Coal dressing decision-making method and system based on offline and online data fusion

Abstract

The invention provides a coal dressing decision-making method and system based on offline and online data fusion, and relates to the technical field of coal washing and dressing processing. According to the method, firstly, a dynamic mapping relation between an offline basic data set and an online real-time data set in the coal dressing process is established; then, a collaborative constraint condition across data types is generated based on the relationship. And then a decision influence evaluation model of the coal dressing process parameters is constructed through the conditions. And generating a coal preparation process adjustment scheme according to a model output result. And finally, the scheme is sent to a coal dressing production control system to be executed, the dynamic mapping relation and the collaborative constraint conditions are optimized according to execution result data, and coal dressing decision optimization is achieved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of intelligent coal washing and sorting processing, and particularly relates to a coal sorting decision method and system based on offline and online data fusion. BACKGROUND

[0002] In the field of coal sorting production, the accuracy and timeliness of coal sorting decision directly affect the efficiency of coal sorting and product quality. At present, there is a common problem of separate processing of offline data and online data in the process of coal sorting production. Offline data is usually manually arranged and archived regularly, and its processing period is relatively long, so it cannot reflect the current state of coal sorting production in real time. Online real-time data, such as real-time sensor collection data, production process data and equipment running state data, can be obtained in real time, but they are often used only for monitoring a single link and are not effectively associated with offline data. In the prior art, some coal sorting decision methods only rely on offline data to set process parameters, which leads to the fact that the decision result cannot adapt to the working condition changes in real-time production. When the composition of raw coal fluctuates, the raw coal ratio parameter adjustment strategy based on historical offline data may not be able to be adjusted in time to ensure product quality. Some other methods attempt to combine online data, but only process the two types of data through simple comparison or superposition, without considering the correlation and consistency problems caused by the difference between data types, so the data fusion effect is not good. At the same time, the existing coal sorting decision method lacks an effective collaborative constraint mechanism, and cannot standardize the association relationship of data from different sources, which leads to the fact that deviations are easy to occur when analyzing the influence of process parameter adjustment on coal sorting efficiency and product quality. In addition, after the decision result is generated, there is no feedback optimization mechanism for the execution result, so the decision model is difficult to continuously improve according to the actual production situation, and the decision accuracy gradually decreases after long-term use. The above problems together lead to the fact that the adaptability and accuracy of the existing coal sorting decision method are insufficient, and it is difficult to meet the demand of modern coal sorting production for efficient and high-quality decision. SUMMARY

[0003] In view of the above, in order to at least partially solve the problems in the prior art, in a first aspect, the present application provides a coal sorting decision method based on offline and online data fusion, which comprises: establishing a dynamic mapping relationship between an offline basic data set and an online real-time data set of a coal sorting process, the offline basic data set comprising historical production record data, laboratory analysis data and equipment maintenance data, and the online real-time data set comprising real-time sensor collection data, production process data and equipment running state data; generating a collaborative constraint condition across data types based on the dynamic mapping relationship; The decision influence evaluation model of the coal preparation process parameters is constructed through the synergistic constraint condition, and is used for analyzing the associated influence of the process parameter adjustment on the coal preparation efficiency and product quality. A coal preparation process adjustment scheme is generated according to the associated influence result output by the decision influence evaluation model; The coal preparation process adjustment scheme is sent to a coal preparation production control system for execution, and execution result data returned by the coal preparation production control system is received, and the dynamic mapping relationship and the synergistic constraint condition are optimized based on the execution result data.

[0004] In a second aspect, the embodiments of the present application further provide a coal preparation decision system based on offline and online data fusion, comprising a processor, a machine readable storage medium, the machine readable storage medium and the processor are connected, the machine readable storage medium is used for storing programs, instructions or codes, and the processor is used for executing the programs, instructions or codes in the machine readable storage medium to realize the coal preparation decision method based on offline and online data fusion.

[0005] In summary, the coal preparation decision method and system based on offline and online data fusion provided by the embodiments of the present application avoid the problem of separate processing of two types of data by establishing a dynamic mapping relationship between the offline basic data set and the online real-time data set, and realize the association of different source data. The dynamic mapping relationship can be adjusted adaptively according to the change of online data, ensuring that the offline data and online data always maintain effective matching, providing a reliable data basis for the generation of the synergistic constraint condition and the construction of the decision influence evaluation model. In addition, the cross-data-type synergistic constraint condition generated based on the dynamic mapping relationship standardizes the association and consistency requirements of different source data, avoiding analysis deviation caused by data differences. Under the support of the synergistic constraint condition, the decision influence evaluation model of the coal preparation process parameters can comprehensively and accurately analyze the associated influence of the process parameter adjustment on the coal preparation efficiency and product quality, covering direct and indirect influences, making the evaluation result more valuable. At the same time, the coal preparation process adjustment scheme generated according to the decision influence evaluation model contains adjustment strategies of the operation parameters of the separation equipment, the raw coal ratio parameters and the reagent addition parameters, and the scheme generation process comprehensively considers multi-objective evaluation, which can balance the demand of coal preparation efficiency and product quality, and meet the decision requirements in different production scenarios. Finally, the result data after the execution of the adjustment scheme is fed back to optimize the dynamic mapping relationship and the synergistic constraint condition, forming a closed-loop optimization mechanism, so that the whole decision method can continuously adapt to production changes, continuously improve the decision accuracy and reliability, and provide efficient and high-quality decision support for coal preparation production in the long term, significantly improving the overall efficiency of coal preparation production.

[0006] Other features and advantages of the present application will be described in the following detailed description. BRIEF DESCRIPTION OF DRAWINGS

[0007] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the embodiment description will be briefly introduced. Obviously, the drawings in the following description are only some embodiments of the present application, and other drawings can be obtained by those skilled in the art without creative effort on the basis of the above drawings.

[0008] In order to more completely understand the present application and its beneficial effects, the following will be described in conjunction with the drawings, wherein the same reference numerals in the following description represent the same parts.

[0009] Figure 1 FIG. 1 is a flow diagram of a coal preparation decision-making method based on offline and online data fusion provided by an embodiment of the present application.

[0010] Figure 2 FIG. 2 is a hardware environment diagram of a coal preparation decision-making system based on offline and online data fusion provided by an embodiment of the present application. DETAILED DESCRIPTION

[0011] The technical solutions in the embodiments of the present application will be described clearly and completely in conjunction with the drawings of the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative effort belong to the protection scope of the present application.

[0012] The present application will be described in detail below in conjunction with the drawings of the specification, Figure 1 FIG. 1 is a flow diagram of a coal preparation decision-making method based on offline and online data fusion provided by an embodiment of the present application. The coal preparation decision-making method based on offline and online data fusion includes the following steps S110-S150, which will be described in detail below.

[0013] Step S110: Establish a dynamic mapping relationship between the offline basic data set and the online real-time data set of the coal preparation process.

[0014] ​As a possible example, the embodiment takes a coal preparation plant as an example. In the daily sound field operation of the coal preparation plant, the offline basic data set contains multiple data types. The historical production record data records various types of information about past production, covering coal handling capacity, yield and quality of coal preparation products, and other related content in different time periods. The laboratory analysis data is the data obtained from various physical and chemical analyses of coal samples, such as composition indicators and characteristic parameters of coal. The equipment maintenance data includes all maintenance records of the equipment since it was put into use, including equipment failure occurrence, repair measures, and other information. For example, the laboratory analysis data in the offline basic data set can also be mainly formed around the periodic or when the raw coal is replaced (such as when the working face is relocated) raw coal sampling and screening float and sink report preparation. This part of data covers relatively comprehensive coal quality information, such as raw coal particle size distribution, density composition, ash content, sulfur content, calorific value, and other detailed indicators, which are important basis for process system adjustment analysis. However, since the coal preparation plant usually needs to update such data periodically (such as 3-6 months), it cannot reflect the dynamic changes of raw coal quality in real time.

[0015] The online real-time data set can also be a combination of multiple data. For example, real-time feedback of various physical quantities in the production site is collected through real-time sensors, such as operating parameters of conveying equipment, material parameters of each processing link, etc. The production process data describes the entire process information of coal from raw coal to final product output, including the sequence of each link, residence time, etc. The equipment running state data shows the current running status of the equipment, such as the running parameters of the key components, the overall running state indicators of the equipment, etc. For example, the online real-time data set can also include data obtained from various online detection sensors installed in the coal preparation plant, such as online ash content meters, electronic belt scales, flow meters, and concentration meters. The above sensors collect key process parameters in the production process in real time, including ash content, yield, flow, concentration, etc. Although these data are real-time, they can only reflect part of the information of coal quality and production process, and cannot fully present the overall picture of coal quality, so they also need to be analyzed in conjunction with offline data for subsequent decision-making of the coal preparation scheme.

[0016] In the embodiment, step S110 can include sub-steps S111-S116, which are described in detail below.

[0017] Step S111: Identify the production cycle characteristics of the historical production record data, the index correlation characteristics of the laboratory analysis data, and the fault mode characteristics of the equipment maintenance data in the offline basic data set, and generate an offline data feature map.

[0018] As a possible example, for the historical production record data, a specific method is used for time series processing. Through the processing, the data can be decomposed into different components, from which the component reflecting the production cycle law is extracted to determine the cycle characteristics of the historical production. For example, through the analysis of the coal handling capacity data in different time periods, the periodic variation law presented in the production process is found, which is the production cycle characteristics.

[0019] For laboratory analysis data, correlation analysis means is used to process each analysis index to determine the correlation between the indexes, including positive correlation and negative correlation, and the correlation degree. The correlation and degree are defined as index correlation characteristics. For example, the mutual influence relationship between different component indexes of coal is analyzed to obtain the correlation characteristics between them.

[0020] For equipment maintenance data, pattern recognition processing is performed on the fault records. The running state parameter sequence of the equipment before the fault occurs and the maintenance measures taken after the fault occurs are extracted from the fault records, and the corresponding relationship between the two is determined as the fault mode characteristics. For example, for the fault record of a certain equipment, the change of the running parameters of the key components of the equipment before the fault occurs and the specific maintenance operation taken after the fault occurs are analyzed to form the fault mode characteristics.

[0021] Further, the production cycle characteristics, index correlation characteristics, and fault mode characteristics are respectively set with characteristic identifiers. Specific identifier symbols are used to distinguish different types of characteristics, which facilitates the subsequent identification and processing of the characteristics. The characteristics and corresponding characteristic identifiers are integrated according to a pre-set atlas structure to generate an offline data characteristic atlas. The atlas structure can be designed according to the hierarchical relationship of the data types and characteristics to clearly show the relationship between the characteristics. Finally, the offline data characteristic atlas is subjected to characteristic integrity verification to ensure that the atlas contains all offline characteristics of the key links of the coal preparation production, and to ensure the data integrity for subsequent analysis and decision-making.

[0022] Step S112: Identify the fluctuation characteristics of real-time sensor collected data, the link connection characteristics of production process data, and the parameter coupling characteristics of equipment running state data in the online real-time data set, and generate an online data characteristic atlas.

[0023] As a possible example, for real-time sensor collected data, the change of the data over a period of time is observed, and the fluctuation, including fluctuation amplitude and frequency, is analyzed. The fluctuation-related information is determined as the fluctuation characteristics. For example, for the data of a certain physical quantity collected by a sensor in the production process, the change at different times is analyzed to obtain the fluctuation characteristics.

[0024] For production process data, the connection between each production link is studied, including the smoothness of material transfer and the matching of the running rhythm of each link. The information related to the connection between links is defined as a link connection feature. For example, by observing the coal conveying process between different processing equipment and analyzing the coordination between the operating parameters of each device, a link connection feature is formed.

[0025] For device operating state data, the mutual relationship between the operating parameters of the device is analyzed, and the coupling between the parameters is determined. The parameter coupling related information is defined as a parameter coupling feature. For example, by analyzing the mutual influence relationship between the operating parameters of multiple key components of the device, a parameter coupling feature is obtained.

[0026] The feature identifiers are set for the fluctuation feature, the link connection feature, and the parameter coupling feature. Different types of online data features are distinguished by specific identifiers. The features and corresponding feature identifiers are integrated according to the pre-set atlas structure to generate an online data feature atlas. The atlas structure can be designed according to the characteristics of online data and the logical relationship between features to clearly show the relationship between various types of online data features. Finally, the feature integrity of the online data feature atlas is checked to ensure that the atlas contains all online features of the real-time link of the coal preparation production, providing complete data support for subsequent data processing and decision-making.

[0027] Step S113: Constructing cross-atlas association dimensions based on the offline data feature atlas and the online data feature atlas.

[0028] As a possible example, from the offline data feature atlas and the online data feature atlas, find the dimensions that exist between the two. For example, it is found that the production cycle feature in the offline data feature atlas and the fluctuation feature in the online data feature atlas have a connection in the time dimension, and a time association dimension of the production cycle and the fluctuation feature can be constructed. At the same time, by analyzing the index association feature in the offline data feature atlas and the link connection feature in the online data feature atlas, it is found that they are associated in the production process, thereby constructing a process association dimension of the index association feature and the link connection feature. In addition, by studying the fault mode feature in the offline data feature atlas and the parameter coupling feature in the online data feature atlas, it is determined that they are related in the device-related aspect, thereby constructing a device association dimension of the fault mode feature and the parameter coupling feature. By constructing the cross-atlas association dimensions, the association between the offline data feature atlas and the online data feature atlas is established, providing a basis for subsequent data similarity calculation and dynamic mapping relationship establishment.

[0029] Step S114: Calculating the data similarity of the offline data feature atlas and the online data feature atlas in each cross-atlas association dimension to generate a dimension similarity vector.

[0030] As a possible example, for each cross-maps correlation dimension, the data similarity of the offline data feature maps and the online data feature maps in this dimension is calculated. For example, in the time correlation dimension, the time-related information involved in the production cycle feature in the offline data feature maps is compared with the time-related information of the fluctuation feature in the online data feature maps, and the similarity of the two in the time dimension is obtained by comparison. In the process correlation dimension, the similarity of the index correlation feature in the offline data feature maps and the link connection feature in the online data feature maps in the production process is analyzed, and the similarity in this dimension is obtained. In the equipment correlation dimension, the similarity of the failure mode feature in the offline data feature maps and the parameter coupling feature in the online data feature maps in the equipment-related aspect is studied, and the similarity in this dimension is calculated. The similarities calculated in each cross-maps correlation dimension are combined to form a vector, i.e. a dimension similarity vector. The vector reflects the similarity of the offline data feature maps and the online data feature maps in different correlation dimensions, and provides a basis for determining the data correlation rules of the dynamic mapping relationship.

[0031] Step S115: determining the data correlation rules of the dynamic mapping relationship according to the dimension similarity vector.

[0032] As a possible example, according to the similarity of each dimension reflected by the dimension similarity vector, the data matching mode and matching accuracy of the offline data and the online data in different correlation dimensions are determined. For example, in the time correlation dimension, if the similarity of this dimension in the dimension similarity vector is high, it indicates that the production cycle feature in the offline data feature maps and the fluctuation feature in the online data feature maps have strong correlation in time, and a more strict data matching mode and a higher matching accuracy can be set to ensure the accurate correspondence of the two data in the time dimension. In the process correlation dimension, according to the similarity of this dimension in the dimension similarity vector, the data matching mode and matching accuracy of the index correlation feature in the offline data feature maps and the link connection feature in the online data feature maps in the production process are determined, so that the two can be reasonably correlated in the process dimension. In the equipment correlation dimension, according to the similarity of this dimension in the dimension similarity vector, the data matching mode and matching accuracy of the failure mode feature in the offline data feature maps and the parameter coupling feature in the online data feature maps in the equipment are set, and the effective data correlation in the equipment dimension is realized. In this way, a complete set of data correlation rules is formed to define the data matching relationship of the offline data and the online data in different correlation dimensions, so as to facilitate the establishment of the dynamic mapping relationship.

[0033] Step S116: establishing the dynamic mapping relationship between the offline basic data set and the online real-time data set based on the data correlation rules, and setting an adaptive updating mechanism for the dynamic mapping relationship.

[0034] As a possible example, according to the determined data association rules, a mapping relationship is established between the offline basic data set and the online real-time data set. For example, in the time association dimension, according to the set data matching mode and matching accuracy, the offline basic data corresponding to the production cycle feature in the offline data feature map is associated with the online real-time data corresponding to the fluctuation feature in the online data feature map. In the process association dimension and the equipment association dimension, the association of the offline basic data and the online real-time data in these two dimensions is completed according to the corresponding data association rules, thereby establishing a dynamic mapping relationship between the offline basic data set and the online real-time data set.

[0035] Meanwhile, an adaptive updating mechanism is set for the dynamic mapping relationship. When the fluctuation feature in the online data feature map changes, for example, the fluctuation amplitude or frequency changes significantly, the adjustment strategy of the data association rule is triggered according to the change frequency. If the fluctuation feature change frequency is high, it indicates that the production process may change greatly, at which time the data association rule is adjusted accordingly to adapt to the new production situation. Moreover, the mapping strength of each association dimension in the dynamic mapping relationship is positively correlated with the similarity value in the dimension similarity vector, that is, the higher the similarity value, the greater the mapping strength of the association dimension, ensuring that the mapping relationship can accurately reflect the actual association degree between the offline data and the online data, and real-time adjustment with the change of production situation, maintaining the effectiveness and accuracy of the dynamic mapping relationship.

[0036] Step S120: generating a cross-data-type collaborative constraint condition based on the dynamic mapping relationship.

[0037] Based on the above established dynamic mapping relationship, a cross-data-type collaborative constraint condition is generated to ensure the consistency and association of different source data in the coal preparation process, and to provide a reliable data basis for subsequent coal preparation process parameter decision-making. In this embodiment, step S120 can include sub-steps S121-S125, which are described in detail below.

[0038] Step S121: analyzing the data association rules in the dynamic mapping relationship and extracting the matching threshold of the offline data and the online data in each association dimension.

[0039] As just one possible example, the matching threshold includes the maximum time deviation in the time-related dimension, the index deviation range in the process-related dimension, and the parameter matching degree threshold in the equipment-related dimension. A thorough analysis is performed on the data association rules in the established dynamic mapping relationship. In the time-related dimension, the maximum allowable time deviation between the timestamp of offline production cycle data and the timestamp of online real-time data is determined based on the data association rules; this is the maximum time deviation matching threshold for the time-related dimension. For example, by parsing the data association rules in the time-related dimension, the maximum acceptable time difference between the start time of a certain production cycle recorded in offline data and the corresponding timestamp in online real-time data is clarified.

[0040] In the process association dimension, the allowable deviation range between offline laboratory analysis indicators and online production process indicators is extracted from the data association rules, including the allowable absolute deviation value and the allowable relative deviation value. This serves as the indicator deviation range matching threshold for the process association dimension. For example, by analyzing the data association rules in the process association dimension, the allowable absolute difference and relative proportion difference between the content of a certain component in coal obtained from offline laboratory analysis and the content of that component detected in the online production process are determined.

[0041] In the device association dimension, the minimum matching requirements between offline fault mode features and online device operating status parameters are defined according to data association rules; that is, the parameter matching degree threshold of the device association dimension. For example, by parsing the data association rules of the device association dimension, the minimum matching degree between the key parameters corresponding to a certain device fault mode recorded offline and the parameters in the online device operating status is determined. By extracting the matching thresholds under each association dimension, specific quantitative basis is provided for the subsequent generation of collaborative constraint rules.

[0042] Step S122: Generate time-coordinated constraint rules based on the maximum time deviation of the time correlation dimension.

[0043] As just one possible example, the time-coordination constraint rule is used to limit the difference between the timestamps of offline production cycle data and the timestamps of online real-time data. This rule is generated based on the maximum time deviation matching threshold of the time correlation dimension. This rule strictly limits the difference between the timestamps of offline production cycle data and the timestamps of online real-time data. For example, if the maximum time deviation matching threshold of the time correlation dimension is set to a specific duration, then the time-coordination constraint rule requires that the time difference between the timestamps of offline production cycle data and the timestamps of online real-time data cannot exceed this set duration. Through such a rule, consistency between offline and online data in the time dimension is ensured, preventing data correlation failure due to excessive time differences, and providing an accurate data foundation for time-series-based analysis and decision-making in coal preparation.

[0044] Step S123: generating a process coordination constraint rule based on the index deviation range of the process correlation dimension. The process coordination constraint rule is used to constrain the deviation between the offline laboratory analysis index and the online production process index within the allowed interval.

[0045] In this embodiment, step S123 can include sub-steps S1231-S1236, which are described in detail below.

[0046] Step S1231: analyzing the data correlation rule of the process correlation dimension in the dynamic mapping relationship to extract the index deviation range including the absolute deviation allowed value and the relative deviation allowed value.

[0047] As only one possible example, the data correlation rule of the process correlation dimension in the dynamic mapping relationship is analyzed. By analyzing the rule content, the allowed absolute deviation allowed value and the relative deviation allowed value between the offline laboratory analysis index and the online production process index are determined to determine the index deviation range. For example, for a certain quality index of coal, the allowed maximum absolute difference between the offline analysis value and the online production process detection value is obtained from the data correlation rule, as well as the allowed maximum relative difference between the two, thereby determining the deviation range of the index, providing specific quantitative parameters for subsequent construction of process coordination constraint rules.

[0048] Step S1232: constructing an absolute deviation constraint sub-rule based on the absolute deviation allowed value to limit the actual numerical difference between the offline laboratory analysis index and the online production process index.

[0049] As only one possible example, an absolute deviation constraint sub-rule is constructed according to the extracted absolute deviation allowed value. The function of this sub-rule is to strictly limit the actual numerical difference between the offline laboratory analysis index and the online production process index. For example, if the absolute deviation allowed value is set to a specific value, the absolute deviation constraint sub-rule requires that the difference between the value of a certain coal quality index obtained by offline laboratory analysis and the value of the index detected in the online production process cannot exceed the set value. Through such a sub-rule, the consistency of offline data and online data is ensured at the actual numerical level, avoiding the influence of too large numerical difference on the control and decision of product quality in the coal preparation process.

[0050] Step S1233: constructing a relative deviation constraint sub-rule based on the relative deviation allowed value and used to limit the ratio deviation between the offline laboratory analysis index and the online production process index.

[0051] As a possible example, a relative deviation constraint sub-rule is constructed using the extracted relative deviation allowance value. The sub-rule aims to limit the deviation of the ratio between the offline laboratory analysis index and the online production process index. For example, if the relative deviation allowance value is set to a certain percentage, the relative deviation constraint sub-rule stipulates that the deviation of the ratio between the value of a certain coal quality index of the offline laboratory analysis and the value of the index detected in the online production process cannot exceed the set percentage. Through the sub-rule, consistency between offline data and online data is ensured from the perspective of relative proportion, providing accurate data support for analysis and decision-making based on proportional relationships in the coal preparation process.

[0052] Step S1234: Identify the key quality control points in the coal preparation production process, and set deviation weighting coefficients for the indexes corresponding to the key quality control points.

[0053] As a possible example, a comprehensive analysis of the coal preparation production process identifies key quality control points that have a critical impact on the quality of the final product. For example, in the coal preparation process, the coal separation link, clean coal dewatering link, etc. can be key quality control points. For the indexes corresponding to the key quality control points, different deviation weighting coefficients are set according to their impact on the quality of the final product. The indexes corresponding to key quality control points that have a greater impact on the quality of the final product are given higher deviation weighting coefficients, while those with a smaller impact are given lower deviation weighting coefficients. The deviation weighting coefficients are positively correlated with the impact of the key quality control points on the quality of the final product. By setting the deviation weighting coefficients, the importance of the key quality control points in the coal preparation process is highlighted, providing a basis for generating more accurate process coordination constraint rules.

[0054] Step S1235: Weighted fusion of the absolute deviation constraint sub-rule and the relative deviation constraint sub-rule with the deviation weighting coefficients to generate a weighted deviation constraint sub-rule.

[0055] As a possible example, the absolute deviation constraint sub-rule and the relative deviation constraint sub-rule constructed earlier are fused with the deviation weighting coefficients set for the key quality control point indexes. For example, according to the degree of influence of the absolute deviation constraint sub-rule and the relative deviation constraint sub-rule on different key quality control point indexes, combined with the corresponding deviation weighting coefficients, the two are fused through a specific weighting method to generate a weighted deviation constraint sub-rule. This weighted deviation constraint sub-rule takes into account the absolute deviation, relative deviation, and importance of the key quality control points, and can more comprehensively and accurately constrain the deviation between the offline laboratory analysis index and the online production process index, providing more effective rule support for quality control in the coal preparation process.

[0056] Step S1236: According to the quality grade requirements of the coal preparation products, set a grade correction coefficient for the weighted deviation constraint sub-rule, generate a process coordination constraint rule containing quality grade differentiation, and use it to match the quality control requirements of different grade products.

[0057] As a possible example, the weighted deviation constraint sub-rule is further optimized according to the different quality grade requirements of the coal preparation products. For different quality grade coal preparation products, set corresponding grade correction coefficients for the weighted deviation constraint sub-rule. For example, for high quality grade coal preparation products, set a relatively strict grade correction coefficient to more strictly control the deviation between offline laboratory analysis indicators and online production process indicators; for lower quality grade products, set a relatively loose grade correction coefficient. In this way, a process coordination constraint rule containing quality grade differentiation is generated, which can accurately match the quality control requirements of different grade products and ensure that the coal preparation process can be effectively controlled and managed according to the quality grade requirements of the products.

[0058] Step S124: Generate a device coordination constraint rule based on the parameter matching degree threshold of the device association dimension.

[0059] As a possible example, the device coordination constraint rule is used to define the minimum matching requirement between offline fault mode characteristics and online device operating state parameters. Based on the parameter matching degree threshold of the device association dimension, a device coordination constraint rule is generated. This rule is used to clearly specify the minimum matching requirement between offline fault mode characteristics and online device operating state parameters. For example, if the parameter matching degree threshold of the device association dimension is set to a certain percentage, the device coordination constraint rule requires that the matching degree between the key parameters corresponding to the offline recorded device fault mode and the parameters in the online device operating state cannot be lower than the set percentage. Through such a rule, the consistency of offline data and online data at the device operating parameter level is ensured, and abnormal situations that may occur during device operation are discovered in a timely manner, providing a guarantee for device maintenance and the stability of coal preparation production.

[0060] Step S125: Logically combine the time coordination constraint rule, the process coordination constraint rule, and the device coordination constraint rule to generate a coordination constraint condition containing multiple rules, and set a constraint priority for each rule in the coordination constraint condition.

[0061] As a possible example, the constraint priority is determined according to safety requirements, efficiency requirements and quality requirements of the coal preparation production. The time coordination constraint rule, the process coordination constraint rule and the equipment coordination constraint rule generated in the foregoing are logically integrated. For example, the three rules are combined together by a specific logical combination manner to form a coordination constraint condition containing multiple rules. Meanwhile, constraint priorities of the rules in the coordination constraint condition are set according to the safety requirements, the efficiency requirements and the quality requirements of the coal preparation production. In the coal preparation production, if the safety requirements are the most critical, the rules related to safety (for example, the part of the equipment coordination constraint rule related to the safe operation parameters of the equipment) are set to a higher constraint priority; if the current production task pays more attention to efficiency, the priority of the rules related to efficiency (for example, the part of the time coordination constraint rule related to the time limit of the production cycle) is increased accordingly. By reasonably setting the constraint priorities, it is ensured that the coordination constraint condition can effectively play a role under different production demands and guarantee the smooth progress of the coal preparation production process.

[0062] Step S130: constructing a decision influence evaluation model of the coal preparation process parameters by the coordination constraint condition.

[0063] As a possible example, based on the coordination constraint condition generated in the foregoing, a decision influence evaluation model of the coal preparation process parameters is constructed to deeply analyze the correlation influence of the process parameter adjustment on the coal preparation efficiency and the product quality, thereby providing a basis for the coal preparation process adjustment. In this embodiment, step S130 can include sub-steps S131-S137, which are described in detail as follows.

[0064] Step S131: determining adjustable process parameters affecting the coal preparation efficiency and the product quality based on the coal preparation process, and generating an initial process parameter set.

[0065] As a possible example, for the coal preparation process of the coal preparation plant, starting from the raw coal entering the coal preparation plant, going through various processing links such as crushing, screening, separation, dehydration and the like, process parameters that can affect the coal preparation efficiency and the product quality in the process are determined. For example, in the separation link, the operation parameters of the separation equipment such as the density and flow rate of the separation medium; in the raw coal preparation link, the raw coal ratio parameter, i.e., the mixing ratio of different types of raw coal; in the reagent adding link, the reagent addition amount, addition time and the like, which can all affect the coal preparation efficiency and the product quality. The adjustable process parameters are collected to form an initial process parameter set, thereby providing basic parameters for subsequent screening and model construction.

[0066] Step S132: screening the initial process parameter set based on the process coordination constraint rule in the coordination constraint condition, removing unadjustable parameters beyond the index deviation range, and retaining adjustable key process parameters.

[0067] As a possible example, the parameters in the initial set of process parameters are screened one by one according to the process synergy constraint rules in the previously generated synergy constraint conditions. The process synergy constraint rules contain the restrictions on the deviation range between the offline laboratory analysis indicators and the online production process indicators. For example, for a process parameter related to coal quality, if according to the process synergy constraint rules, the allowed deviation range is within a certain numerical interval, and the value of the parameter in the initial set of process parameters exceeds this range and cannot be adjusted to the allowed range under the current production conditions, the parameter is removed from the initial set of process parameters. Through such a screening process, the key process parameters that can both affect the coal separation efficiency and product quality and be adjusted within the allowed range of the process synergy constraint rules are retained, ensuring that the parameters relied on by the subsequently constructed model have practical operability and effectiveness.

[0068] Step S133: Collect the historical adjustment records of the key process parameters in the offline basic data set and the corresponding historical coal separation efficiency, historical product quality data to generate a historical process result data set.

[0069] As a possible example, in the offline basic data set of the coal preparation plant, the historical adjustment records of the previously screened key process parameters are collected. For example, the time of each adjustment of the operation parameters of the separation equipment, the adjustment amplitude, and the corresponding coal separation efficiency and product quality data are recorded. The data reflects the results of coal separation production under different key process parameter settings in the past. The historical adjustment records and the corresponding historical coal separation efficiency, historical product quality data are sorted and associated to form a historical process result data set. The data set contains the historical correspondence between the changes of the key process parameters and the changes of the coal separation efficiency and the product quality, providing historical data support for the subsequent analysis of the causal relationship between the key process parameters and the coal separation efficiency and the product quality.

[0070] Step S134: Associate the historical process result data set with the real-time production data in the online real-time data set through the dynamic mapping relationship to generate a fusion process result data set containing offline historical experience and online real-time state.

[0071] As a possible example, the historical process result data set is associated with real-time production data in the online real-time data set by using the dynamic mapping relationship between the offline basic data set and the online real-time data set. For example, according to the time correlation dimension, the process correlation dimension and the equipment correlation dimension in the dynamic mapping relationship, the time, the production process and the equipment information corresponding to a certain key process parameter adjustment record in the historical process result data set are matched and associated with the real-time production data of the corresponding time, the process and the equipment in the online real-time data set. In this way, the offline historical experience is combined with the online real-time state to generate a fusion process result data set. The data set not only contains the influence information of the key process parameter adjustment on the coal separation efficiency and the product quality in history, but also integrates the data of the current real-time production state, so as to more accurately analyze the relationship between the key process parameter and the coal separation efficiency and the product quality.

[0072] Step S135: Based on the fusion process result data set, a causal inference method is used to analyze the causal relationship between the key process parameter and the coal separation efficiency and the product quality, and to identify the direct causal path and the indirect causal path.

[0073] In this embodiment, step S135 can include sub-steps S1351-S1357, which will be described in detail below.

[0074] Step S1351: The key process parameter, the coal separation efficiency and the product quality data in the fusion process result data set are subjected to variable standardization processing, so that each variable has a unified data scale.

[0075] As a possible example, the key process parameter, the coal separation efficiency and the product quality data involved in the fusion process result data set are processed. Since the data can have different dimensions and value ranges, in order to facilitate subsequent analysis and comparison, a specific method is used to perform variable standardization processing. For example, the value range of the medium density in the key process parameter can be in a certain interval, while the coal separation efficiency can be expressed in percentage, and the product quality can be expressed in a certain quality index value. Through standardization processing, the different types of data are converted into a form with a unified data scale, so as to be compared and analyzed under the same standard.

[0076] Step S1352: Based on the standardized variables, an initial structure of a causal relationship graph is constructed, and the initial structure includes potential association edges between each key process parameter and the coal separation efficiency and the product quality.

[0077] As a possible example, an initial structure of the causal relationship diagram is constructed based on the standardized key process parameters, coal preparation efficiency and product quality data. In the initial structure, a potential association edge is established between each key process parameter and the coal preparation efficiency and product quality, indicating that there may be a causal relationship between them. For example, a potential association edge is established between the separation medium density in the separation equipment operation parameter and the coal preparation efficiency, implying that the change of the separation medium density may affect the coal preparation efficiency; similarly, a potential association edge is established between the raw coal ratio parameter and the product quality, indicating that the change of the raw coal ratio may affect the product quality.

[0078] Step S1353: The initial structure is subjected to causal relationship test by using a causal inference algorithm to calculate the causal effect value of each potential association edge, which is used to represent the degree of influence of the cause variable on the result variable.

[0079] Exemplarily, the initial structure of the causal relationship diagram constructed in the foregoing is tested by using a specific causal inference algorithm. For each potential association edge, the causal inference algorithm is used to calculate its causal effect value. For example, for the potential association edge between the separation medium density and the coal preparation efficiency, the causal inference algorithm is used to calculate the degree of influence of the separation medium density as the cause variable on the coal preparation efficiency as the result variable, combined with the historical data and real-time data of these two variables in the fusion process result data set, and the degree of influence is the causal effect value of the potential association edge. By calculating the causal effect value of each potential association edge, the strength of the possible causal relationship between the key process parameters and the coal preparation efficiency and product quality can be quantified.

[0080] Step S1354: Significant association edges are screened according to the significance level of the causal effect value, and association edges with no significant causal effect are removed to generate a simplified causal relationship diagram.

[0081] Exemplarily, all potential association edges are screened according to the significance level of the causal effect value. A significance level standard is set, for example, a certain specific numerical value or percentage. If the causal effect value of a potential association edge reaches or exceeds the standard, it is considered that the association edge has a significant causal effect and is retained; if the causal effect value does not reach the standard, it is considered that the association edge has no significant causal effect and is removed from the causal relationship diagram. Through such a screening process, association edges that have no significant influence on the coal preparation efficiency and product quality in actual production are removed, and a simplified causal relationship diagram is generated. The simplified causal relationship diagram more clearly shows the causal relationships between the key process parameters and the coal preparation efficiency and product quality that have a significant influence.

[0082] Step S1355: In the simplified causal relationship diagram, the association edge directly pointing from the key process parameter to the coal preparation efficiency or product quality is identified, and the association edge is defined as a direct causal path.

[0083] In the simplified causal relationship diagram, the associated edges directly connecting the key process parameters to the coal cleaning efficiency or product quality are found. For example, if the separation medium density is found to be directly connected to the coal cleaning efficiency in the diagram, and this associated edge is confirmed to have a significant causal effect in the previous screening process, this associated edge is defined as a direct causal path. The direct causal path indicates that the change of the key process parameter can directly affect the coal cleaning efficiency or product quality, and provides a clear path for analyzing the direct effect of the key process parameter on the coal cleaning production result.

[0084] Step S1356: identifying the associated edge sequence indirectly pointing from each key process parameter to the coal cleaning efficiency or product quality through other key process parameters, and defining the associated edge sequence as an indirect causal path.

[0085] For example, assuming that the key process parameter A is connected to the coal cleaning efficiency through the key process parameter B, and these associated edges are all confirmed to have a significant causal effect in the previous screening process, the associated edge sequence from the key process parameter A to the coal cleaning efficiency through the key process parameter B is defined as an indirect causal path. The indirect causal path reflects the mutual influence between the key process parameters, and the indirect effect of the key process parameters on the coal cleaning efficiency or product quality.

[0086] Step S1357: calculating the total causal effect of the indirect causal path, and adding the causal direction identifier and causal effect value label to the direct causal path and indirect causal path to form a causal relationship network containing path characteristics. The total causal effect is the product of the causal effect values of the associated edges in the path.

[0087] In one embodiment, the total causal effect of each indirect causal path is calculated. Since the indirect causal path is composed of multiple associated edges, the total causal effect of the indirect causal path is the product of the causal effect values of the associated edges in the path. For example, if an indirect causal path is composed of three associated edges, and the causal effect values of the three associated edges are A, B and C respectively, the total causal effect of the indirect causal path is A multiplied by B multiplied by C. At the same time, the causal direction identifier is added to the direct causal path and indirect causal path to clearly indicate the relationship direction between the cause variable and the result variable, and the causal effect value label is added to mark the causal effect size of each path. Through such processing, a causal relationship network containing path characteristics is formed. The network clearly shows the direct and indirect causal relationships between the key process parameters and the coal cleaning efficiency or product quality, as well as the strength and direction of the relationships, and provides detailed causal relationship information for constructing a decision influence evaluation model of the coal cleaning process parameters.

[0088] Step S136: Constructing a multi-factor influence network according to the direct causal path and the indirect causal path.

[0089] In an embodiment, the nodes in the multi-factor influence network are the key process parameters, the coal separation efficiency and the product quality, and the edges are the causal influence relationships and the influence strengths. Based on the identified direct causal path and the indirect causal path, a multi-factor influence network is constructed. In the network, the key process parameters, the coal separation efficiency and the product quality are taken as nodes. For example, the key process parameters such as the separation medium density and the raw coal ratio parameter are taken as nodes, and the coal separation efficiency and the product quality are also taken as nodes. The causal influence relationships represented by the direct causal path and the indirect causal path are taken as edges to connect the corresponding nodes, and the strengths of the causal influence relationships, i.e. the causal effect values calculated above, are marked on the edges. Through such a construction manner, a multi-factor influence network is formed, which intuitively shows the complex causal relationships and mutual influence strengths between the key process parameters, the coal separation efficiency and the product quality.

[0090] Step S137: Integrating the equipment coordination constraint rules and the time coordination constraint rules in the coordination constraint conditions into the multi-factor influence network as network constraint conditions to generate a decision influence evaluation model of the coal separation process parameters.

[0091] In an embodiment, the equipment coordination constraint rules and the time coordination constraint rules in the coordination constraint conditions generated above are taken as network constraint conditions and integrated into the multi-factor influence network. The equipment coordination constraint rules limit the relationships between the equipment-related key process parameters and the equipment operating states, such as the matching requirements between the separation equipment operating parameters and the equipment operating parameters. The limit conditions are taken as constraint conditions of the equipment-related nodes and edges in the network to ensure that the adjustment of the key process parameters is within the range of equipment operation. The time coordination constraint rules limit the production cycle time and other aspects, which are taken as constraint conditions related to time in the network to ensure that the coal separation process is carried out within a reasonable time frame. By integrating the two constraint rules into the multi-factor influence network, a decision influence evaluation model of the coal separation process parameters is generated. The model comprehensively considers the causal relationships between the process parameters, the equipment operation limitations and the time limitations and other factors, and can more accurately analyze the associated influence of the process parameter adjustment on the coal separation efficiency and the product quality.

[0092] Step S140: Generating a coal separation process adjustment scheme according to the associated influence results output by the decision influence evaluation model.

[0093] In one embodiment, based on the decision impact evaluation model output of the previously constructed coal preparation process parameter relevance impact results, a coal preparation process adjustment scheme is generated to optimize the coal preparation production process and improve the coal preparation efficiency and product quality. In this embodiment, step S140 can include sub-steps S141-S149, which are described in detail below.

[0094] Step S141: Analyze the relevance impact results output by the decision impact evaluation model, and extract the influence coefficients of each key process parameter on the coal preparation efficiency and product quality through direct and indirect causal paths. Among them, the influence coefficients include direct and indirect influence coefficients.

[0095] In one embodiment, for each key process parameter in the relevance impact results output by the decision impact evaluation model, the influence coefficient of the key process parameter on the coal preparation efficiency and product quality through the direct causal path, i.e. the direct influence coefficient, is extracted. For example, for the key process parameter of separation medium density, the coefficient value of the direct influence of the key process parameter on the coal preparation efficiency is obtained from the model output results. At the same time, the influence coefficient of the key process parameter on the coal preparation efficiency and product quality through the indirect causal path, i.e. the indirect influence coefficient, is extracted. For example, if the separation medium density affects another key process parameter, which in turn affects the coal preparation efficiency, the influence coefficient value of this indirect causal path is obtained from the model output results. By extracting the influence coefficients, the direct and indirect influence degrees of each key process parameter on the coal preparation efficiency and product quality are determined, providing data support for subsequent calculation of the comprehensive influence index.

[0096] Step S142: Calculate the comprehensive influence index of each key process parameter based on the influence coefficients, and the comprehensive influence index is the weighted sum of the direct and indirect influence coefficients. Among them, the weight is positively correlated with the significance level of the causal path.

[0097] In one embodiment, the comprehensive influence index of each key process parameter is calculated based on the previously extracted direct and indirect influence coefficients. The comprehensive influence index is the weighted sum of the direct and indirect influence coefficients. For a causal path with a higher significance level, a higher weight is given to the corresponding influence coefficient; for a causal path with a lower significance level, a lower weight is given to the corresponding influence coefficient. For example, if a key process parameter affects the coal preparation efficiency through a direct causal path with a significant causal effect, and at the same time affects the coal preparation efficiency through an indirect causal path with a relatively weak causal effect, then in the calculation of the comprehensive influence index, the weight of the direct influence coefficient will be relatively high, and the weight of the indirect influence coefficient will be relatively low. Through such a weighting method, the comprehensive influence index of each key process parameter is calculated, which comprehensively reflects the overall influence degree of the key process parameter on the coal preparation efficiency and product quality through direct and indirect paths.

[0098] Step S143: Sort the key process parameters according to the comprehensive influence index, and determine the priority order of parameter adjustment. For example, the parameter with high comprehensive influence index is adjusted first.

[0099] For example, according to the calculated comprehensive influence index of each key process parameter, the key process parameters are sorted. The key process parameters with higher comprehensive influence index are placed in front, and the key process parameters with lower comprehensive influence index are placed in back. For example, assuming that after calculation, the comprehensive influence index of the separation medium density is higher than that of the raw coal ratio parameter, then the separation medium density is placed before the raw coal ratio parameter in the sorting. According to the sorting result, the priority order of parameter adjustment is determined, that is, the parameter with high comprehensive influence index is adjusted first. Such sorting and adjustment order can ensure that in the process of coal preparation process adjustment, the key process parameters that have greater influence on coal preparation efficiency and product quality are adjusted first, thereby improving the pertinence and effectiveness of process adjustment.

[0100] Step S144: Determine the allowed adjustment range of each key process parameter based on the process coordination constraint rule in the coordination constraint condition, and the allowed adjustment range corresponds to the index deviation range.

[0101] For example, according to the process coordination constraint rule in the coordination constraint condition generated in the foregoing, the allowed adjustment range of each key process parameter is determined. The process coordination constraint rule contains the limitation of the deviation range between the offline laboratory analysis index and the online production process index, and the key process parameter is closely related to the index. For example, for the key process parameter related to coal quality, the allowed adjustment range of the key process parameter is determined according to the deviation range of the quality index specified in the process coordination constraint rule. If the process coordination constraint rule specifies that the allowed deviation range of a certain coal quality index is within a certain numerical interval, then the adjustment range of the key process parameter related to the index must be ensured to be within the allowed deviation range after adjustment. In this way, it is ensured that the adjustment of the key process parameter is within a reasonable range, and the situation of inconsistent data or unqualified product quality in the production process is avoided.

[0102] Step S145: Generate multiple candidate adjustment values for each key process parameter within the allowed adjustment range, and the interval of the candidate adjustment value is determined according to the sensitivity of the parameter. For example, the interval of the parameter with high sensitivity is small to realize fine adjustment.

[0103] For example, a plurality of candidate adjustment values are generated for each key process parameter within the allowable adjustment range of the determined key process parameters. The interval of the candidate adjustment values is determined according to the sensitivity of different key process parameters. A key process parameter with high sensitivity may have a slight change that can have a great impact on the coal separation efficiency and product quality, and therefore a smaller interval of candidate adjustment values is set for such a parameter to achieve fine adjustment. For example, the separation medium density can be a key process parameter with high sensitivity, and a plurality of candidate adjustment values are set at a smaller interval within the allowable adjustment range, for example, starting from the lower limit value, a candidate adjustment value is set every smaller value within the allowable adjustment range. For a key process parameter with low sensitivity, a relatively larger interval of candidate adjustment values is set. In this way, a series of suitable candidate adjustment values are generated for each key process parameter, providing multiple choices for generating an initial process adjustment scheme.

[0104] Step S146: combining the candidate adjustment values of different key process parameters to generate a plurality of initial process adjustment schemes, each initial process adjustment scheme containing a complete set of parameter adjustment combinations.

[0105] For example, the candidate adjustment values of different key process parameters are combined comprehensively. For example, assume that there are three key process parameters A, B and C, the key process parameter A has three candidate adjustment values a1, a2 and a3, the key process parameter B has two candidate adjustment values b1 and b2, and the key process parameter C has three candidate adjustment values c1, c2 and c3. Through the combination method, a plurality of initial process adjustment schemes are generated, such as (a1, b1, c1), (a1, b1, c2), (a1, b2, c1), etc. Each initial process adjustment scheme contains a complete set of parameter adjustment combinations, i.e. corresponding adjustment values are set for all key process parameters. Through this combination method, a plurality of different initial process adjustment schemes are generated, providing a basis for subsequent screening of the optimal coal separation process adjustment scheme.

[0106] Step S147: inputting the initial process adjustment scheme into the decision influence evaluation model to obtain the predicted values of the coal separation efficiency and product quality corresponding to each scheme.

[0107] By way of example only, each of the previously generated initial process adjustment schemes is input into the decision impact assessment model of the coal preparation process parameters. The model, based on the adjustment values of the key process parameters in the input initial process adjustment schemes, in combination with the established causal relationship network and constraint conditions, predicts the coal preparation efficiency and product quality. For example, for a certain initial process adjustment scheme, the model analyzes the impact of the adjustment of the key process parameters in the scheme on the coal preparation process, and through causal relationship calculation and constraint condition limitation, outputs the corresponding predicted value of the coal preparation efficiency and the predicted value of the product quality of the scheme. By inputting all the initial process adjustment schemes into the model, the predicted values corresponding to each scheme are obtained.

[0108] Step S148: A multi-objective evaluation function containing the efficiency target weight and the quality target weight is constructed, and the evaluation scores of the initial process adjustment schemes are calculated based on the predicted values of the coal preparation efficiency and the product quality.

[0109] By way of example only, the efficiency target weight and the quality target weight are determined according to the current demand of the coal preparation production. When the production task prioritizes yield, the efficiency target weight is greater than the quality target weight; when the production task pays more attention to product quality, the quality target weight is greater than the efficiency target weight. And the sum of the efficiency target weight and the quality target weight is a set proportion. For example, if the current production task prioritizes yield, the efficiency target weight is set to a certain large proportion, and the quality target weight is set to a corresponding small proportion, and the sum of the two is 100%.

[0110] The predicted values of the coal preparation efficiency and the product quality are normalized respectively. The predicted value of the coal preparation efficiency is converted into an efficiency evaluation index within a set interval, so that the efficiency evaluation index is positively correlated with the predicted value of the coal preparation efficiency, i.e. the higher the predicted value of the coal preparation efficiency, the higher the efficiency evaluation index. Similarly, the predicted value of the product quality is converted into a quality evaluation index within a set interval, so that the quality evaluation index is positively correlated with the predicted value of the product quality.

[0111] A multi-objective evaluation function is constructed based on the efficiency target weight, the quality target weight, the efficiency evaluation index and the quality evaluation index. The output of the multi-objective evaluation function is the product of the efficiency evaluation index and the efficiency target weight plus the product of the quality evaluation index and the quality target weight. For example, if the efficiency evaluation index is E, the efficiency target weight is W1, the quality evaluation index is Q, and the quality target weight is W2, the value of the multi-objective evaluation function is E multiplied by W1 plus Q multiplied by W2. Through the multi-objective evaluation function, the evaluation scores of the initial process adjustment schemes are calculated, which comprehensively considers the importance of the coal preparation efficiency and the product quality under the current production demand, and provides a quantitative evaluation standard for selecting the optimal scheme.

[0112] Step S149: Select the initial process adjustment scheme with the highest evaluation score as the coal preparation process adjustment scheme. The coal preparation process adjustment scheme includes specific adjustment values of each key process parameter and an adjustment sequence.

[0113] By way of example only, the evaluation scores of all initial process adjustment schemes calculated by the multi-objective evaluation function are compared. The initial process adjustment scheme with the highest evaluation score is selected as the final coal preparation process adjustment scheme. The coal preparation process adjustment scheme includes specific adjustment values of each key process parameter and an adjustment sequence determined according to the previously determined parameter adjustment primary and secondary sequence. For example, the coal preparation process adjustment scheme can specify adjusting the separation medium density to a specific value first, and then adjusting other key process parameters to corresponding specific values in a certain sequence. By selecting the scheme with the highest evaluation score, the coal preparation process adjustment scheme can maximize the optimization of coal preparation efficiency and product quality under current production requirements, and effectively improve the coal preparation production process.

[0114] Step S150: Send the coal preparation process adjustment scheme to the coal preparation production control system for execution, and receive the execution result data returned by the coal preparation production control system. Based on the execution result data, optimize the dynamic mapping relationship and the coordination constraint condition.

[0115] By way of example only, the coal preparation process adjustment scheme generated earlier is sent to the coal preparation production control system for execution, and the dynamic mapping relationship and coordination constraint condition established earlier are optimized according to the execution result data returned by the system, forming a closed-loop optimization mechanism to continuously improve the accuracy of coal preparation decisions. In this embodiment, step S150 can include sub-steps S151-S158, which are described in detail below.

[0116] Step S151: Convert the separation equipment operation parameters, raw coal ratio parameters, and reagent addition parameters in the coal preparation process adjustment scheme into a control instruction format recognizable by the coal preparation production control system. The control instruction format includes parameter name, target adjustment value, adjustment rate, and execution timestamp.

[0117] By way of example only, the separation equipment operation parameters, raw coal ratio parameters, and reagent addition parameters in the coal preparation process adjustment scheme are processed. The parameters are converted into a control instruction format that can be recognized by the coal preparation production control system. For example, for the separation medium density adjustment value in the separation equipment operation parameters, it is converted into a control instruction format that includes the parameter name (such as "separation medium density"), the target adjustment value (the specific value to be adjusted to), the adjustment rate (the speed at which the adjustment is made), and the execution timestamp (the time at which the adjustment begins to be executed). Similarly, the raw coal ratio parameters and reagent addition parameters are also converted in a similar manner to ensure that the coal preparation production control system can accurately understand and execute the adjustment instructions.

[0118] Step S152: Send the control instruction to the central controller of the coal preparation production control system through the industrial communication network, trigger the central controller to execute the parameter adjustment operation.

[0119] For example, the converted control instruction is sent to the central controller of the coal preparation production control system through the industrial communication network. The industrial communication network ensures that the control instruction can be accurately and quickly transmitted to the central controller. For example, the control instruction containing the adjustment of the sorting equipment operation parameter, the raw coal proportioning parameter and the reagent adding parameter is sent to the central controller through the wired or wireless industrial communication network. After the central controller receives the control instruction, it triggers the corresponding parameter adjustment operation according to the parameter name, target adjustment value, adjustment rate and execution timestamp in the instruction, and starts to adjust the related parameters in the coal preparation production process.

[0120] Step S153: During the execution of the parameter adjustment operation, real-time receive the execution state data returned by the coal preparation production control system, which contains the adjusted equipment running state data, production process data and intermediate product quality detection data.

[0121] For example, during the execution of the parameter adjustment operation by the coal preparation production control system, real-time receive the execution state data returned by the system. The execution state data contains the adjusted equipment running state data, such as the running parameter change of the sorting equipment after adjusting the sorting medium density; production process data, such as whether the flow of raw coal in each processing link is changed due to parameter adjustment; intermediate product quality detection data, such as the quality index change of the intermediate product after a certain processing link. By real-time receiving the execution state data, the execution of the parameter adjustment operation in the production process can be understood in time.

[0122] Step S154: When the parameter adjustment operation is completed, receive the execution result data returned by the coal preparation production control system, which contains the adjusted coal preparation efficiency actual value, product quality actual value and equipment energy consumption data.

[0123] For example, when the parameter adjustment operation is completed, receive the execution result data returned by the coal preparation production control system. The execution result data contains the adjusted coal preparation efficiency actual value, i.e. the achieved coal preparation efficiency in the actual coal preparation process; the product quality actual value, i.e. the actual quality index of the final product; and the equipment energy consumption data, i.e. the energy consumption of the equipment in the running process after the parameter adjustment. The execution result data reflects the implementation effect of the coal preparation process adjustment scheme, and provides data for further analysis and optimization.

[0124] Step S155: Calculate the efficiency deviation between the actual value of the coal preparation efficiency in the execution result data and the predicted value of the coal preparation efficiency output by the decision influence evaluation model, and the quality deviation between the actual value of the product quality and the predicted value of the product quality.

[0125] For example, the efficiency deviation between the actual value of the coal preparation efficiency in the execution result data and the predicted value of the coal preparation efficiency output by the decision influence evaluation model is calculated by comparing the two values. For example, if the actual value of the coal preparation efficiency is a certain value and the predicted value of the coal preparation efficiency output by the decision influence evaluation model is another value, the difference between the two values is calculated by a specific calculation method, which is the efficiency deviation. Similarly, the quality deviation is calculated by comparing the actual value of the product quality and the predicted value of the product quality. The deviation value reflects the difference between the prediction result of the decision influence evaluation model and the actual execution result.

[0126] Step S156: Compare the efficiency deviation and the quality deviation with the preset deviation threshold value, and if any deviation exceeds the corresponding threshold value, optimize the data association rules in the dynamic mapping relationship based on the execution result data and adjust the association weight of the association dimension.

[0127] For example, the efficiency deviation and the quality deviation calculated above are compared with the preset deviation threshold value. The preset deviation threshold value is set according to the actual needs of coal preparation and the acceptable error range. For example, if the preset efficiency deviation threshold value is a certain value, when the calculated efficiency deviation exceeds the threshold value, it means that there is a large difference between the prediction of the decision influence evaluation model on the coal preparation efficiency and the actual situation. At this time, the data association rules in the dynamic mapping relationship are optimized based on the execution result data. For example, the data matching method and matching accuracy of the offline basic data set and the online real-time data set in each association dimension are reviewed, and the association weight of the association dimension is adjusted according to the actual situation in the execution result data, so that the dynamic mapping relationship can more accurately reflect the data association in the actual production. Similarly, if the quality deviation exceeds the preset quality deviation threshold value, similar optimization operations are also performed.

[0128] Step S157: Regenerate the collaborative constraint condition based on the optimized dynamic mapping relationship, and adjust the constraint threshold value and the constraint priority of each constraint item in the collaborative constraint condition.

[0129] For example only, after optimizing the dynamic mapping relationship, the collaborative constraint conditions are regenerated based on the optimized dynamic mapping relationship. For example, according to the optimized data association rules, the matching threshold of offline data and online data under each association dimension is re-determined, and then the time collaborative constraint rules, process collaborative constraint rules and device collaborative constraint rules are regenerated. At the same time, according to the new situation reflected by the execution result data in the coal preparation process, the constraint threshold and constraint priority of each constraint item in the collaborative constraint conditions are adjusted. For example, if the execution result data shows that a production link is more critical to product quality, the priority of the constraint item related to the link is correspondingly increased, and its constraint threshold is adjusted to be more stringent or more lenient to adapt to the actual production demand, so as to ensure that the collaborative constraint conditions can better guarantee the stability of the coal preparation process and the product quality.

[0130] Step S158: The optimized dynamic mapping relationship and the collaborative constraint conditions are used for decision impact assessment of the next round of coal preparation process parameters, forming a closed-loop optimization mechanism to continuously improve the decision accuracy.

[0131] For example only, the optimized dynamic mapping relationship and the collaborative constraint conditions are applied to the decision impact assessment process of the next round of coal preparation process parameters. In the next round of decision-making process, according to the previous steps, the data connection is established based on the optimized dynamic mapping relationship, the decision impact assessment model of the coal preparation process parameters is constructed according to the regenerated collaborative constraint conditions, and then the coal preparation process adjustment scheme is generated. In this way, a closed-loop optimization mechanism is formed. With the feedback of the execution result data in each round of coal preparation process and the optimization of the dynamic mapping relationship and the collaborative constraint conditions, the accuracy of the coal preparation decision can be continuously improved, and the coal preparation process can be continuously optimized to improve the coal preparation efficiency and product quality.

[0132] Figure 2 A schematic diagram of exemplary hardware and software components of the coal preparation decision system 100 based on offline and online data fusion that can implement the inventive concept provided by some embodiments of the present application is shown. For example, the processor 120 can be used in the coal preparation decision system 100 based on offline and online data fusion, and used to perform the functions in the present application.

[0133] The coal preparation decision system 100 based on offline and online data fusion can be a general server or a special-purpose server, both of which can be used to implement the method of the present application. Although only one server is shown in the present application, for the sake of convenience, the functions described in the present application can be implemented in a distributed manner on multiple similar platforms to balance the processing load.

[0134] For example, the coal preparation decision system 100 based on offline and online data fusion can include a network port 110 connected to a network, one or more processors 120 for executing program instructions, a communication bus 130, and different forms of storage media 140, such as a disk, a ROM, or a RAM, or any combination thereof. Illustratively, the coal preparation decision system 100 based on offline and online data fusion can also include program instructions stored in a ROM, a RAM, or other types of non-transitory storage media, or any combination thereof. The method of the present application can be implemented according to the above program instructions. The coal preparation decision system 100 based on offline and online data fusion also includes an input / output (I / O) interface 150 between the computer and other input / output devices.

[0135] For ease of illustration, only one processor is described in the coal preparation decision system 100 based on offline and online data fusion. However, it should be noted that the coal preparation decision system 100 based on offline and online data fusion in the present application can also include multiple processors, so the steps performed by one processor described in the present application can also be jointly performed or separately performed by multiple processors. For example, if the processor of the coal preparation decision system 100 based on offline and online data fusion performs steps A and B, it should be understood that steps A and B can also be jointly performed by two different processors or separately performed in one processor. For example, a first processor performs step A, a second processor performs step B, or the first processor and the second processor jointly perform steps A and B.

[0136] In summary, the coal selection decision-making method and system based on offline and online data fusion provided by the embodiments of the present application avoid the problem of separate processing of two types of data by establishing a dynamic mapping relationship between the offline basic data set and the online real-time data set, realizing the association of different source data. The dynamic mapping relationship can be adaptively adjusted according to the changes of online data, ensuring that the offline data and online data always maintain effective matching, providing a reliable data basis for the generation of subsequent collaborative constraint conditions and the construction of a decision impact evaluation model. In addition, the cross-data type collaborative constraint conditions generated based on the dynamic mapping relationship standardize the association and consistency requirements of different source data, avoiding analysis bias caused by data differences. Under the support of the collaborative constraint conditions, the constructed coal selection process parameter decision impact evaluation model can comprehensively and accurately analyze the associated influence of process parameter adjustment on coal selection efficiency and product quality, covering direct and indirect influences, making the evaluation results more valuable. At the same time, the coal selection process adjustment scheme generated according to the decision impact evaluation model includes adjustment strategies for sorting equipment operation parameters, raw coal ratio parameters and reagent addition parameters, and the scheme generation process comprehensively considers multi-objective evaluation, which can balance the demand for coal selection efficiency and product quality, meet the decision requirements in different production scenarios. Finally, the result data after the execution of the adjustment scheme is fed back to optimize the dynamic mapping relationship and collaborative constraint conditions, forming a closed-loop optimization mechanism, so that the entire decision-making method can continuously adapt to production changes, continuously improve the decision accuracy and reliability, and provide efficient and high-quality decision support for coal selection production in the long term, significantly improving the overall efficiency of coal selection production.

[0137] It should be noted that, in order to simplify the expression of the disclosure of the present application and help understand one or more embodiments of the present application, in the foregoing description of the embodiments of the present application, various features are sometimes combined into one embodiment, figure or description thereof.

[0138] In the above embodiments, the description of each embodiment focuses on different aspects, and the parts not described in detail in a certain embodiment can be referred to the related description of other embodiments. The embodiments of the present application, the implementation manners and the related technical features can be combined, replaced or modified without conflict. The above is only a preferred embodiment of the present application, and does not limit the present application in any form. Any simple modification, equivalent change and modification made to the above embodiments according to the technical essence of the present application still falls within the scope of the technical solution of the present application.

Claims

1. A coal preparation decision-making method based on the fusion of offline and online data, characterized in that, The method includes: A dynamic mapping relationship is established between the offline basic data set and the online real-time data set of the coal preparation process. The offline basic data set includes historical production record data, laboratory analysis data, and equipment maintenance data, while the online real-time data set includes real-time sensor acquisition data, production process data, and equipment operating status data. Based on the dynamic mapping relationship, cross-data type collaborative constraints are generated; A decision impact assessment model for coal preparation process parameters is constructed using the aforementioned collaborative constraints. This model is used to analyze the correlation between process parameter adjustments and coal preparation efficiency and product quality. Based on the correlation impact results output by the decision impact assessment model, a coal preparation process adjustment plan is generated; The coal preparation process adjustment plan is sent to the coal preparation production control system for execution, and the execution result data returned by the coal preparation production control system is received. Based on the execution result data, the dynamic mapping relationship and the collaborative constraint conditions are optimized.

2. The coal preparation decision-making method based on offline and online data fusion according to claim 1, characterized in that, The establishment of a dynamic mapping relationship between the offline basic data set and the online real-time data set for the coal preparation process includes: Identify the production cycle characteristics of historical production record data, the indicator correlation characteristics of laboratory analysis data, and the fault mode characteristics of equipment maintenance data in the offline basic data set, and generate an offline data feature map. Identify the fluctuation characteristics of real-time sensor data, the link connection characteristics of production process data, and the parameter coupling characteristics of equipment operating status data in the online real-time data set, and generate an online data feature map. Based on the offline data feature map and the online data feature map, a cross-map association dimension is constructed. The cross-map association dimension includes the time association dimension of production cycle and fluctuation characteristics, the process association dimension of indicator association characteristics and link connection characteristics, and the equipment association dimension of failure mode characteristics and parameter coupling characteristics. Calculate the data similarity between the offline data feature map and the online data feature map in each cross-map association dimension, and generate a dimension similarity vector; Data association rules that determine dynamic mapping relationships based on the dimensional similarity vector are used to define the data matching methods and matching accuracy between offline and online data under different association dimensions. Based on the data association rules, a dynamic mapping relationship is established between the offline basic data set and the online real-time data set, and an adaptive update mechanism for the dynamic mapping relationship is set.

3. The coal preparation decision-making method based on offline and online data fusion according to claim 2, characterized in that, The process of identifying the production cycle characteristics of historical production record data, the indicator correlation characteristics of laboratory analysis data, and the failure mode characteristics of equipment maintenance data in the offline basic data set, and generating an offline data feature map, includes: The historical production record data is decomposed into a time series to extract the periodic and trend components in the data. Based on the periodic components, the cyclical pattern of historical production is determined, and production cycle characteristics are generated. Correlation analysis is performed on the various analytical indicators in the laboratory analysis data to identify positive and negative correlation pairs between the indicators. The correlation relationship and degree of correlation of the positive and negative correlation pairs are defined as the indicator correlation characteristics. Pattern recognition is performed on the fault records of the equipment maintenance data to extract the equipment status parameter sequence before the fault occurred and the maintenance measures after the fault occurred. The correspondence between the equipment status parameter sequence and the maintenance measures is defined as the fault mode feature. Feature identifiers are set for the production cycle features, indicator correlation features, and failure mode features respectively. The production cycle features, indicator correlation features, failure mode features and their corresponding feature identifiers are integrated according to a preset map structure to generate an offline data feature map. The offline data feature map is subjected to feature integrity verification to ensure that the offline data feature map contains all offline features of key links in coal preparation production.

4. The coal preparation decision-making method based on offline and online data fusion according to claim 1, characterized in that, The generation of cross-data type collaborative constraints based on the dynamic mapping relationship includes: The data association rules in the dynamic mapping relationship are analyzed, and the matching thresholds of offline data and online data under each association dimension are extracted. The matching thresholds include the maximum time deviation of the time association dimension, the index deviation range of the process association dimension, and the parameter matching degree threshold of the device association dimension. Based on the maximum time deviation of the time correlation dimension, a time coordination constraint rule is generated. The time coordination constraint rule is used to limit the difference between the time stamp of offline production cycle data and the data timestamp of online real-time data. Based on the deviation range of the indicators of the process association dimension, process collaboration constraint rules are generated. These rules are used to constrain the deviation between offline laboratory analysis indicators and online production process indicators to be within the allowable range. Based on the parameter matching degree threshold of the device association dimension, device collaboration constraint rules are generated. These rules define the minimum matching requirements between offline fault mode features and online device operating status parameters. The time coordination constraint rules, process coordination constraint rules, and device coordination constraint rules are logically combined to generate a coordination constraint condition containing multiple rules, and a constraint priority is set for each rule in the coordination constraint condition.

5. The coal preparation decision-making method based on offline and online data fusion according to claim 4, characterized in that, The generation of process collaboration constraint rules based on the deviation range of the indicators in the process correlation dimension includes: Analyze the data association rules of the process association dimension in the dynamic mapping relationship, and extract the index deviation range containing the absolute deviation allowable value and the relative deviation allowable value; Based on the absolute deviation allowable value, an absolute deviation constraint sub-rule is constructed to limit the actual numerical difference between offline laboratory analysis indicators and online production process indicators; Based on the relative deviation allowable value, a relative deviation constraint sub-rule is constructed and used to limit the ratio deviation between offline laboratory analysis indicators and online production process indicators; Identify key quality control points in the coal preparation production process and set deviation weighting coefficients for the indicators corresponding to the key quality control points; The absolute deviation constraint sub-rules and the relative deviation constraint sub-rules are weighted and fused with the deviation weighting coefficients to generate weighted deviation constraint sub-rules; Based on the quality grade requirements of coal preparation products, a grade correction coefficient is set for the weighted deviation constraint sub-rule to generate a process collaborative constraint rule that includes quality grade differences, which is then used to match the quality control requirements of different grade products.

6. The coal preparation decision-making method based on offline and online data fusion according to claim 1, characterized in that, The decision impact assessment model for coal preparation process parameters constructed through the aforementioned collaborative constraints includes: Based on the coal preparation process flow, determine the adjustable process parameters that affect coal preparation efficiency and product quality, and generate an initial process parameter set; Based on the process coordination constraint rules in the aforementioned coordination constraint conditions, the initial process parameter set is filtered to obtain adjustable key process parameters; Collect historical adjustment records of the key process parameters in the offline basic data set, as well as the corresponding historical coal preparation efficiency and historical product quality data, and generate a historical process result dataset. The historical process result dataset is associated with the real-time production data in the online real-time dataset through the dynamic mapping relationship, generating a fused process result dataset that includes offline historical experience and online real-time status. Based on the fused process result dataset, the causal inference method is used to analyze the causal relationship between key process parameters and coal preparation efficiency and product quality, and to identify direct and indirect causal paths. A multi-factor influence network is constructed based on the direct and indirect causal paths. The nodes in the multi-factor influence network are key process parameters, coal preparation efficiency, and product quality, and the edges represent causal influence relationships and influence intensity. The equipment coordination constraint rules and time coordination constraint rules in the coordination constraint conditions are incorporated into the multi-factor influence network as network constraint conditions to generate a decision influence assessment model for coal preparation process parameters.

7. The coal preparation decision-making method based on offline and online data fusion according to claim 6, characterized in that, Based on the fused process result dataset, a causal inference method is used to analyze the causal relationship between key process parameters and coal preparation efficiency and product quality, identifying direct and indirect causal paths, including: The key process parameters, coal preparation efficiency, and product quality data in the fusion process result dataset are standardized. An initial structure for a causal relationship graph is constructed based on standardized variables. The initial structure includes potential correlation edges between each key process parameter and coal preparation efficiency and product quality. The initial structure is tested for causal relationships using a causal inference algorithm, and the causal effect value of each potential associated edge is calculated. The causal effect value is used to represent the degree of influence of the causal variable on the outcome variable. Based on the significance level of the causal effect value, significant correlation edges are selected, and correlation edges without significant causal effects are removed to generate a simplified causal relationship graph. In the simplified causal relationship diagram, the associated edges that directly point from key process parameters to coal preparation efficiency or product quality are identified, and these associated edges are defined as direct causal paths. Identify the sequence of related edges that indirectly point from each key process parameter to coal preparation efficiency or product quality through other key process parameters, and define the sequence of related edges as an indirect causal path; The total causal effect of the indirect causal path is calculated, and causal direction identifiers and causal effect value labels are added to the direct and indirect causal paths to form a causal relationship network containing path features.

8. The coal preparation decision-making method based on offline and online data fusion according to claim 1, characterized in that, The step of generating a coal preparation process adjustment plan based on the correlation impact results output by the decision impact assessment model includes: The correlation impact results output by the decision impact assessment model are analyzed, and the impact coefficients of each key process parameter on coal preparation efficiency and product quality through direct and indirect causal paths are extracted. The comprehensive impact index of each key process parameter is calculated based on the aforementioned impact coefficients. The comprehensive impact index is the weighted sum of the direct impact coefficients and the indirect impact coefficients. The key process parameters are ranked according to the comprehensive impact index to determine the order of parameter adjustments. Based on the process coordination constraint rules in the aforementioned coordination constraint conditions, the allowable adjustment range of each key process parameter is determined, and the allowable adjustment range corresponds to the index deviation range. Within the allowable adjustment range, multiple candidate adjustment values ​​are generated for each key process parameter, and the interval between the candidate adjustment values ​​is determined according to the sensitivity of the parameter; Candidate adjustment values ​​of different key process parameters are combined to generate multiple initial process adjustment schemes. Each initial process adjustment scheme contains a complete set of parameter adjustment combinations. Input the initial process adjustment scheme into the decision impact assessment model to obtain the predicted coal preparation efficiency and product quality values ​​corresponding to each scheme. A multi-objective evaluation function containing efficiency target weights and quality target weights is constructed, and the evaluation score of each initial process adjustment scheme is calculated based on the predicted coal preparation efficiency value and the predicted product quality value. The initial process adjustment scheme with the highest evaluation score is selected as the coal preparation process adjustment scheme. The coal preparation process adjustment scheme includes the specific adjustment values ​​and adjustment order of each key process parameter.

9. The coal preparation decision-making method based on offline and online data fusion according to claim 1, characterized in that, The step of sending the coal preparation process adjustment plan to the coal preparation production control system for execution, and receiving the execution result data returned by the coal preparation production control system, and optimizing the dynamic mapping relationship and the collaborative constraint conditions based on the execution result data, includes: The operating parameters of the sorting equipment, the raw coal ratio parameters, and the reagent addition parameters in the coal preparation process adjustment scheme are converted into a control instruction format that can be recognized by the coal preparation production control system. The control instruction format includes parameter name, target adjustment value, adjustment rate, and execution timestamp. The control commands are sent to the central controller of the coal preparation production control system via an industrial communication network, triggering the central controller to perform parameter adjustment operations. During the parameter adjustment operation, the execution status data returned by the coal preparation production control system is received in real time. The execution status data includes the adjusted equipment operating status data, production process data, and intermediate product quality inspection data. After the parameter adjustment operation is completed, the execution result data returned by the coal preparation production control system is received. The execution result data includes the actual value of the adjusted coal preparation efficiency, the actual value of the product quality, and the equipment energy consumption data. Calculate the efficiency deviation between the actual coal preparation efficiency and the predicted coal preparation efficiency output by the decision impact assessment model in the execution result data, and the quality deviation between the actual product quality and the predicted product quality. The efficiency deviation and quality deviation are compared with preset deviation thresholds. If either deviation exceeds the corresponding threshold, the data association rules in the dynamic mapping relationship are optimized based on the execution result data, and the association weight of the association dimension is adjusted. Based on the optimized dynamic mapping relationship, the collaborative constraints are regenerated, and the constraint thresholds and priorities of each constraint term in the collaborative constraints are adjusted. The optimized dynamic mapping relationship and collaborative constraints will be used to assess the impact of the next round of coal preparation process parameters on decision-making.

10. A coal preparation decision-making system based on the fusion of offline and online data, characterized in that, The method includes a processor and a machine-readable storage medium connected to the processor. The machine-readable storage medium is used to store programs, instructions, or code, and the processor is used to execute the programs, instructions, or code in the machine-readable storage medium to implement the coal preparation decision method based on offline and online data fusion as described in any one of claims 1-9.

Images