A human-model-platform integrated approach to tailings design process

Abstract

This invention discloses an integrated human-model-platform method for tailings design, relating to the field of digital collaborative processing technology. The method includes receiving basic tailings design data and task objectives through an industrial cloud platform, extracting, classifying, and associating tailings design constraints to form a tailings design intent list; extracting matching solution support information by connecting a knowledge base, a standard base, and a case base along the constraint order and professional orientation corresponding to the tailings design intent list; and using a solution deduction model to perform solution deduction based on the solution support information to form a candidate solution set, which is then sent to the adjudication link. This invention improves the continuity of constraint transmission and the systematic nature of candidate solution formation by extracting, classifying, and associating tailings design constraints from basic tailings design data through an industrial cloud platform and combining them with solution support information for solution deduction.

Description

Technical Field

[0001] This invention relates to the field of digital collaborative processing technology, and in particular to an integrated human-model-platform method for tailings design process. Background Technology

[0002] With the continuous advancement of digitalization in mining engineering and the application of industrial cloud platforms, the tailings design process is evolving from decentralized manual processing to platform-based, collaborative, and intelligent processing. Around the aspects of tailings dam layout, water conveyance and return connection, safety control, professional collaboration, and output, related technologies have gradually acquired the capabilities of data management, knowledge retrieval, standard retrieval, case reuse, and output generation, enabling tailings design to begin to show a development trend of full-process organization and multi-professional linkage.

[0003] Existing methods have some shortcomings. Tailings design and processing methods relying on industrial cloud platforms mainly rely on data retrieval, clause searching, or content generation. It is difficult to uniformly extract, classify, and associate constraint information from multi-source tailings design basic data around the task objectives. This can easily lead to scattered access to supporting information, discontinuous transmission of constraints, and unclear hierarchy of candidate solutions. In addition, related technologies lack an orderly adjudication mechanism for key differences such as technology comparison, parameter selection, process disagreement, and risk assessment. It is difficult to stably correspond the differences, related evidence, and adjudication results, resulting in unclear boundaries between adopted and unadopted content, insufficient traceability of evidence, and weak consistency of results. Summary of the Invention

[0004] In view of the aforementioned existing problems, the present invention is proposed.

[0005] Therefore, this invention provides a human-model-platform integrated method for tailings design process to solve the problems of insufficient organization of constraint information and insufficient orderliness of collaborative decision-making.

[0006] To solve the above-mentioned technical problems, the present invention provides the following technical solution: This invention provides a human-model-platform integrated method for tailings design, comprising: receiving basic tailings design data and task objectives through an industrial cloud platform, extracting, classifying, and associating tailings design constraints to form a tailings design intent list; extracting matching scheme support information by connecting a knowledge base, a standard base, and a case base along the constraint order and professional orientation corresponding to the tailings design intent list, and performing scheme deduction by a scheme deduction model in conjunction with the scheme support information to form a candidate scheme set, and sending the candidate scheme set into the adjudication link; performing collaborative adjudication on the technical comparison nodes, parameter selection nodes, process divergence nodes, and risk assessment nodes within the candidate scheme set in the adjudication link, and backfilling the adjudication results into the candidate scheme set, consolidating unadopted content, and retaining adopted content and related evidence to form an execution plan; and using the industrial cloud platform to organize and generate tailings design specifications, process connection diagrams, building connection diagrams, and material statistics along the process relationships, professional relationships, and basis relationships determined in the execution plan, and writing the basis identifiers in the execution plan into the output results to form an integrated tailings design result.

[0007] As a preferred embodiment of the human-model-platform integrated method for tailings design process described in this invention, the specific steps for forming the tailings design intent list are as follows: The industrial cloud platform receives basic tailings design data and task objectives, and collects and organizes the basic tailings design data to form a basic record set. Extract tailings design constraints related to the task objectives from the basic record set, and classify the tailings design constraints to form a tailings design constraint set. The various items in the tailings design constraint set are correlated with the task objectives, and then organized according to the constraint order and professional orientation to form a tailings design intent list.

[0008] As a preferred embodiment of the human-model-platform integrated method for tailings design process described in this invention, the specific steps of forming a candidate solution set and sending the candidate solution set into the adjudication link are as follows: Following the constraint order and professional orientation corresponding to the tailings design intent list, the items in the tailings design intent list are sequentially expanded to form an extraction order. The knowledge base, standard base, and case base are connected according to the extraction order. Relevant content that matches each item in the tailings design intent list is extracted and the relevant content is organized into scheme support information that corresponds to each item in the tailings design intent list. A scheme deduction model is formed by constructing an input correspondence layer, a sequential arrangement layer, a fusion deduction layer, and a result processing layer. The tailings design intent list and scheme support information are fed into the scheme simulation model to simulate the scheme and generate candidate branch results. The candidate branch results are merged and organized according to process relationship, professional relationship and basis relationship to form a candidate solution set, and the candidate solution set is sent to the decision link.

[0009] As a preferred embodiment of the human-model-platform integrated method for tailings design process described in this invention, the specific steps for forming the execution plan are as follows: Extract technology comparison nodes, parameter selection nodes, process disagreement nodes, and risk assessment nodes from the candidate solution set, and organize each node into a set of nodes to be decided. The set of adjudication nodes is arranged according to the constraint order and professional orientation corresponding to the tailings design intent list, forming an adjudication sequence; The participants in the organizational design process execute collaborative decisions along the decision sequence, performing decisions on each node in the set of nodes to be decided in the decision chain, to form a decision result. The ruling results are then filled into the corresponding content in the candidate solution set. The unadopted content in the candidate solution set is collected, and the adopted content and related evidence are retained to form an implementation plan.

[0010] As a preferred embodiment of the human-model-platform integrated method for tailings design process described in this invention, the specific steps for forming the integrated tailings design result are as follows: The process relationships, professional relationships, and basis relationships in the implementation plan are mapped to the tailings design specifications, process connection diagrams, building connection diagrams, and material statistics to form an output organization list; Extract the corresponding content from the implementation plan along the results organization list, and organize the corresponding content into tailings design description, process connection diagram, building connection diagram and material statistics to form a set of output results; Write the basis identifiers in the execution plan into the corresponding positions in the output result set, and merge and organize the output result set to form an integrated tailings design result.

[0011] As a preferred embodiment of the human-model-platform integrated method for tailings design process described in this invention, the specific steps for forming a scheme deduction model by constructing an input correspondence layer, a sequential arrangement layer, a fusion deduction layer, and a result processing layer are as follows: Each tailings design intent in the tailings design intent list is matched with the supporting information of the scheme to form an input correspondence layer; The corresponding contents in the input layer are arranged according to the constraint order and professional orientation to form a sequential arrangement layer; The corresponding items in the sequential arrangement layer are continuously deduced to form a fusion deduction layer; The different inference contents formed by the fusion inference layer are merged and identified to form the result processing layer; The input correspondence layer serves as the basis for the arrangement of the sequential arrangement layer, the sequential arrangement layer serves as the basis for the deduction of the fusion deduction layer, and the fusion deduction layer serves as the basis for the organization of the result organization layer. The input correspondence layer, the sequential arrangement layer, the fusion deduction layer, and the result organization layer are connected sequentially to form a scheme deduction model.

[0012] As a preferred embodiment of the human-model-platform integrated method for tailings design process described in this invention, the adjudication link refers to an ordered processing path formed after the candidate scheme set is formed, which takes into account the differences, related basis and corresponding relationships in the candidate scheme set, and organizes the subsequent adjudication processing in sequence according to the constraint order and professional orientation in the tailings design intent list.

[0013] As a preferred embodiment of the human-model-platform integrated method for tailings design process described in this invention, the organizational design participants perform collaborative adjudication along the adjudication sequence at each node in the set of nodes to be adjudicated in the adjudication chain. The specific steps are as follows: The decision-making process involves locating each node to be decided sequentially along the decision-making chain, and retrieving the candidate direction, related basis, and decision result corresponding to the previous node to be decided for each node. The design participants are determined according to the professional orientation of each node to be adjudicated, and the design participants are responsible for comparing and judging the differences. The content to be retained and the content not adopted are determined according to the node type of the node to be decided, a decision result is formed, and the decision result is passed to the next node to be decided along the decision sequence.

[0014] As a preferred embodiment of the human-model-platform integrated method for tailings design process described in this invention, the basis identifier refers to the basis source category identifier, the basis source file identifier, the basis location identifier, and the corresponding decision node identifier.

[0015] As a preferred embodiment of the human-model-platform integrated method for tailings design process described in this invention, the step of writing the basis identifier in the execution plan into the corresponding position in the output result set refers to writing the basis source category identifier, basis source document identifier, basis location identifier, and corresponding decision node identifier into the corresponding position in the output result set according to the identifier relationship between each corresponding position in the result organization list and the basis identifier, so that the explanatory content in the tailings design description, the illustrative content in the process connection diagram, the illustrative content in the building connection diagram, and the statistical content in the material statistics content are respectively aligned with the corresponding basis identifier.

[0016] The beneficial effects of this invention are as follows: By extracting, classifying, and associating tailings design constraints from the basic tailings design data through an industrial cloud platform, a tailings design intent list is formed. Combined with supporting information, scheme deduction is then performed, solving the problem of insufficient constraint information organization and improving the continuity of constraint transmission and the systematic nature of candidate scheme formation. Furthermore, by constructing an adjudication link through the industrial cloud platform, collaborative adjudication is performed on discrepancies in candidate schemes, and the adjudication results are backfilled into the candidate scheme set. Unadopted content is consolidated, while adopted content and related evidence are retained, solving the problem of insufficient orderliness in collaborative adjudication. This improves the accuracy of scheme formation, the clarity of evidence traceability, and the consistency of output results. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 A flowchart of a human-model-platform integrated approach for tailings design process.

[0019] Figure 2 Create a flowchart for the tailings design intent list.

[0020] Figure 3 A flowchart for forming a set of candidate solutions.

[0021] Figure 4 A flowchart for generating implementation plans and organizing results.

[0022] Figure 5 This is a comparison chart of performance metrics data for different task sample groups.

[0023] Figure 6 Heatmaps of performance metrics for different task complexities and solution types. Detailed Implementation

[0024] To make the above-mentioned objects, features and advantages of the present invention more apparent and understandable, the specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings.

[0025] Many specific details are set forth in the following description in order to provide a full understanding of the invention. However, the invention may also be practiced in other ways different from those described herein, and those skilled in the art can make similar extensions without departing from the spirit of the invention. Therefore, the invention is not limited to the specific embodiments disclosed below.

[0026] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The phrase "in one embodiment" appearing in different places in this specification does not necessarily refer to the same embodiment, nor is it a single or selective embodiment that is mutually exclusive with other embodiments.

[0027] Reference Figures 1-6 As one embodiment of the present invention, this embodiment provides a human-model-platform integrated method for tailings design process, including the following steps: S1. Receive basic tailings design data and task objectives through the industrial cloud platform, and extract, classify and associate tailings design constraints to form a tailings design intent list.

[0028] S1.1 Receive tailings design basic data and task objectives through the industrial cloud platform, and collect and organize the tailings design basic data to form a set of basic records corresponding to the task objectives.

[0029] It should be noted that the industrial cloud platform receives the basic tailings design data and task objectives corresponding to the current tailings design task. The basic tailings design data includes the design commission, test report, geological survey report, historical scheme drawings, and calculation tables, equipment lists, material statistics tables, and specifications related to the current tailings design task. The task objectives include the target content extracted from the design commission requirements, deliverable requirements, and the current processing scope. The correspondence between the basic tailings design data and the task objectives is verified. The basic tailings design data belonging to the same task objective are grouped into the same processing scope and collected according to the data source and content form. The text content, table content, and drawing annotations are organized into sequentially readable record items. The correspondence between each record item and the original basic tailings design data is retained to form a basic record set corresponding to the task objectives.

[0030] S1.2 Extract tailings design constraint elements related to the task objectives from the basic record set, and classify the tailings design constraint elements to form a tailings design constraint element set.

[0031] It should be noted that, within the basic record set, records directly related to the task objective and capable of characterizing the tailings design task's scope, boundaries, conditions, and requirements are located, and these records are extracted as tailings design constraint elements. These tailings design constraint elements are then categorized according to their scope, boundaries, conditions, and requirements. Tailings design constraint elements with consistent content, identical target objects, or those jointly defining the same task objective are grouped into the same category. Simultaneously, the correspondence between each tailings design constraint element and its corresponding record in the basic record set is preserved, forming a set of tailings design constraint elements corresponding to the task objective, expressed as: ; in, Indicates the relationship with the first The set of tailings design constraints corresponding to each task objective; Indicates the first Tailings design constraints; Indicates the first One basic record item; Indicates the relationship with the first The basic record set corresponding to the content of each task objective; Indicates the basic record item Design constraints for tailings Extraction marker, when the underlying record item There exist elements that can characterize the design constraints of tailings. When recording content, Otherwise, it is 0; An index representing the content of the task objective; Index representing the design constraints of tailings; Indicates the index of the underlying record item.

[0032] S1.3. Associate each item in the tailings design constraint set with the task objectives, and organize them according to the constraint order and professional orientation to form a tailings design intent list.

[0033] It should be noted that each tailings design constraint element in the tailings design constraint element set is matched with the target content in the task objectives one by one to determine the limiting relationship between each tailings design constraint element and each target content, and to establish corresponding associations for tailings design constraint elements that jointly limit the same target content; the tailings design constraint elements are arranged in order of their restriction on the task objectives to form a constraint order; the constraint order is determined according to logical dependencies. When a tailings design constraint element constitutes the input boundary, standard premise, safety premise, layout premise, or calculation premise of the corresponding task objective content, this tailings design constraint element is arranged first; for tailings design constraint elements that do not have a direct dependency relationship, the processing boundary type content, the standard mandatory type content, and the safety control type content are arranged before the scheme selection type content and the parameter adjustment type content; for content that is still at the same level, it is maintained in the order of appearance in the basic record set.

[0034] Each tailings design constraint element is grouped and identified according to its professional category to form a professional orientation. The professional orientation is the subsequent processing assignment determined by the professional category to which the tailings design constraint element belongs. It represents the professional processing link that the tailings design constraint element should enter in the subsequent process of extracting supporting information for the scheme, deducing the scheme, making collaborative decisions and organizing results. The professional categories include process disciplines, general layout or layout disciplines, building structure disciplines and comprehensive verification disciplines. The professional categories represent the professional classification to which the tailings design constraint element belongs.

[0035] The corresponding association results with constraint order and professional orientation are organized sequentially into a tailings design intent list, expressed as: ; in, This indicates the tailings design list; Indicates the first The content of each task objective; Representing tailings design constraints With the content of the task objectives The corresponding associated markers, when tailings design constraints Regarding the content of the task objectives When forming a limiting relationship, Otherwise, it is 0; Representing tailings design constraints Regarding the content of the task objectives The order in which constraints are formed, i.e., the constraint sequence; Representing tailings design constraints Regarding the content of the task objectives The professional category to which it belongs, i.e., the professional orientation; This operation involves sorting according to the constraint order, and then merging and arranging the contents that are in the same constraint order according to their professional orientation.

[0036] S2. Following the constraint order and professional orientation corresponding to the tailings design intent list, connect the knowledge base, standard base and case base to extract matching scheme support information, and use the scheme deduction model to combine the scheme support information to deduce the scheme, form a candidate scheme set, and send the candidate scheme set into the adjudication link.

[0037] S2.1. Following the constraint order and professional orientation corresponding to the tailings design intent list, the items in the tailings design intent list are sequentially expanded to form the extraction order corresponding to the tailings design intent list.

[0038] It should be noted that the tailings design intent items with constraints and professional orientations in the tailings design intent list are expanded sequentially and arranged in order of constraint, so that the tailings design intent items in the earlier positions precede the tailings design intent items in the later positions. The tailings design intent items in the same constraint order are marked side by side according to professional orientation, thereby clarifying the sequential position and professional orientation of each tailings design intent item in the subsequent extraction process. After completing the sequential arrangement and side by side marking, the tailings design intent items are organized into an extraction order that maintains a one-to-one correspondence with the tailings design intent list.

[0039] S2.2 Connect the knowledge base, standard base and case base according to the extraction order, extract relevant content that matches each item in the tailings design intent list, and organize the relevant content into scheme support information that corresponds to each item in the tailings design intent list.

[0040] It should be noted that the knowledge base, standard library, and case library are pre-stored in the industrial cloud platform. The knowledge base is formed by compiling and organizing historical project data, technical data, and reusable experience; the standard library is formed by compiling and organizing standard data related to the current tailings design task; and the case library is formed by compiling and organizing historical scheme drawings and historical results. The knowledge base, standard library, and case library are sequentially connected according to the order of extraction of tailings design intent content and their professional focus. In the knowledge base, knowledge content matching the corresponding tailings design intent content is located; in the standard library, standard content matching the corresponding tailings design intent content is located; and in the case library, case content matching the corresponding tailings design intent content is located. A one-to-one correspondence is established between the extracted knowledge content, standard content, and case content and the corresponding tailings design intent content. The knowledge content, standard content, and case content corresponding to the same tailings design intent content are merged, retaining the correspondence between each content and the knowledge base, standard library, case library, and tailings design intent list, forming scheme support information corresponding to each item in the tailings design intent list. The expression is: ; in, Indicates the relationship with the first Supporting information for each tailings design intent item; Indicates the first The design intent of the tailings dam; This represents all the knowledge content in the knowledge base; This represents all the specifications in the specification library; This represents all case content in the case library; This represents a single piece of knowledge content in the knowledge base. This represents a single specification item in the specification library; This represents the content of a single case in the case library; This indicates a matching tagging function that matches knowledge content, specification content, or case study content with tailings design intent. The function value is 1 if a match is found, and 0 otherwise. An index representing the design intent of the tailings.

[0041] S2.3. A scheme deduction model is formed by constructing an input correspondence layer, a sequential arrangement layer, a fusion deduction layer, and a result processing layer. The tailings design intent list and scheme support information are sent into the scheme deduction model to perform scheme deduction and form candidate branch results.

[0042] It should be noted that the scheme deduction model is a sequence deduction model based on neural networks. It is formed by constructing the tailings design intent list and scheme support information in layers. The scheme deduction model includes an input correspondence layer, a sequence arrangement layer, a fusion deduction layer and a result processing layer.

[0043] The tailings design intents in the tailings design intent list are linked to the corresponding knowledge content, standard content, and case content in the scheme support information. Each tailings design intent, corresponding knowledge content, corresponding standard content, and corresponding case content is organized into corresponding input units, and all corresponding input units constitute the input correspondence layer.

[0044] Based on the corresponding input units in the corresponding layer, the corresponding input units are arranged in front and behind according to the constraint order in the tailings design intent list. Then, the corresponding input units in the same constraint order are arranged side by side according to the professional orientation, so that the corresponding input units form a deduction sequence consistent with the tailings design intent list. The deduction sequence constitutes the sequential arrangement layer.

[0045] Based on the deduction sequence in the sequential arrangement layer, the deduction features formed by the corresponding input unit in the previous position are transferred to the corresponding input unit in the next position. The current corresponding input unit is then combined with the corresponding knowledge content, standard content, and case content for fusion processing. This allows parallel corresponding input units in the same constraint sequence to form different deduction branches. During the formation of each deduction branch, process relationships, professional relationships, and basis relationships are established. All deduction branches and their corresponding relationships constitute the fusion deduction layer.

[0046] The various deduction branches formed by the fusion deduction layer are grouped and identified according to the process relationship, professional relationship and basis relationship. Deduction branches that can be connected from beginning to end and have complete basis are grouped into the same candidate direction. Deduction branches with different deduction content or different basis are retained as different candidate directions. The candidate branch results corresponding to each item in the tailings design intent list are output. The result sorting layer is composed of each candidate direction and its corresponding candidate branch results.

[0047] The sequential arrangement layer is based on the item-by-item correspondence in the input correspondence layer. The sequential arrangement layer is based on the front-to-back arrangement and parallel arrangement relationship in the sequential arrangement layer. The different deduction contents formed by the fusion deduction layer are based on the result organization layer. The input correspondence layer, sequential arrangement layer, fusion deduction layer and result organization layer are connected in sequence to form a scheme deduction model.

[0048] The simulation model is trained, and the training process includes training sample construction, training sample organization, training execution, training verification, and training consolidation. Among them, training sample construction involves extracting sample content related to the tailings design task from historical project data, technical data, specification data, historical scheme drawings, and historical results. Candidate branch results that have been formed and adopted in historical projects are organized into supervision samples to ensure that the training samples maintain the correspondence between the tailings design intent list, scheme supporting information, and candidate branch results. Manual review and correction records are used to mark the difference location, difference type, and correction direction between candidate branch results and actual adopted results, which serve as the basis for deviation correction during the training process.

[0049] The training samples are cleaned up by removing duplicate, missing, erroneous, and irrelevant content, and the training samples are divided into training sample set, validation sample set, and test sample set.

[0050] Training is performed by feeding training samples batch by batch into the scheme deduction model to generate training candidate branches, and then updating the network parameters in the scheme deduction model according to the total loss function; the expression for the total loss function is: ; in, Represents the total loss function; This represents the branch generation loss, which characterizes the content difference between the training candidate branch results and the corresponding candidate branch results in the supervised samples; The relation consistency loss represents the degree of consistency between the process relations, professional relations, and basis relations in the training candidate branch results and the corresponding relations in the supervision samples. The alignment loss represents the degree of correspondence between the training candidate branch results and the supporting information of the scheme.

[0051] Training and validation involves feeding the validation sample set into the scheme deduction model after training is completed, performing accuracy, completeness, and correspondence checks on the candidate branch results corresponding to the validation sample set, and making a joint judgment based on the test results corresponding to the test sample set.

[0052] Training and solidification involves solidifying and saving the current parameter configuration after the scheme deduction model meets the requirements for generating candidate branch results. The actual adoption results generated in new projects and their corresponding difference markers are then added to the training samples for subsequent training processing. This ensures that the scheme deduction model maintains its scheme deduction capability consistent with the results called from the knowledge base, specification base, and case base.

[0053] When the tailings design intent list and supporting scheme information are fed into the scheme deduction model for scheme deduction, the input corresponding layer processes each tailings design intent content and the knowledge content, standard content, and case content in the supporting scheme information to form item-by-item corresponding input content. The sequence arrangement layer sorts the item-by-item corresponding input content according to the constraint order and arranges them in parallel according to professional orientation to form a deduction sequence. The fusion deduction layer processes each input content sequentially along the deduction sequence, passing the deduction content formed by the previous input content to the deduction process of the subsequent input content, and forming different deduction content for input content in the same constraint order position. The result processing layer organizes the different deduction content according to process relationship, professional relationship, and basis relationship, and outputs the candidate branch results corresponding to each item in the tailings design intent list. The expression is: ; , ; ; in, Indicates processing up to the th A set of deductive content formed when considering the design intent of a tailings mine; Indicates processing up to the th A set of deductive content formed when considering the design intent of a tailings mine; Indicates the first The design intent of the tailings dam; Indicates the relationship with the first Supporting information for each tailings design intent item; This refers to the calculation that connects and updates the previous simulation content with the current tailings design intent and corresponding supporting information to form new simulation content; Represents the initial set of deduction content, and It is an empty set; This represents the final set of deduced content formed after processing all tailings design intent content; Indicates the first Candidate branch results corresponding to each candidate direction; Indicates the content of the deduction Corresponding process relationship identifiers; Indicates the content of the deduction Corresponding professional relationship identifier; Indicates the content of the deduction Corresponding relation identifier; , , They represent the first Each candidate direction corresponds to a process relationship identifier, a professional relationship identifier, and a basis relationship identifier. Indicates the total number of candidate directions; This represents the set of candidate branch results.

[0054] S2.4. The candidate branch results are merged and organized according to process relationship, professional relationship and basis relationship to form a candidate solution set, and the candidate solution set is sent to the decision link.

[0055] It should be noted that, using candidate branch results as the merging object, the corresponding connections between the preceding and following content in each candidate branch result are checked according to the process relationship. Candidate branch results that can be connected sequentially are grouped into the same candidate direction. Candidate branch results grouped into the same candidate direction are merged in parallel according to professional relationship. Candidate branch results belonging to the same professional orientation and able to jointly represent the same tailings design intent are organized into corresponding candidate contents. The knowledge content, standard content, and case content in each corresponding candidate content are correspondingly identified according to the basis relationship. Candidate branch results with complete and consistent basis are retained in the same candidate direction. Candidate branch results with differences in basis or differences in deduction content are retained as different candidate directions, forming a candidate scheme set containing multiple candidate directions, expressed as: ; in, Represents the set of candidate solutions; This function represents a processing function that merges and organizes candidate branch results according to process relationships, professional relationships, and basis relationships.

[0056] The candidate options in the tailings design intent list are organized sequentially according to the constraints and professional orientations in the list, ensuring that the candidate options list corresponds to the tailings design intent list. The candidate options list is then sent to the adjudication link. The adjudication link is an ordered processing path formed after the candidate options list is formed. It is used to take into account the differences, related evidence, and their correspondences in the candidate options list, and to organize the subsequent adjudication processing in the order of constraints and professional orientations in the tailings design intent list. This ensures that the content to be adjudicated in the candidate options list maintains the connection, professional correspondence, and evidence correspondence in the subsequent processing.

[0057] It should also be noted that existing technologies typically generate complete solution outputs through data retrieval and matching or direct generation, which can easily lead to problems such as unclear process connections, mixed professional correspondences, insufficient supporting evidence, and unclear candidate result hierarchy. This solution extracts and organizes solution supporting information according to the constraint order and professional orientation of the tailings design intent list, and uses a solution deduction model to continuously deduce and merge to form a candidate solution set. This makes the candidate solutions more orderly, the differences clearer, the supporting evidence more complete, and improves the accuracy of subsequent decisions and the systematic nature of the solution output.

[0058] S3. In the adjudication chain, perform collaborative adjudication on the technology comparison nodes, parameter selection nodes, process disagreement nodes, and risk assessment nodes within the candidate solution set, and backfill the adjudication results into the candidate solution set. Unadopted content is collected, and adopted content and related evidence are retained to form an execution plan.

[0059] S3.1 Extract technology comparison nodes, parameter selection nodes, process disagreement nodes, and risk assessment nodes from the candidate solution set, and organize the extracted nodes into a set of nodes to be decided.

[0060] It should be noted that, according to the item-by-item correspondence in the tailings design intent list, each candidate direction corresponding to the same tailings design intent content in the candidate scheme is grouped into the same comparison range, and the process relationship, professional relationship and basis relationship of each candidate direction in the same comparison range are compared to determine the positions where there are differences in content or basis.

[0061] The differences at each location are categorized by type. When the difference lies in the technical route, processing method, or scheme organization, the corresponding location is extracted as a technology comparison node. When the difference lies in parameter selection, parameter boundaries, or parameter combination relationships, the corresponding location is extracted as a parameter selection node. When the difference lies in the connection between preceding and following processes, the arrangement of parallel processes, or the connection between specialties, the corresponding location is extracted as a process divergence node. When the difference lies in the content of safety control, the basis of risk control, or the direction of risk treatment, the corresponding location is extracted as a risk assessment node. All types of nodes are then grouped according to the item-by-item correspondence in the tailings design intent list and arranged in the order of constraints to form a set of nodes to be adjudicated.

[0062] S3.2 Arrange the set of nodes to be adjudicated according to the constraint order and professional orientation corresponding to the tailings design intent list to form an adjudication sequence.

[0063] It should be noted that the technology comparison nodes, parameter selection nodes, process disagreement nodes, and risk assessment nodes in the set of nodes to be decided are matched one by one with the tailings design intent contents in the tailings design intent list. The nodes are arranged in the order of constraints in the tailings design intent list so that each node corresponding to the preceding tailings design intent content is kept before each node corresponding to the following tailings design intent content.

[0064] Nodes in the same constraint order position are grouped together according to their professional orientation, so that nodes belonging to the same professional orientation are arranged together, and nodes with different professional orientations are arranged separately. Nodes in the same constraint order position and belonging to the same professional orientation are maintained in the set of nodes to be adjudicated, so as to clarify the sequential position of each node in the adjudication chain and form an adjudication sequence corresponding to the tailings design intent list.

[0065] S3.3. The participants in the organization design perform collaborative adjudication along the adjudication sequence to each node in the set of nodes to be adjudicated in the adjudication chain, forming an adjudication result.

[0066] It should be noted that the design participants are pre-configured in the industrial cloud platform. The design participants include process designers, general layout or arrangement designers, architectural structure designers, and reviewers. Along the adjudication sequence, each node to be adjudicated in the set of nodes to be adjudicated is located sequentially in the adjudication chain, and the candidate directions, related evidence, and existing adjudication results formed by the preceding nodes to be adjudicated are retrieved for each node to be adjudicated. According to the professional orientation of each node to be adjudicated, the design participants corresponding to the professional orientation are determined as the participants of the current node to be adjudicated, and the participants of the current node to be adjudicated compare and judge the differences.

[0067] If the current node awaiting adjudication is a technical comparison node, retrieve the relevant content and corresponding basis for tailings transportation, water return organization, flood discharge organization and building layout corresponding to different candidate directions, verify the completeness of the content with the task objectives, previous adjudication results and basis, retain the candidate content that meets the task objectives, is consistent with the previous and subsequent decisions and has complete basis, and determine the remaining content as unadopted content.

[0068] If the current node to be decided is a parameter selection node, retrieve the parameter selection, parameter boundaries, parameter combination relationships and corresponding basis for different candidate directions, check the compliance with the standard documents, case content and previously retained content, retain the parameter content that meets the standard requirements, corresponds to the previously retained content and will not cause subsequent conflicts, and determine the remaining content as unadopted content.

[0069] If the current node to be decided is a process divergence node, retrieve the corresponding process connections, parallel process arrangements, professional interface connections, and corresponding basis for different candidate directions, check the corresponding relationships in the tailings design description, process connection diagram, building connection diagram, and material statistics, retain the candidate content that is process-continuous, professionally clear, and based on complete basis, and determine the remaining content as unadopted content.

[0070] If the current pending decision node is a risk assessment node, retrieve the risk control content, risk source basis, risk impact location, and risk disposal content corresponding to different candidate directions, verify the correspondence with the normative materials, case content, and currently retained solutions, retain candidate content with clear basis, clear impact location, and disposal content consistent with the retained solutions, and determine the remaining content as unadopted content.

[0071] The retained content, unadopted content, participant identifiers, and associated evidence corresponding to the current pending node are combined to form the ruling result of the current pending node. The ruling result of the current pending node is then passed along the ruling sequence to the ruling process of the subsequent pending nodes, so that each pending node in the set of pending nodes forms a ruling result in sequence under the connection relationship.

[0072] S3.4. Fill the corresponding content in the candidate solution set with the adjudication results, consolidate the unadopted content in the candidate solution set, and retain the adopted content and related evidence to form an implementation plan.

[0073] It should be noted that, according to the order of each node to be decided in the decision sequence, the decision results formed by each node to be decided are backfilled into the candidate direction and corresponding candidate content in the candidate scheme set, so that the differences in each candidate direction correspond to the decision results.

[0074] The technical routes, processing methods, scheme organization methods, parameter selection, parameter boundaries, parameter combination relationships, process connection relationships, process arrangement relationships, professional connection relationships, safety control content, risk control basis and risk handling direction that are determined to be unadopted in the candidate scheme set will be removed and will be stopped from participating in subsequent combinations.

[0075] For candidate directions, corresponding candidate contents, and related supporting evidence that are determined to be retained, they are retained in their respective positions to maintain a correspondence between the retained contents and the related supporting evidence. Based on the item-by-item correspondence, constraint order, and professional orientation in the tailings design intent list, the retained candidate directions and corresponding candidate contents are checked for continuity and parallelism. Retained contents that are coherent and have complete supporting evidence are sequentially combined to form an execution plan that maintains a correspondence with the tailings design intent list. The expression is: ; in, Indicate the execution plan; Indicates the first The set of unadopted content corresponding to each node awaiting adjudication; Indicates the first The set of adopted content corresponding to each node awaiting adjudication; Indicates the total number of nodes pending adjudication; Indicates the function that forms the execution plan; This represents the index of the node to be adjudicated, with a value ranging from 1 to... .

[0076] Figure 5 This shows the changes in key performance indicators under different task sample groups; by Figure 5 As can be seen, all indicators remain at a high level overall, indicating that this embodiment can improve the continuity of constraint transmission, the systematic nature of candidate solution formation, the accuracy of execution solution formation, the clarity of basis tracing, and the consistency of output results; the magnified local area further shows that this embodiment still has good stability under local complex sample group conditions.

[0077] Figure 5 The curves in the figure represent the continuity rate of constraint transmission, the orderliness rate of candidate solutions, the accuracy rate of execution solution formation, the completeness rate of basis traceability, and the consistency rate of output results, respectively. They are used to characterize the degree of continuous transmission of constraint information, the clarity of candidate solution organization, the consistency between execution solution and standard solution, the completeness of basis identification, and the consistency between various output results.

[0078] It should also be noted that existing technologies typically determine the solution through manual comparison or simple screening, which is prone to problems such as inaccurate difference identification, unclear adjudication order, and insufficient retention of evidence. This solution extracts various nodes to be adjudicated and forms an adjudication sequence, performs collaborative adjudication along the adjudication link and backfills the adjudication results, consolidates unadopted content, and retains adopted content and related evidence, making difference positioning more accurate, adjudication process more orderly, and connection clearer, and improving the accuracy and systematicity of the implementation plan.

[0079] S4. Based on the technological, professional, and basis relationships determined in the implementation plan, use the industrial cloud platform to generate tailings design specifications, process connection diagrams, building connection diagrams, and material statistics. Write the basis identification in the implementation plan into the output results to form an integrated tailings design result.

[0080] S4.1. Map the technological relationships, professional relationships, and basis relationships in the implementation plan to the tailings design description, process connection diagram, building connection diagram, and material statistics to form an output organization list.

[0081] It should be noted that, according to the item-by-item correspondence in the tailings design intent list, each retained item and related basis in the implementation plan will be expanded item by item, and the corresponding positions of each retained item and related basis in the tailings design description, process connection diagram, building connection diagram, and material statistics will be determined. Specifically, the content representing the connection between the preceding and following processes will be mapped to the corresponding positions in the tailings design description and process connection diagram, the content representing the professional correspondence will be mapped to the corresponding positions in the building connection diagram, and the content related to material statistics will be mapped to the corresponding positions in the material statistics content.

[0082] The basis relationships corresponding to each retained item are organized into basis identifiers, which include basis source category identifiers, basis source document identifiers, basis location identifiers, and corresponding adjudication node identifiers. The basis source category identifier is used to indicate that the corresponding content comes from the knowledge base, standard base, case base, or adjudication chain. The basis source document identifier is used to indicate the source document to which the corresponding content belongs. The basis location identifier is used to indicate the position of the corresponding content in the clause, drawing, table, or record item of the source document. The corresponding adjudication node identifier is used to indicate the adjudication position of the corresponding content in the adjudication chain. The basis identifiers are then linked to their corresponding positions to establish an identifier relationship, so that the process relationship, professional relationship, and basis relationship in the implementation plan are respectively mapped to the tailings design description, process connection diagram, building connection diagram, and material statistics, forming a deliverable organization list.

[0083] S4.2 Extract the corresponding content from the execution plan along the results organization list, and organize the corresponding content into tailings design description, process connection diagram, building connection diagram and material statistics to form an output results set.

[0084] It should be noted that, according to the order of the corresponding positions in the outcome organization list, the retained content and related evidence matching each corresponding position are extracted from the implementation plan, and the extracted corresponding content is written item by item into the corresponding position indicated in the outcome organization list; among them, the corresponding content representing process connection and process treatment is organized sequentially into the explanatory content in the tailings design description and the graphic content in the process connection diagram; the corresponding content representing professional correspondence and the correspondence between buildings is organized sequentially into the graphic content in the building connection diagram; and the corresponding content representing material category, material attribution, and material correspondence is organized sequentially into the statistical content in the material statistics content. This ensures that each corresponding content maintains a consistent correspondence and arrangement with the outcome organization list in the tailings design description, process connection diagram, building connection diagram, and material statistics content. The resulting tailings design description, process connection diagram, building connection diagram, and material statistics content are then compiled into an output outcome set, expressed as: ; in, Indicates the first The class outputs a collection of results. Indicates the first in the execution plan Items to be retained; Reserved content The destination of the results; Indicates the output result category label, when At that time, in accordance with the tailings design specifications, when At that time, the corresponding process connection diagram, when At that time, the corresponding building connection diagram, when At that time, the corresponding material statistics content; Indicates an index for reserved content.

[0085] S4.3 Write the basis identifier in the execution plan into the corresponding position in the output result set, and merge and organize the output result set to form an integrated tailings design result.

[0086] It should be noted that, according to the identification relationship between the corresponding positions and the basis identifiers in the results organization list, the basis identifiers in the implementation plan are written into the corresponding positions in the output results set item by item, so that the explanatory content in the tailings design description, the graphic content in the process connection diagram, the graphic content in the building connection diagram, and the statistical content in the material statistics content are respectively aligned with the corresponding basis identifiers. After completing the writing of the basis identifiers, the consistency of the tailings design description, process connection diagram, building connection diagram, and material statistics content in the output results set is checked to ensure that the correspondence, constraint order, and professional orientation of each item are consistent with the tailings design intent list item by item, and the corresponding explanatory content, graphic content, statistical content, and basis identifiers are merged to form an integrated tailings design result.

[0087] Figure 6 The thermal distribution of key performance indicators is shown under different task complexities and scheme types; by Figure 6 As can be seen, under different task complexity conditions, the corresponding indicators of this embodiment are generally at a high level, indicating that this embodiment can maintain good constraint information organization ability, orderly collaborative decision-making and consistency of output results.

[0088] Figure 6 The vertical columns represent the constraint transmission continuity rate, candidate solution orderliness rate, execution solution formation accuracy rate, basis traceability completeness rate, and output consistency rate, respectively. The horizontal columns represent different combinations of task complexity and solution type, and the color intensity indicates the corresponding indicator values. Among them, the solution of this invention is the solution that adopts the complete technical process of this invention, the control solution A is the solution that only performs data retrieval and matching or conventional content organization, and the control solution B is the solution that does not adopt the continuous deduction, orderly decision-making, and complete basis identification writing mechanism.

[0089] In summary, this invention addresses the problem of insufficient constraint information organization by: extracting, classifying, and associating tailings design constraints from basic tailings design data through an industrial cloud platform to form a tailings design intent list; and then combining this list with supporting information to perform scheme deduction. This improves the continuity of constraint transmission and the systematic nature of candidate scheme formation. Furthermore, by constructing an adjudication link through the industrial cloud platform, collaborative adjudication is performed on discrepancies in candidate schemes, and the adjudication results are backfilled into the candidate scheme set. Unadopted content is consolidated, while adopted content and its associated evidence are retained. This solves the problem of insufficient orderliness in collaborative adjudication, thereby improving the accuracy of scheme formation, the clarity of evidence traceability, and the consistency of output results.

[0090] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A human-model-platform integrated method for tailings design process, characterized in that, include: The industrial cloud platform receives basic tailings design data and task objectives, and extracts, classifies and associates tailings design constraints to form a tailings design intent list. Following the constraint order and professional orientation corresponding to the tailings design intent list, the knowledge base, standard base and case base are connected to extract matching scheme support information. The scheme deduction model is combined with the scheme support information to deduce the scheme, form a candidate scheme set, and send the candidate scheme set into the adjudication link. In the adjudication chain, collaborative adjudication is performed on the technical comparison nodes, parameter selection nodes, process disagreement nodes, and risk assessment nodes within the candidate solution set, and the adjudication results are backfilled into the candidate solution set. Unadopted content is collected, and adopted content and related evidence are retained to form an execution plan. Based on the technological, professional, and supporting relationships determined in the implementation plan, the industrial cloud platform is used to generate tailings design specifications, process connection diagrams, building connection diagrams, and material statistics. The supporting identifiers from the implementation plan are written into the output results to form an integrated tailings design result.

2. The integrated human-model-platform method for tailings design process as described in claim 1, characterized in that, The specific steps for forming the tailings design intent list are as follows: The industrial cloud platform receives basic tailings design data and task objectives, and collects and organizes the basic tailings design data to form a basic record set. Extract tailings design constraints related to the task objectives from the basic record set, and classify the tailings design constraints to form a set of tailings design constraints. The various items in the tailings design constraint set are associated with the task objectives and organized according to the constraint order and professional orientation to form a tailings design intent list.

3. The integrated human-model-platform method for tailings design process as described in claim 1, characterized in that, The specific steps for forming a candidate solution set and sending the candidate solution set into the decision link are as follows: Following the constraint order and professional orientation corresponding to the tailings design intent list, the items in the tailings design intent list are sequentially expanded to form an extraction order. The knowledge base, standard base, and case base are connected according to the extraction order. Relevant content that matches each item in the tailings design intent list is extracted and the relevant content is organized into scheme support information that corresponds to each item in the tailings design intent list. A scheme deduction model is formed by constructing an input correspondence layer, a sequential arrangement layer, a fusion deduction layer, and a result processing layer. The tailings design intent list and scheme support information are fed into the scheme simulation model to simulate the scheme and generate candidate branch results. The candidate branch results are merged and organized according to process relationship, professional relationship and basis relationship to form a candidate solution set, and the candidate solution set is sent to the decision link.

4. The integrated human-model-platform method for tailings design process as described in claim 1, characterized in that, The specific steps for forming the execution plan are as follows: Extract technology comparison nodes, parameter selection nodes, process disagreement nodes, and risk assessment nodes from the candidate solution set, and organize each node into a set of nodes to be decided. The set of adjudication nodes is arranged according to the constraint order and professional orientation corresponding to the tailings design intent list, forming an adjudication sequence; The participants in the organizational design process execute collaborative decisions along the decision sequence, performing decisions on each node in the set of nodes to be decided in the decision chain, to form a decision result. The ruling results are then filled into the corresponding content in the candidate solution set. The unadopted content in the candidate solution set is collected, and the adopted content and related evidence are retained to form an implementation plan.

5. The integrated human-model-platform method for tailings design process as described in claim 1, characterized in that, The specific steps to form the integrated tailings design are as follows: The process relationships, professional relationships, and basis relationships in the implementation plan are mapped to the tailings design specifications, process connection diagrams, building connection diagrams, and material statistics to form an output organization list; Extract the corresponding content from the implementation plan along the results organization list, and organize the corresponding content into tailings design description, process connection diagram, building connection diagram and material statistics to form a set of output results; Write the basis identifiers in the execution plan into the corresponding positions in the output result set, and merge and organize the output result set to form an integrated tailings design result.

6. The integrated human-model-platform method for tailings design process as described in claim 3, characterized in that, The scheme deduction model is formed by constructing an input correspondence layer, a sequential arrangement layer, a fusion deduction layer, and a result processing layer. The specific steps are as follows: Each tailings design intent in the tailings design intent list is matched with the supporting information of the scheme to form an input correspondence layer; The corresponding contents in the input layer are arranged according to the constraint order and professional orientation to form a sequential arrangement layer; The corresponding items in the sequential arrangement layer are continuously deduced to form a fusion deduction layer; The different inference contents formed by the fusion inference layer are merged and identified to form the result processing layer; The input correspondence layer serves as the basis for the arrangement of the sequential arrangement layer, the sequential arrangement layer serves as the basis for the deduction of the fusion deduction layer, and the fusion deduction layer serves as the basis for the organization of the result organization layer. The input correspondence layer, the sequential arrangement layer, the fusion deduction layer, and the result organization layer are connected sequentially to form a scheme deduction model.

7. The integrated human-model-platform method for tailings design process as described in claim 3, characterized in that, The adjudication chain refers to the orderly processing path formed after the candidate scheme set is formed, which takes into account the differences, related evidence and corresponding relationships in the candidate scheme set, and organizes the subsequent adjudication processing in sequence according to the constraint order and professional orientation in the tailings design intent list.

8. The human-model-platform integrated method for tailings design process as described in claim 4, characterized in that, The organizational design participants perform collaborative adjudication along the adjudication sequence, executing collaborative adjudication on each node in the set of nodes to be adjudicated in the adjudication chain. The specific steps are as follows: The decision-making process involves locating each node to be decided sequentially along the decision-making chain, and retrieving the candidate direction, related basis, and decision result corresponding to the previous node to be decided for each node. The design participants are determined according to the professional orientation of each node to be adjudicated, and the design participants are responsible for comparing and judging the differences. The content to be retained and the content not adopted are determined according to the node type of the node to be decided, a decision result is formed, and the decision result is passed to the next node to be decided along the decision sequence.

9. The human-model-platform integrated method for tailings design process as described in claim 5, characterized in that, The basis identifiers refer to the basis source category identifier, the basis source document identifier, the basis location identifier, and the corresponding adjudication node identifier.

10. The human-model-platform integrated method for tailings design process as described in claim 5, characterized in that, The step of writing the basis identifier in the execution plan into the corresponding position in the output result set means writing the basis source category identifier, basis source document identifier, basis location identifier, and corresponding adjudication node identifier into the corresponding position in the output result set according to the identifier relationship between each corresponding position in the result organization list and the basis identifier, so that the explanatory content in the tailings design description, the graphic content in the process connection diagram, the graphic content in the building connection diagram, and the statistical content in the material statistics content are respectively aligned with the corresponding basis identifier.

Images