reflect the performance of sorting algorithms under different conditions, especially when the production environment requirements change or the ore properties are complex. Proxy indicators or single indicators cannot provide users with enough information to optimize sorting strategies. In addition, the existing technology lacks the understanding of the upper limit of sorting, and users cannot know the best combination of recovery rate and waste rate that can be achieved under ideal conditions, making it difficult to make effective production decisions. The limitations of proxy indicators are very obvious. For example, sorting algorithm A uses the mean square error (MSE) indicator, while sorting algorithm B uses binary accuracy. At this time, the performance of the algorithms cannot be directly compared. In addition, even if algorithm C also uses binary accuracy as a performance evaluation indicator, the different definitions of binary classification between algorithms B and C will also cause the indicators to be unable to be directly compared. Single indicators, such as the commonly used "recovery rate-waste rate" combination, also have their limitations. This is mainly because the higher the recovery rate and waste rate, the better, but they are often mutually exclusive (throwing more ore will increase the waste rate, but at the same time reduce the recovery rate). For example, the recovery rate and scrap rate of sorting algorithm A are 70% and 60%, while those of sorting algorithm B are 90% and 30%. It is difficult to conduct a comprehensive evaluation of the two algorithms based on this single indicator. Summary of the invention

[0003] Technical problem to be solved by the present invention: In view of the above-mentioned problems of the prior art, a method, system and medium for evaluating the performance of ore sorting are provided. The present invention aims to quantify the actual performance and ideal performance of the ore sorting algorithm by constructing a "recovery rate-waste rate" curve, and provide comprehensive ore sorting algorithm evaluation results to guide users to select appropriate ore sorting algorithms.

[0004] In order to solve the above technical problems, the technical solution adopted by the present invention is:

[0005] A method for evaluating ore separation performance comprises the following steps:

[0006] S1, obtain the collected ore sample data;

[0007] S2, using the ore sorting algorithm to be evaluated to predict the predicted grade of the ore sample, calculate the recovery rate and discard rate under different grade thresholds, and construct a "recovery rate-discard rate" curve based on the predicted grade;

[0008] S3, calculating the ratio of the area enclosed by the curve of the "recovery rate-waste rate" curve based on the predicted grade to the area enclosed by the curve of the preset "recovery rate-waste rate" curve based on the assay grade as the evaluation result of the ore sorting algorithm to be evaluated.

[0009] Optionally, the X-axis of the "recovery rate-waste rate" curve is the waste rate and the Y-axis is the recovery rate, or the X-axis of the "recovery rate-waste rate" curve is the recovery rate and the Y-axis is the waste rate.

[0010] Optionally, step S1 comprises: randomly extracting from the target ore a plurality of components, grades and forms of The bulk ore samples form a sample collection:

[0011] ,

[0012] In the above formula, For the sample collection, ~ They are No. 1~ lump ore sample; sample collection Each ore sample in Get their test grades respectively and weight Form the grade vector and weight vector:

[0013] ,

[0014] ,

[0015] In the above formula, is the taste vector, ~ They are No. 1~ Assay grade of lump ore samples, is the weight vector, ~ They are No. 1~ The weight of the lump ore sample, Indicates transpose.

[0016] Optionally, the ore sorting algorithm to be evaluated is an ore sorting algorithm that predicts the grade of ore samples based on single-energy XRT images and performs ore sorting, and after step S1 and before step S2, it also includes for each ore sample Get its characterization data , the characterization data Refers to ore samples The single-energy XRT image is obtained to obtain the characterization vector :

[0017] ,

[0018] In the above formula, ~ They are No. 1~ Characterization data of lump ore samples.

[0019] Optionally, step S2 includes:

[0020] S2.1, adjust the hyperparameters of the ore sorting algorithm, obtain the predicted grade of the ore sample predicted by the ore sorting algorithm to be evaluated under different hyperparameter combinations, and calculate the recovery rate and discard rate under different grade thresholds :

[0021] ,

[0022] In the above formula, For the The rejection rate under the hyperparameter combination, For the The recovery rate under the hyperparameter combination, is the number of hyperparameter combinations;

[0023] S2.2, using the Pareto front method from the set The non-dominated points of hyperparameter combinations whose recovery rate and discard rate are lower than those of other hyperparameter combination data points are screened and filtered out, and the non-dominated points of hyperparameter combinations are screened and filtered out according to the set Plot a recovery-waste curve based on predicted grade.

[0024] Optionally, the calculation function expression of the area ratio in step S3 is:

[0025] ,

[0026] In the above formula, Indicates the area ratio, The area enclosed by the preset “recovery rate-waste rate” curve based on the assay grade. is the area enclosed by the “recovery-waste rate” curve based on the predicted grade, and:

[0027] ,

[0028] ,

[0029] In the above formula, is the number of data points in the recovery-rejection curve based on the assay grade, is the number of data points in the recovery-waste curve based on the predicted grade, and are the recoveries of the nth and n+1th data points in the preset “recovery rate-waste rate” curve based on the assay grade, and are the discard rates of the nth and n+1th data points in the preset “recovery rate-discard rate” curve based on the assay grade, respectively. and are the recovery rates of the mth and m+1th data points of the “recovery rate-waste rate” curve based on the predicted grade, and They are the rejection rates of the mth and m+1th data points of the “recovery rate-rejection rate” curve based on the predicted grade, respectively.

[0030] Optionally, before step S3, the method further includes obtaining a preset "recovery rate-waste rate" curve based on the assay grade:

[0031] S101, the taste vector Sort in ascending order to get the sorted ore sample index set:

[0032] ,

[0033] In the above formula, ~ After sorting The index of the lump ore sample that satisfies ,in ~ They are ranked 1st to The grade of the bulk ore sample;

[0034] S102, for each grade threshold Calculate the corresponding scrap rate according to the following formula and recovery rate :

[0035] ,

[0036] ,

[0037] In the above formula, After sorting The serial number of the lump ore sample, After sorting The weight of the lump ore sample, After sorting The grade of the bulk ore sample;

[0038] S103, all grade thresholds Corresponding scrap rate and recovery rate As data points to form a data set :

[0039] ,

[0040] S104, according to the collection A "recovery rate-waste rate" curve based on the assay grade is drawn.

[0041] In addition, the present invention also provides an ore sorting performance evaluation system, comprising a microprocessor and a memory connected to each other, wherein the microprocessor is programmed or configured to execute the ore sorting performance evaluation method.

[0042] In addition, the present invention also provides a computer-readable storage medium, in which a computer program or instruction is stored. The computer program or instruction is programmed or configured to execute the ore sorting performance evaluation method through a processor.

[0043] In addition, the present invention also provides a computer program product, comprising a computer program or instructions, which are programmed or configured to execute the ore sorting performance evaluation method through a processor.

[0044] Compared with the prior art, the present invention mainly has the following advantages: the present invention includes acquiring collected ore sample data; using the ore sorting algorithm to be evaluated to predict the predicted grade of the ore sample, calculating the recovery rate and the waste rate under different grade thresholds, and constructing a "recovery rate-waste rate" curve based on the predicted grade; calculating the area ratio of the curve enclosed area of ​​the "recovery rate-waste rate" curve based on the predicted grade and the area ratio of the curve enclosed area of ​​the preset "recovery rate-waste rate" curve based on the assay grade as the evaluation result of the ore sorting algorithm to be evaluated. The present invention comprehensively evaluates the ideal sorting performance of the ore and the actual performance of the sorting algorithm by constructing the "recovery rate-waste rate" curve. The present invention can help users determine whether the ore is suitable for sorting, select a suitable sorting threshold, and quantify the comprehensive performance of different ore sorting algorithms. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 Schematic diagram of the basic flow of the method of the embodiment of the present invention.

[0046] Figure 2 Detailed flowchart of the method of the embodiment of the present invention.

[0047] Figure 3 Schematic diagram for comparing two “recovery rate-waste rate” curves in an embodiment of the present invention.

[0048] Figure 4 Schematic diagram of the Pareto front in an embodiment of the present invention.

[0049] Figure 5 Schematic diagram of comparison of “recovery rate-waste rate” curves of different algorithms in an embodiment of the present invention.

[0050] Figure 6 This is a schematic diagram of "recovery rate-waste rate" for guiding on-site selection in the embodiments of the present invention.

[0051] Figure 7 This is a schematic diagram of a batch of ores that are considered unsuitable for separation through an evaluation method in an embodiment of the present invention. DETAILED DESCRIPTION

[0052] The present invention is further described in detail below in conjunction with specific processes, drawings and embodiments.

[0053] like Figure 1 and Figure 2 As shown, this embodiment provides a method for evaluating ore separation performance, comprising the following steps:

[0054] S1, obtain the collected ore sample data;

[0055] S2, using the ore sorting algorithm to be evaluated to predict the predicted grade of the ore sample, calculate the recovery rate and discard rate under different grade thresholds, and construct a "recovery rate-discard rate" curve based on the predicted grade;

[0056] S3, calculating the ratio of the area enclosed by the curve of the "recovery rate-waste rate" curve based on the predicted grade to the area enclosed by the curve of the preset "recovery rate-waste rate" curve based on the assay grade as the evaluation result of the ore sorting algorithm to be evaluated.

[0057] The ore sorting performance evaluation method of this embodiment can use the preset "recovery rate-waste rate" curve based on the assay grade to help users determine whether the ore is suitable for sorting and the expected maximum sorting efficiency; use the "recovery rate-waste rate" curve based on the predicted grade to guide users to select a suitable combination of recovery rate and waste rate to optimize the ore sorting algorithm. Wherein, the X-axis of the "recovery rate-waste rate" curve is the waste rate and the Y-axis is the recovery rate, or the X-axis of the "recovery rate-waste rate" curve is the recovery rate and the Y-axis is the waste rate, the waste rate refers to the ratio of waste rock discarded in ore sorting, that is, the ratio of the total weight of ore identified as waste ore to the total weight of all ores, and the recovery rate refers to the ratio of metal minerals recovered in ore sorting, that is, the ratio of the total weight of target minerals in ore identified as concentrate to the total weight of target minerals in all ores. Figure 3 : is a comparative schematic diagram of two kinds of "recovery rate-waste rate" curves in this embodiment, wherein "based on assay grade" represents the "recovery rate-waste rate" curve based on assay grade, and "based on XRT algorithm predicted grade" represents the "recovery rate-waste rate" curve based on predicted grade. Figure 3 The area enclosed by the "recovery rate-waste rate" curve based on the predicted grade and the area enclosed by the "recovery rate-waste rate" curve based on the preset assay grade account for 85.01%.

[0058] In this embodiment, step S1 includes: randomly extracting minerals covering various compositions, grades and forms from the target ore. The bulk ore samples form a sample collection:

[0059] ,

[0060] In the above formula, For the sample collection, ~ They are No. 1~ lump ore sample; sample collection Each ore sample in Get their test grades respectively and weight Form the grade vector and weight vector:

[0061] ,

[0062] ,

[0063] In the above formula, is the taste vector, ~ They are No. 1~ Assay grade of lump ore samples, is the weight vector, ~ They are No. 1~ The weight of the lump ore sample, Indicates transpose.

[0064] In this embodiment, the ore sorting algorithm to be evaluated is an ore sorting algorithm (XRT algorithm) that predicts the grade of ore samples based on single-energy XRT images and performs ore sorting. It should be noted that the ore sorting algorithm is the evaluation object in this embodiment. The construction of the ore sorting algorithm is not the innovation of this embodiment, so the introduction of the details of the algorithm is omitted. After step S1 and before step S2, it also includes for each ore sample Get its characterization data , the characterization data Refers to ore samples The single-energy XRT image is obtained to obtain the characterization vector :

[0065] ,

[0066] In the above formula, ~ They are No. 1~ Characterization data of lump ore samples.

[0067] like Figure 2 As shown, step S2 of this embodiment includes:

[0068] S2.1, adjust the hyperparameters of the ore sorting algorithm (e.g., by grid search, etc.), obtain the predicted grade of the ore sample predicted by the ore sorting algorithm to be evaluated under different hyperparameter combinations, and calculate the recovery rate and discard rate under different grade thresholds :

[0069] ,

[0070] In the above formula, For the The rejection rate under the hyperparameter combination, For the The recovery rate under the hyperparameter combination, is the number of hyperparameter combinations;

[0071] S2.2, using the Pareto front method from the set The non-dominated points of hyperparameter combinations whose recovery rate and discard rate are lower than those of other hyperparameter combination data points are screened and filtered out, and the non-dominated points of hyperparameter combinations are screened and filtered out according to the set Draw a "recovery-waste rate" curve based on the predicted grade. When the "recovery-waste rate" set of the ore sorting algorithm to be evaluated is not Pareto optimal, that is, there are some data points whose recovery and waste rates are lower than other data points, also known as non-dominated points. At this time, the Pareto Frontier method can be used to screen out non-dominated points, and the dominated points retained in the set form the "recovery-waste rate" set of the new algorithm, which can be expressed as: .

[0072] The calculation function expression of the area ratio in step S3 of this embodiment is:

[0073] ,

[0074] In the above formula, Indicates the area ratio, The area enclosed by the preset “recovery rate-waste rate” curve based on the assay grade. is the area enclosed by the “recovery-waste rate” curve based on the predicted grade, and:

[0075] ,

[0076] ,

[0077] In the above formula, is the number of data points in the recovery-rejection curve based on the assay grade, is the number of data points in the recovery-waste curve based on the predicted grade, and are the recoveries of the nth and n+1th data points in the preset “recovery rate-waste rate” curve based on the assay grade, and are the discard rates of the nth and n+1th data points in the preset “recovery rate-discard rate” curve based on the assay grade, respectively. and are the recovery rates of the mth and m+1th data points of the “recovery rate-waste rate” curve based on the predicted grade, and are the discard rates of the mth and m+1th data points of the “recovery rate-discard rate” curve based on the predicted grade. and The value is used to evaluate the comprehensive performance of the sorting algorithm. The shape and position of the algorithm curve and the assay curve are combined to analyze the performance of the algorithm under different recovery rates and waste rates.

[0078] Before step S3 of this embodiment, the method further includes obtaining a preset "recovery rate-waste rate" curve based on the assay grade:

[0079] S101, the taste vector Sort in ascending order to get the sorted ore sample index set:

[0080] ,

[0081] In the above formula, ~ After sorting The index of the lump ore sample that satisfies ,in ~ They are ranked 1st to The grade of the bulk ore sample;

[0082] S102, for each grade threshold Calculate the corresponding scrap rate according to the following formula and recovery rate :

[0083] ,

[0084] ,

[0085] In the above formula, After sorting The serial number of the lump ore sample, After sorting The weight of the lump ore sample, After sorting The grade of the bulk ore sample;

[0086] S103, all grade thresholds Corresponding scrap rate and recovery rate As data points to form a data set :

[0087] ,

[0088] S104, according to the collection A "recovery rate-waste rate" curve based on the assay grade is drawn.

[0089] This embodiment can be used for the following guiding applications: judging whether the ore is suitable for sorting: Figure 7 As shown, the assay curve (the "recovery-waste rate" curve based on the assay grade) is used to determine whether the recovery rate and waste rate required for production can be achieved under optimal conditions. If the assay curve cannot meet the requirements, the ore may not be suitable for sorting. Select a suitable sorting strategy: Combine the assay curve and the algorithm curve (the "recovery-waste rate" curve based on the predicted grade), select the recovery rate and waste rate combination points that meet the production needs, and determine the corresponding algorithm parameters or sorting thresholds based on the selected combination points. Optimize the sorting algorithm: If the actual curve is significantly different from the assay curve, consider optimizing the sorting algorithm or trying a new algorithm to improve the sorting performance.

[0090] In this embodiment, the actual application of the ore sorting performance evaluation method has demonstrated significant results. Through the evaluation of multiple ore sorting algorithms, it is found that the performance differences of different ore sorting algorithms can be intuitively quantified by comparing the actual curve with the ideal curve. Taking Algorithm 1 and Algorithm 2, two known ore sorting algorithms for sorting in a certain mine, as an example, when the ore sorting performance evaluation method proposed in this embodiment is not adopted, since the performance of Algorithm 1 and Algorithm 2 is difficult to compare, people choose Algorithm 1 based on experience. However, after evaluating Algorithm 1 and Algorithm 2 using the "recovery rate-waste rate" curve proposed by the ore sorting performance evaluation method proposed in this embodiment, it is found that for Algorithm 1, the area ratio as a performance indicator is The value of is 73%, while the area ratio of Algorithm 2 as a performance indicator is The value reached 87%. Taking a batch of ore sorting algorithms as an example, when the ore sorting performance evaluation method proposed in this embodiment is not adopted, the algorithm research encounters a bottleneck, and the algorithm performance is still difficult to improve despite continuous research. However, after adopting the ore sorting performance evaluation method proposed in this embodiment, it is found that the "recovery rate-waste rate" curve of the algorithm is very close to the assay curve, proving that the algorithm has reached its limit. Taking a mine sorting as an example, when the ore sorting performance evaluation method proposed in this embodiment is not adopted, the researchers have no comprehensive evaluation method for the ideal sorting effect of the ore, and can only judge whether a batch of ore is suitable for sorting based on experience. However, after adopting the ore sorting performance evaluation method proposed in this embodiment, the ideal "recovery rate-waste rate" curve of the ore can be obtained. It is found that for a batch of ore, even the assay curve is poor and cannot meet the minimum requirements on site, so it is judged that this batch of ore has no meaning of sorting. In addition, the ore sorting performance evaluation method proposed in this embodiment can also provide users with optimized decision-making suggestions through actual curves without clear requirements or specific recovery rate and waste rate targets. For example, a certain mine site improved the sorting efficiency by using the maximum recovery rate-waste rate combination point recommended by the present invention, and ultimately achieved a result that exceeded expectations.

[0091] In one embodiment, 190 ore samples were collected from a mine, and the actual grade and weight of each sample were obtained. According to the embodiment method, the assay curve (the "recovery rate-waste rate" curve based on the assay grade) was calculated and drawn as follows: Figure 3 As shown. For a certain algorithm 1 to be evaluated, the algorithm parameters were adjusted by using different hyperparameter combinations to obtain multiple sets of recovery rate and waste rate combination points. The Pareto frontier method was used to screen out non-dominated points and construct the algorithm curve (the "recovery rate-waste rate" curve based on the predicted grade) as shown Figure 4 Calculate the area ratio of the curve enclosed by the "recovery rate-waste rate" curve based on the predicted grade and the curve enclosed by the preset "recovery rate-waste rate" curve based on the assay grade, and get the area ratio , indicating that the performance of Algorithm 1 needs to be improved. If the value is low, it is recommended to try a new sorting algorithm to obtain better sorting results.

[0092] In another embodiment, for a certain algorithm 2 to be evaluated, the same ore sample data and assay curve as those of algorithm 1 are used. By adjusting the hyperparameters of algorithm 2, different recovery rate and waste rate combination points are obtained. Since the data points generated by algorithm 2 are all non-dominated points, the algorithm curve is directly constructed. The final "recovery rate-waste rate" curves of algorithm 1 and algorithm 2 are shown in Figure 5 As shown in the figure, the area ratio of the curve enclosed area of ​​the "recovery rate-waste rate" curve based on the predicted grade and the curve enclosed area of ​​the preset "recovery rate-waste rate" curve based on the assay grade is calculated to obtain the area ratio , indicating that the performance of Algorithm 2 is close to the ideal state. Guidance application: Due to the area ratio The value is low, so it is recommended to try a new sorting algorithm to obtain better sorting results. By comparing the "recovery rate-waste rate" curve and area ratio of algorithm 1 and algorithm 2 , it is concluded that the performance of Algorithm 2 is better than that of Algorithm 1 and is more suitable for the sorting of this batch of ore.

[0093] In another embodiment, a ore pile in a mine was sampled and assayed, and the assay curve of the invention was drawn. The mine staff set an "ideal" recovery rate target for sorting. and scrap rate targets So, on the test curve, find the point closest to the target point. , based on Euclidean distance:

[0094] ,

[0095] According to the nearest point on the test data curve ( The smallest point) found that the ideal scrap rate can reach 40.90% and the recovery rate can reach 98.91%. The corresponding sorting grade threshold at this time is 0.5%, such as Figure 6 As shown. Use As the grade threshold, the ore samples were classified and the binary classification model was trained. The verification found that the final effect of the trained binary classification model met the requirements and was close to the expected effect.

[0096] In another embodiment, an experiment was conducted on a certain ore pile in a certain mine, and the algorithm's "recovery rate-waste rate" curve was obtained by adjusting the algorithm's hyperparameters. The mine staff set a "minimum" recovery rate target for sorting. and scrap rate targets So, in satisfying and Under the condition of The verification found that the actual sorting results met the requirements of the minimum recovery rate and waste rate and optimized the sorting performance.

[0097] In another embodiment, a sample of a ore pile in a mine was tested and the test curve of the invention was drawn. Before the algorithm and the sorting machine were developed, the on-site staff wanted to determine whether the ore was suitable for sorting. The test constructed by the invention found that even under ideal conditions, the combination of recovery rate and discard rate could not meet the production requirements, such as Figure 7 Based on the analysis results, it was decided not to sort the ore, thus avoiding unnecessary resource investment.

[0098] In summary, the ore sorting performance evaluation method of this embodiment proposes a method of constructing an assay recovery rate-waste rate curve to help users understand the best sorting effect that can be achieved under ideal conditions. This is an evaluation method that has not been widely adopted in the prior art. In the construction of the algorithm curve, the ore sorting performance evaluation method of this embodiment combines the characteristics of different sorting algorithms and proposes a method of using the Pareto Frontier algorithm to screen non-dominated points to construct the algorithm curve, which is particularly suitable for sorting algorithms with multi-parameter adjustment. The ore sorting performance evaluation method of this embodiment proposes a new sorting algorithm performance index E by calculating the area ratio between the assay curve and the actual curve. This index can quantify the gap between the algorithm and the ideal state, and provide a more comprehensive evaluation standard than the traditional single index. There is no direct corresponding implementation method for this evaluation system in the prior art.

[0099] In addition, the present embodiment also provides an ore sorting performance evaluation system, comprising a microprocessor and a memory connected to each other, wherein the microprocessor is programmed or configured to execute the ore sorting performance evaluation method. In addition, the present embodiment also provides a computer-readable storage medium, wherein a computer program or instruction is stored in the computer-readable storage medium, wherein the computer program or instruction is programmed or configured to execute the ore sorting performance evaluation method through a processor. In addition, the present embodiment also provides a computer program product, comprising a computer program or instruction, wherein the computer program or instruction is programmed or configured to execute the ore sorting performance evaluation method through a processor.

[0100] Those skilled in the art should understand that the technical solutions provided by the embodiments of the present application may be in the form of methods, systems, or computer program products. Therefore, the present application may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present application may take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. 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 produce 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 process Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be stored in a computer-readable memory that can guide a computer or other programmable data processing device to work in a specific way, so that the instructions stored in the computer-readable memory produce a product including an instruction device, which implements the functions specified in the process. Figure 1 A process or multiple processes and / or boxes Figure 1 These computer program instructions can also be loaded onto a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer-implemented process, so that the instructions executed on the computer or other programmable device provide for implementing the process in 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.

[0101] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.

Claims

1. A method for evaluating ore sorting performance, characterized in that: The steps include: S1, obtain the collected ore sample data; S2, using the ore sorting algorithm to be evaluated to predict the predicted grade of the ore sample, calculate the recovery rate and waste rate under different grade thresholds, and construct a "recovery rate-waste rate" curve based on the predicted grade; S3, calculating the area ratio of the curve enclosed area of ​​the "recovery rate-waste rate" curve based on the predicted grade and the curve enclosed area of ​​the preset "recovery rate-waste rate" curve based on the assay grade as the evaluation result of the ore sorting algorithm to be evaluated; the calculation function expression of the area ratio is: ， In the above formula, Indicates the area ratio, is the area enclosed by the preset "recovery rate-waste rate" curve based on the assay grade, is the area enclosed by the "recovery-waste rate" curve based on the predicted grade, and: ， ， In the above formula, is the number of data points of the "recovery-waste" curve based on the assay grade, is the number of data points in the "recovery-waste" curve based on the predicted grade, and are the recoveries of the nth and n+1th data points in the preset "recovery rate-waste rate" curve based on the assay grade, respectively. and are the discard rates of the nth and n+1th data points in the preset "recovery rate-discard rate" curve based on the assay grade, respectively. and are the recoveries of the mth and m+1th data points of the "recovery-waste rate" curve based on the predicted grade, and are the rejection rates of the mth and m+1th data points of the "recovery rate-rejection rate" curve based on the predicted grade, respectively.

2. The ore separation performance evaluation method according to claim 1, characterized in that: The X-axis of the "recovery rate-waste rate" curve is the waste rate, and the Y-axis is the recovery rate, or the X-axis of the "recovery rate-waste rate" curve is the recovery rate, and the Y-axis is the waste rate.

3. The ore separation performance evaluation method according to claim 1, characterized in that: Step S1 includes: randomly extracting minerals covering various compositions, grades and forms from the target ore; The bulk ore samples form a sample collection: ， In the above formula, For the sample collection, ~ They are No. 1~ lump ore sample; sample collection Each ore sample in Get their test grades respectively and weight Form the grade vector and weight vector: ， ， In the above formula, is the taste vector, ~ They are No. 1~ Assay grade of lump ore samples, is the weight vector, ~ They are No. 1~ The weight of the lump ore sample, Indicates transpose.

4. The ore separation performance evaluation method according to claim 3, characterized in that: The ore sorting algorithm to be evaluated is an ore sorting algorithm that predicts the predicted grade of ore samples based on single-energy XRT images and performs ore sorting. After step S1 and before step S2, it also includes respectively performing a Get its characterization data , the characterization data Refers to ore samples The single-energy XRT image is obtained to obtain the characterization vector : ， In the above formula, ~ They are No. 1~ Characterization data of lump ore samples.

5. The ore separation performance evaluation method according to claim 4, characterized in that: Step S2 includes: S2.1, adjust the hyperparameters of the ore sorting algorithm, obtain the predicted grade of the ore sample using the ore sorting algorithm to be evaluated under different hyperparameter combinations, and calculate the recovery rate and discard rate under different grade thresholds : ， In the above formula, For the The rejection rate under the hyperparameter combination, For the The recovery rate under the hyperparameter combination, is the number of hyperparameter combinations; S2.2, using the Pareto front method from the set The non-dominated points of hyperparameter combinations whose recovery rate and discard rate are lower than those of other hyperparameter combination data points are screened and filtered out, and the non-dominated points of hyperparameter combinations are screened and filtered out according to the set Plot a "recovery-waste" curve based on predicted grade.

6. The ore separation performance evaluation method according to claim 3, characterized in that: Before step S3, the method also includes obtaining a preset "recovery rate-waste rate" curve based on the assay grade: S101, the taste vector Sort in ascending order to get the sorted ore sample index set: ， In the above formula, ~ After sorting The index of the lump ore sample that satisfies ,in ~ They are ranked 1st to The grade of the bulk ore sample; S102, for each grade threshold Calculate the corresponding scrap rate according to the following formula and recovery rate : ， ， In the above formula, After sorting The serial number of the lump ore sample, After sorting The weight of the lump ore sample, After sorting The grade of the bulk ore sample; S103, all grade thresholds Corresponding scrap rate and recovery rate As data points to form a data set : ， S104, according to the collection A "recovery rate-waste rate" curve based on the assay grade was drawn.

7. An ore sorting performance evaluation system, comprising a microprocessor and a memory connected to each other, characterized in that: The microprocessor is programmed or configured to execute the ore sorting performance evaluation method according to any one of claims 1 to 6.

8. A computer-readable storage medium having a computer program or instruction stored therein, characterized in that: The computer program or instruction is programmed or configured to execute the ore sorting performance evaluation method according to any one of claims 1 to 6 through a processor.

9. A computer program product comprising a computer program or instructions, characterized in that The computer program or instruction is programmed or configured to execute the ore sorting performance evaluation method according to any one of claims 1 to 6 through a processor.

Images