A system for quickly calculating beneficiation cost and evaluating schemes

Abstract

The present application belongs to the technical field of mineral processing cost data processing and scheme evaluation, and discloses a kind of mineral processing cost fast measurement and scheme evaluation system.The system converts raw ore properties, process combination, product indicators, resource prices and cost items into unified measurement fields, and constructs a process cost coupling graph containing process nodes, material flow transfer relationships and cost collection relationships;Within the graph, material balance forward derivation and product indicator reverse verification are performed, affected cost items are recalculated based on node association relationships, and historical sub-item residuals are corrected within the scope of constraint propagation results.The system links mineral quantity, grade, recovery rate, cost collection and historical deviation correction in the same data link, reduces the problems of process non-closure, cost repeated collection and non-linked calculation after parameter change, and forms traceable scheme evaluation data.

Description

Technical Field

[0001] This invention belongs to the field of mineral processing cost data processing and scheme evaluation technology, and discloses a mineral processing cost rapid calculation and scheme evaluation system. Background Technology

[0002] Current mineral processing cost calculations typically rely on spreadsheets, cost ledgers, or expense aggregation tables within enterprise management systems. When calculating costs, technicians input data such as raw ore processing volume, raw ore grade, concentrate grade, recovery rate, reagent consumption, electricity consumption, water consumption, media consumption, labor costs, and ore disposal costs into a preset template. Then, they calculate the cost per ton of ore, concentrate cost, and unit metal cost by multiplying unit consumption by unit price, summing individual costs, and dividing the total cost by the processing volume or metal quantity. While this method has a relatively intuitive calculation logic and can meet the periodic accounting needs of a stable operating process, its data organization is centered on expense categories. The material transfer relationships between operations such as crushing, grinding, grading, beneficiation, concentration and dewatering, and tailings discharge in the mineral processing process are usually only reflected through manually entered intermediate values. The calculation tables lack a unified expression of the relationships between ore quantity, grade, recovery rate, and cost items.

[0003] In conventional scheme evaluation scenarios, technicians create multiple calculation versions for different combinations of mineral processing flows. These versions may adjust grinding fineness, reagent regime, number of separation stages, concentrate indicators, or tailings disposal methods, and then calculate the cost results for each version separately. Existing systems often generate scheme comparison tables by copying templates, adjusting fields, and re-summarizing, displaying results such as cost per ton of ore, recovery rate, concentrate grade, and processing capacity side-by-side. This approach primarily reflects differences in input fields and summary results, lacking a record of the transmission path from process nodes to cost items. When a parameter changes, the system typically cannot automatically determine the affected downstream process nodes and cost items, easily leading to situations where local parameters have changed but downstream ore quantities, tailings quantities, reagent costs, or electricity costs have not been updated synchronously.

[0004] Some existing cost estimation systems incorporate historical production or process data to empirically adjust the cost of new schemes. These systems typically select similar historical cases based on ore type, processing scale, or process category, and then adjust the current estimated value using historical average unit consumption, historical average cost, or manually set coefficients. While this method can reflect actual production fluctuations to some extent, historical data often suffers from differences in statistical periods, cost definitions, and unit consumption aggregation objects. If historical residuals are directly added to the current scheme's cost without being constrained by material balance, cost aggregation definitions, and process node relationships, historical production deviations may propagate to the new scheme, causing the cost adjustment result to deviate from the current scheme's process constraints and making it difficult to explain the specific cost source corresponding to the adjusted value.

[0005] The main technical problem with existing technologies is that the process constraints and cost calculation logic are disconnected in the rapid cost calculation process for mineral processing. This makes it difficult for the system to maintain consistent relationships between ore quantity transfer, grade recovery, cost aggregation, and historical residual correction within the same data link after changes in scheme parameters. Specifically, existing tabular or subject-based calculation methods can summarize costs, but cannot construct a unified relational structure for raw ore input, process nodes, product output, tailings output, and cost items. Existing historical data correction methods can utilize existing cost data, but cannot guarantee that the correction value is within the material balance and cost aggregation constraints of the current scheme. As a result, rapid calculation results are prone to problems such as incomplete process closure, duplicate or omitted cost items, failure to recalculate in response to changes in local parameters, and difficulty in tracing the source of costs. Summary of the Invention

[0006] The purpose of this invention is to provide a system for rapid calculation of mineral processing costs and evaluation of solutions, which can effectively solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the technical solution adopted by the present invention is as follows: A rapid cost calculation and scheme evaluation system for mineral processing includes a scheme data structured processing module, a process cost coupling map construction module, a two-way constraint propagation calculation module, an associated cost incremental recalculation module, and a historical residual constraint correction module; The data structuring module of the solution converts the properties of raw ore, process combinations, product indicators, resource prices, and cost items into unified calculation fields; The process cost coupling graph construction module generates a coupling graph containing mineral processing process nodes, material flow relationships, and cost aggregation relationships based on the unified calculation fields. The bidirectional constraint propagation calculation module performs forward derivation of material balance and reverse verification of product indicators within the coupling graph. The associated cost increment recalculation module recalculates the affected cost items according to the node association relationship; The historical residual constraint correction module performs residual correction on the sub-item costs within the range of constraint propagation results.

[0008] Preferably, the process cost coupling graph construction module sets the raw ore input node, crushing node, grinding node, classification node, beneficiation node, concentration and dewatering node, concentrate output node, tailings output node and cost item node as heterogeneous nodes, and establishes directed associations according to ore quantity transfer edge, grade recovery edge, unit consumption edge, cost collection edge and indicator constraint edge; The heterogeneous nodes are configured with attributes such as processing capacity, grade, recovery rate, yield, unit consumption, and unit price. The directed association configuration of source node, target node, constraint type, and recalculation flag ensures that cost items and corresponding mineral processing process nodes maintain the same data reference link.

[0009] Preferably, the bidirectional constraint propagation calculation module includes forward material recursion processing and reverse index verification processing; The forward material recursion process calculates the amount of ore entering, the amount of ore exiting, the amount of metal, and the amount of tailings at each process node according to the process topology order. The reverse index verification process uses the target concentrate grade, target recovery rate, and allowable tailings metal loss as boundaries to back-transmit and mark the recovery rate, enrichment ratio, and loss rate of the beneficiation nodes. When the forward recursion result is inconsistent with the reverse verification boundary, the constraint conflict location and the corresponding set of process nodes are generated, and subsequent cost calculations are restricted to only referencing data links that have passed the consistency verification.

[0010] Preferably, the associated cost increment recalculation module establishes a parameter influence index table, which records the correspondence between grinding fineness, feed grade, throughput, reagent system, recovery rate, concentrate yield, tailings yield and power consumption cost, reagent cost, media cost, water consumption cost, labor cost, and tailings disposal cost. When any scheme field changes, the associated cost incremental recalculation module extracts the affected process nodes and affected cost items based on the parameter impact index table, and performs local recalculation along the downstream direction of the coupling graph, starting from the affected process node.

[0011] Preferably, the process cost coupling map construction module sets a cost aggregation caliber field for the cost item node, and the cost aggregation caliber field includes aggregation by raw ore quantity, aggregation by concentrate quantity, aggregation by operation quantity, and aggregation by tailings quantity; When the same cost item node is connected to multiple mineral processing nodes at the same time, a unique primary aggregation edge and at least one reference association edge are generated based on the cost aggregation caliber field. The unique primary aggregation edge participates in cost aggregation, the reference association edge participates in parameter influence index construction, and duplicate aggregation relationships and missing aggregation relationships are marked with graph consistency before cost aggregation.

[0012] Preferably, the bidirectional constraint propagation calculation module configures a conflict propagation boundary for the constraint conflict location, and the conflict propagation boundary is jointly determined by the nearest upstream material closure node, the nearest downstream product indicator node, and the associated cost item node; For process nodes located within the conflict propagation boundary, the ore quantity, grade, and recovery rate attributes are recalculated using the current scheme fields; For process nodes located outside the conflict propagation boundary, retain the attribute values ​​that have passed the consistency check; The cost recalculation starting point is formed by intersecting the conflict propagation boundary with the set of affected process nodes of the associated cost increment recalculation module.

[0013] Preferably, the associated cost increment recalculation module sets a trigger expression for the affected cost item, and the trigger expression is composed of process node attributes, cost aggregation caliber field and resource price version; When the resource price version changes but the process node attributes remain unchanged, the recalculation scope is limited to cost item nodes with the same price reference identifier. When the process node attributes change but the resource price version remains unchanged, the recalculation scope is limited to cost item nodes connected along the mineral quantity transfer edge and the unit consumption edge. When both change simultaneously, the two recalculation ranges will be merged and duplicate cost item nodes will be removed.

[0014] Preferably, the historical residual constraint correction module extracts residual samples from the historical mineral processing scheme library based on ore property similarity, process topology similarity, cost item similarity, and data caliber consistency. The residual samples are categorized by power consumption, reagents, media, water consumption, labor allocation, and tailings disposal. The historical residual constraint correction module maps the sub-item residuals to the corresponding cost item nodes in the coupled graph, and uses the material balance result, cost aggregation caliber field and resource price version as correction boundaries to remove or scale residual samples that exceed the correction boundaries.

[0015] Preferably, the associated cost incremental recalculation module establishes a cost version chain before generating the cost recalculation result. The cost version chain records the unified calculation fields before the change, the unified calculation fields after the change, the recalculation starting point, the recalculation path, the unique main aggregation edge participating in the aggregation, and the reference associated edges that do not participate in the aggregation. When the same cost item node is accessed in different recalculation paths, the path record closest to the changed field is retained according to the cost version chain, and the remaining path records are set as bypass verification records. When summarizing costs, only the retained path records and their corresponding cost item node values ​​are referenced.

[0016] Preferably, the historical residual restricted correction module establishes a residual write-back table, which records the source of historical residual samples, the type of sub-item residual, the corresponding cost item node, the correction boundary processing method, and the cost version identifier after correction. When obtaining the measured cost data in the future, the measured cost data will be split according to the cost item nodes, and the difference will be calculated with the item correction value corresponding to the corrected cost version identifier. The difference calculation results are updated with residual sample weights according to the consistency of historical residual sample sources and data caliber, and the updated residual sample weights are made available for the next scheme evaluation and rapid calculation.

[0017] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. By combining structured data processing, process cost coupling graphs, bidirectional constraint propagation, incremental recalculation of associated costs, and limited correction of historical residuals, the properties of raw ore, process combinations, product indicators, resource prices, and cost items in mineral processing cost estimation are transformed into a unified calculation field system. This system is further mapped into a coupling graph that includes mineral processing process nodes, material flow relationships, and cost aggregation relationships. Forward derivation of material balance ensures that the ore quantity, metal quantity, and tailings quantity at each process node form a continuous calculation chain along the process direction. Reverse verification of product indicators ensures that the constraints on concentrate grade, recovery rate, and tailings loss can be transmitted back to relevant nodes. Incremental recalculation of associated costs limits the recalculation scope according to node relationships, avoiding data inconsistencies caused by full table recalculation or manual selection of recalculation items. Historical residual correction is limited to the constraint propagation results, ensuring that historical measured deviations can only be corrected for corresponding cost items, and cannot cover the current process calculation results. This ensures that the rapid calculation results remain consistent across process closure, cost aggregation, and residual correction.

[0018] 2. Regarding secondary technical effects, the ore quantity transfer edge, grade recovery edge, unit consumption edge, cost aggregation edge, and indicator constraint edge in the process cost coupling graph ensure a clear data reference link between cost items and corresponding beneficiation process nodes, facilitating the identification of cost change sources. The settings of the cost aggregation caliber field, unique primary aggregation edge, and reference association edge enable the differentiation between summary and reference relationships when the same cost item connects multiple process nodes, reducing duplicate and missing aggregations. The coordination of the parameter-affected index table, trigger expression, recalculation path, and cost version chain ensures that after changes to fields such as processing volume, feed grade, grinding fineness, reagent system, and resource price version, the system retains the fields before and after the change, the recalculation starting point, the recalculation path, and the aggregation edges participating in the aggregation. The residual write-back table records the source of residual samples, the type of sub-items residual, the correction boundary processing method, and the cost version identifier, enabling subsequent measured costs to be broken down by cost item and the residual sample weights to be written back, forming an auditable data correction link. Attached Figure Description

[0019] Figure 1 This is a flowchart of the overall processing flow of the mineral processing cost rapid calculation and scheme evaluation system of the present invention; Figure 2This is a flowchart of the process cost coupling graph construction and cost aggregation relationship processing of the present invention; Figure 3 This is a flowchart of the bidirectional constraint propagation and associated cost increment recalculation process of the present invention; Figure 4 This is a flowchart of the historical residual constraint correction and residual write-back process of the present invention. Detailed Implementation

[0020] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0021] Please refer to Figure 1This embodiment provides a rapid cost calculation and scheme evaluation system for mineral processing, including a scheme data structuring module, a process cost coupling graph construction module, a two-way constraint propagation calculation module, a related cost increment recalculation module, and a historical residual constraint correction module. The scheme data structuring module converts the raw ore properties, process combinations, product indicators, resource prices, and cost items into unified calculation fields. The process cost coupling graph construction module generates a coupling graph containing mineral processing process nodes, material flow relationships, and cost aggregation relationships based on the unified calculation fields. The two-way constraint propagation calculation module performs forward derivation of material balance and reverse verification of product indicators within the coupling graph. The related cost increment recalculation module recalculates affected cost items according to node relationships. The historical residual constraint correction module performs residual correction on sub-item costs within the constraint propagation result range. In specific implementation, the scheme data structuring module receives raw ore processing volume, raw ore grade, target concentrate grade, target recovery rate, process route, and operation nodes. The system includes a sequence of data points, unit consumption of reagents, unit consumption of electricity, unit consumption of media, unit consumption of water, unit price of tailings disposal, and resource price versions. Data from calculation tables, historical scheme databases, or manual input interfaces are converted into the same field code. The process cost coupling graph construction module maps the unified calculation fields to node attributes and edge attributes. The bidirectional constraint propagation calculation module calculates the ore quantity and metal quantity along the process direction in the graph and uses product indicators to check the constraint status of the beneficiation node and tailings node. The associated cost increment recalculation module only performs local calculations on cost items affected by parameter changes. The historical residual limited correction module limits historical measured deviations to the corresponding cost item node and cost collection caliber. The working principle of this embodiment is to replace isolated cost tables with graph relationships, limit available calculation links with constraint propagation, and absorb historical deviations with limited residual correction. This embodiment puts the raw ore input, process nodes, product output, tailings output, and cost items in the same data link, reducing the situation of non-closed process, repeated cost collection, and failure to recalculate when local parameters change.

[0022] In this embodiment, the data structuring module preprocesses the input data according to field source, measurement object, and calculation caliber. The raw ore properties are broken down into raw ore processing volume, metal grade, particle size range, and optional minerality labels. The process combination is broken down into a sequence of operation nodes and the connection relationships between nodes. Product indicators are broken down into concentrate grade boundaries, recovery rate boundaries, and tailings metal loss boundaries. Resource prices are broken down into price items, price units, price effective versions, and price reference identifiers. Cost items are broken down into electricity consumption costs, reagent costs, media costs, water consumption costs, labor allocation costs, and tailings disposal costs. During field conversion, a field identifier, unit identifier, caliber identifier, and version identifier are generated for the same data field. The unit identifier is used for tons and thousands of tons. The conversion relationships between watt-hours, cubic meters, reagent quality, and amount are defined. The caliber identifier is used to distinguish different calculation benchmarks collected by raw ore quantity, concentrate quantity, operation quantity, and tailings quantity. The version identifier is used to limit the referential range of resource prices and historical cost samples. If there are multiple source values ​​for the same cost item, the scheme data structuring module uses fields with consistent measurement caliber, similar time versions, and high process matching degree as the current field, and retains the remaining fields as bypass verification fields. Bypass verification fields do not directly participate in cost summarization, but are used for subsequent anomaly location and residual sample write-back. This embodiment ensures that the data caliber before mineral processing cost calculation is uniformly expressed, avoiding the same-named cost fields from entering the same calculation link due to different units, periods, or collection objects.

[0023] In this embodiment, the forward derivation of the material balance is described by the following formula: ; ; ; ; in, This represents the amount of ore input at the i-th process node. This represents the set of upstream nodes of the i-th process node. This represents the ore output of the j-th upstream process node. This represents the mineral distribution coefficient from upstream node j to node i. This represents the amount of metal corresponding to node i. This indicates the ore grade at node i. This represents the metal recovery rate at node i. This represents the product grade of node i. This represents the amount of product ore at node i. This represents the tailings quantity at node i. Physically, it represents the processing baseline formed by the upstream ore flows after distribution. For example, if the upstream ore quantity at a certain beneficiation node is 1000, the distribution coefficient is 1, the feed grade is 0.02, the recovery rate is 0.9, and the product grade is 0.4, then the product ore quantity is 45, and the tailings quantity is 955. This value allows the tailings disposal cost to be calculated based on the derived tailings quantity, avoiding the disruption of the material closure relationship caused by manually entering the tailings quantity separately.

[0024] In a preferred embodiment, reference Figure 2 The process cost coupling graph construction module sets the raw ore input node, crushing node, grinding node, classification node, beneficiation node, thickening and dewatering node, concentrate output node, tailings output node, and cost item node as heterogeneous nodes, and establishes directed associations according to ore quantity transfer edge, grade recovery edge, unit consumption edge, cost aggregation edge, and indicator constraint edge. The heterogeneous nodes are configured with processing capacity, grade, recovery rate, yield, unit consumption, and unit price attributes. The directed associations are configured with source node, target node, constraint type, and recalculation mark, so that the cost item and the corresponding beneficiation process node maintain the same data reference link. In specific implementation, the raw ore input node stores the raw ore processing capacity and raw ore grade, the crushing node stores the crushing operation volume and crushing power consumption caliber, and the grinding node stores the grinding operation volume, grinding fineness, and media. The system uses a multi-level approach, where each node stores return and input quantities, each separation node stores reagent regime, recovery rate, concentrate grade, and tailings loss, each concentration and dewatering node stores concentrate moisture content changes and dewatering operation volume, each concentrate output node stores concentrate quantity and product grade, each tailings output node stores tailings quantity and tailings disposal criteria, and each cost item node stores expense name, unit consumption value, unit value, aggregation criteria, and price version. Ore quantity transfer nodes connect to operational process nodes, grade recovery nodes connect to metal quantity change nodes, unit consumption nodes connect operational nodes to consumption-related cost items, cost aggregation nodes connect cost items to expense summary nodes, and indicator constraint nodes connect product indicator nodes to relevant process nodes. This embodiment converts the process data and cost data in the mineral processing flow into a traversable, verifiable, and partially recalculated relational structure.

[0025] Furthermore, heterogeneous nodes and directed associations are stored using both a graph data table and a field index table. The graph data table records node identifiers, node types, node attributes, and edge identifiers. The field index table records the binding relationships between unified calculation fields and node attributes. When a scheme field changes, the field index table is used to locate the node attribute corresponding to the modified field, and then the graph data table is used to obtain the connecting edges and downstream node sets. For process nodes such as crushing, grinding, grading, and beneficiation, the processing capacity in the node attribute is generated by combining the source node's transmitted value and the current node's allocation coefficient. The grade attribute is generated by the ratio of upstream metal quantity to current ore quantity. The recovery rate attribute is generated in the beneficiation process. The system performs calculations on other nodes and allocates them to concentrate output nodes and tailings output nodes. For cost item nodes, the unit consumption attribute is bound to the corresponding operation node, the unit price attribute is bound to the resource price version, and the aggregation caliber attribute is bound to the cost summary caliber. After the graph is generated, the system performs consistency marking on non-raw ore nodes without incoming edges, non-output nodes without outgoing edges, cost nodes without master aggregation edges, and duplicate master aggregation edges, and feeds the marking results back to the scheme data structuring processing module for field completion or caliber verification. This embodiment makes the graph construction process have a data closed loop, and can exclude process nodes and cost nodes that do not have a complete reference basis before cost calculation.

[0026] In this embodiment, the breakdown of cost item nodes is described using the following formula:

[0027] ; in, This represents the itemized cost of the k-th cost category. This represents the set of process nodes associated with the k-th type of cost item. This value indicates whether process node i is considered a primary aggregation relation for cost item k. A value of 1 indicates participation in the aggregation, while a value of 0 indicates only a reference association. This represents the unit consumption of cost item k corresponding to process node i. This represents the processing volume of process node i. Indicates the conversion factor for the scope of collection. This represents the unit price of the resource corresponding to cost item k. This represents the set of all cost items, where C represents the total value of all cost items. For example, if the grinding node's processing capacity is 1000, the media unit consumption is 0.5, the aggregation conversion factor is 1, and the media unit price is 6, then... The media cost is 3000. If the same media cost node has a reference correlation with the tier node and... The hierarchical nodes do not participate in the cost aggregation of this medium. This value ensures that when the same cost item is connected to multiple process nodes, a unique aggregation benchmark is still maintained.

[0028] In a preferred embodiment, reference Figure 3 The bidirectional constraint propagation calculation module includes forward material recursion processing and reverse index verification processing. The forward material recursion processing calculates the input, output, metal content, and tailings content of each process node according to the process topology. The reverse index verification processing uses the target concentrate grade, target recovery rate, and allowable tailings metal loss as boundaries to back-transmit and mark the recovery rate, enrichment ratio, and loss rate of the beneficiation nodes. When the forward recursion result is inconsistent with the reverse verification boundary, a constraint conflict location and the corresponding set of process nodes are generated, and subsequent cost calculations are restricted to using only data links that have passed consistency verification. Specifically, the forward material recursion processing starts from the raw ore input node and proceeds according to the crushing, grinding, and... The topological sequence of grading, beneficiation, concentration and dewatering, product output, and tailings output generates the input and output quantities for each node. The metal quantity is obtained by multiplying the ore quantity by the grade, and the tailings quantity is obtained by subtracting the product ore quantity from the input ore quantity at each node. The reverse index verification process starts from the concentrate output node and the tailings output node, and reversely transmits the target concentrate grade, target recovery rate, and allowable tailings metal loss to the beneficiation node and the grading node. Conflict markers are generated for nodes that do not meet the constraint boundaries. The conflict markers include node identifier, field identifier, conflict type, and associated cost items. This embodiment establishes a mutual verification relationship between the forward calculation results and the product target, preventing the derivation based solely on the process sequence while ignoring the terminal index constraints.

[0029] Furthermore, the bidirectional constraint propagation calculation module does not terminate all calculations upon discovering a constraint conflict. Instead, it limits the conflict to a local graph range with a directed correlation to the conflicting field. The ore quantity generated by forward recursion and the boundary value generated by reverse verification are stored as calculated values ​​and constraint values ​​respectively in the same node attribute. The calculated values ​​are used to record the results derived from the current scheme field, and the constraint values ​​are used to record the boundaries obtained from the product index feedback. If there is an inconsistency between the calculated value and the constraint value in the node attribute, the adjacent upstream edge and the adjacent downstream edge are added to the conflict candidate set. For crushing and grinding nodes, conflicts are mainly transmitted to the cost node through processing volume, operation volume, and unit consumption. For beneficiation nodes, conflicts are simultaneously transmitted to the product output node, tailings output node, and reagent cost node through recovery rate, enrichment ratio, concentrate grade, and tailings loss. For concentration and dewatering nodes, conflicts are transmitted to the dewatering cost node through concentrate quantity and operation volume. This embodiment ensures a clear correspondence between the conflict location and the affected cost item, avoiding the recalculation of irrelevant cost items due to local index conflicts.

[0030] In this embodiment, the degree of constraint conflict is described by the following formula:

[0031] in, This indicates the degree of constraint conflict at process node i. This represents the amount of minerals obtained from the forward recursion. This indicates the boundary of the node's mineral quantity returned after reverse verification. This represents the node grade obtained through forward recursion. This indicates the node grade boundary for reverse verification and feedback. This indicates the boundary of the node recycling rate for reverse verification and feedback. , and These represent the weights of ore quantity, grade, and recovery rate, respectively, and the sum of their values ​​is 1. This represents absolute value operations, such as those performed on a selector node. For 1000, It is 980. It is 0.02. It is 0.021. It is 0.9. If the value is 0.92 and the three weights are 0.4, 0.3, and 0.3 respectively, then the degree of conflict is formed by the three types of biases. Based on this, the system will include the conflicting nodes in the local recalculation range, instead of including unrelated price version fields in the same conflict range.

[0032] In a preferred embodiment, the associated cost incremental recalculation module establishes a parameter impact index table. This index table records the correspondence between grinding fineness, feed grade, throughput, reagent system, recovery rate, concentrate yield, tailings yield, and electricity cost, reagent cost, media cost, water cost, labor cost, and tailings disposal cost. When any scheme field changes, the associated cost incremental recalculation module extracts the affected process nodes and affected cost items based on the parameter impact index table, and performs a local recalculation starting from the affected process node and proceeding downstream along the coupling graph. Specifically, the parameter impact index table consists of field identifiers, associated node types, and related... The recalculation is structured around the edge type, cost item type, and recalculation direction. The grinding fineness field is bound to the grinding node, grading node, and beneficiation node, and is associated with power consumption cost, media cost, and reagent cost through the unit consumption edge. The feed grade field is bound to the beneficiation node and is associated with concentrate output node and tailings output node through the grade recovery edge. The processing volume field is bound to the raw ore input node and is associated with all downstream operation nodes through the ore quantity transfer edge. The reagent regime field is bound to the beneficiation node and is associated with reagent cost through the unit consumption edge. The resource price version field is bound to the cost item node and the recalculation scope is limited by the price reference identifier. In this embodiment, the cost recalculation after parameter changes is triggered by the index relationship, avoiding manual judgment of the recalculation scope.

[0033] Furthermore, the incremental cost recalculation module employs a processing logic of changing field locking, node set expansion, cost item extraction, and summary replacement when performing partial recalculation. Changing field locking is used to determine the currently modified unified calculation field. Node set expansion is used to find downstream nodes along the ore quantity transfer edge, grade recovery edge, and unit consumption edge. Cost item extraction is used to find cost item nodes that are directly or indirectly connected in the node set. Summary replacement is used to replace the old version value of the corresponding cost item in the cost summary table with the recalculated value. If the changed field is the resource price version and the process node attribute has not changed, only cost item nodes with the same price reference identifier are updated. If the changed field is the process node attribute and the resource price version has not changed, only cost item nodes affected along the process relationship are updated. If both types of fields change simultaneously, the two types of node sets are merged and duplicates are removed. After the recalculation is completed, the system generates a calculation timestamp, field version identifier, and graph path identifier for the recalculated nodes. Unaffected cost items retain their original calculated values ​​and are marked as being in use. This embodiment makes the cost recalculation granularity correspond to the actual parameter influence range, reducing the risk of intermediate result overwriting caused by full recalculation.

[0034] In this embodiment, the set of affected cost items is described by the following formula: ; in, This represents the set of affected cost items triggered by the change in field x. Represents the set of all process nodes. This indicates whether the changed field is associated with a process node. A value of 1 indicates that there is an association, and a value of 0 indicates that there is no association. Indicates whether process node i is associated with cost item k. Indicates whether the changed field is directly associated with the cost item. This indicates that it exists. This represents the union of sets, for example, when the field is grinding fineness, if Corresponding grinding nodes and grading nodes, and the relevant nodes pass through Connecting power consumption costs, medium costs, and reagent costs, these three types of costs then enter... If the field is the electricity price version, If only the electricity cost is considered, then only the electricity cost enters the calculation. This value allows changes in process parameters and price parameters to create different recalculation paths.

[0035] In a preferred embodiment, the process cost coupling graph construction module sets a cost aggregation caliber field for the cost item nodes. This cost aggregation caliber field includes aggregation by raw ore quantity, by concentrate quantity, by processing volume, and by tailings quantity. When the same cost item node is connected to multiple beneficiation process nodes simultaneously, a unique primary aggregation edge and at least one reference association edge are generated based on the cost aggregation caliber field. The unique primary aggregation edge participates in cost aggregation, and the reference association edge participates in parameter influence index construction. Before cost aggregation, duplicate and missing aggregation relationships are marked for graph consistency. Specifically, cost items aggregated by raw ore quantity use the raw ore input node or the main processing volume node as the source of the primary aggregation edge, and cost items aggregated by concentrate quantity use the concentrate output node as the primary aggregation edge. For cost items aggregated by workload, the primary aggregation edge source is the crushing, grinding, classification, beneficiation, or concentration and dewatering node where the actual operation occurs. For cost items aggregated by tailings volume, the primary aggregation edge source is the tailings output node. When a reagent cost is associated with both grinding and beneficiation nodes, if the cost aggregation caliber is based on beneficiation workload, a unique primary aggregation edge is generated between the beneficiation node and the reagent cost node, while only a reference association edge is generated between the grinding node and the reagent cost node. When the system detects that the same cost item has two primary aggregation edges or no primary aggregation edge, a consistency flag is generated before cost aggregation and the aggregation value of that cost item is stopped from being referenced. This embodiment separates the cost aggregation relationship from the parameter influence relationship, reducing the duplicate inclusion of the same cost item caused by multiple node associations.

[0036] Furthermore, the bidirectional constraint propagation calculation module configures conflict propagation boundaries for the constraint conflict locations. These boundaries are jointly determined by the nearest upstream material closure node, the nearest downstream product indicator node, and the associated cost item node. For process nodes located within the conflict propagation boundary, the ore quantity, grade, and recovery rate attributes are recalculated using the current scheme fields. For process nodes located outside the conflict propagation boundary, attribute values ​​that have passed consistency verification are retained. The cost recalculation starting point is formed by intersecting the conflict propagation boundary with the set of affected process nodes in the associated cost increment recalculation module. Specifically, the nearest upstream material closure node refers to the point where, when tracing back upstream from the conflict node, the ore quantity input, ore quantity output, and metal quantity have all passed a certain threshold. For consistency verification process nodes, the nearest downstream product indicator node refers to the node that directly receives the boundary of concentrate grade, recovery rate, or tailings metal loss when traversing downstream from the conflict node. The associated cost item node refers to the cost node connected to the conflict node through the unit consumption edge or cost aggregation edge. The system uses the subgraph enclosed by the above three types of nodes as the conflict propagation boundary and recalculates the node attributes within the subgraph. The node attributes outside the subgraph are not overwritten. When the associated cost incremental recalculation module has provided the set of affected process nodes, the system takes the intersection of the nodes within the conflict propagation boundary and the set of affected process nodes as the actual recalculation starting point. This embodiment makes the conflict correction range and the cost recalculation range mutually calibrated to avoid isolated conflict markers causing irrelevant node recalculation.

[0037] In a preferred embodiment, the associated cost increment recalculation module sets a trigger expression for the affected cost items. This trigger expression is composed of process node attributes, cost aggregation caliber fields, and resource price versions. When the resource price version changes but the process node attributes remain unchanged, the recalculation scope is limited to cost item nodes with the same price reference identifier. When the process node attributes change but the resource price version remains unchanged, the recalculation scope is limited to cost item nodes connected along the ore quantity transfer edge and the unit consumption edge. When both change simultaneously, the two recalculation scopes are merged, and duplicate cost item nodes are removed. Specifically, the trigger expression records the basic variables upon which the cost item depends; the electricity consumption cost trigger expression includes workload and unit electricity consumption. The electricity price version, the reagent cost trigger expression includes versions for beneficiation operation volume, reagent system, and reagent price; the media cost trigger expression includes versions for grinding operation volume, media unit consumption, and media price; the water consumption cost trigger expression includes versions for water consumption at operation nodes and water price; the labor allocation cost trigger expression includes the aggregation caliber of treatment volume or operation volume; and the tailings disposal cost trigger expression includes versions for tailings volume and disposal unit price. Each trigger expression stores a variable reference identifier and a graph path identifier. The variable reference identifier is used to determine whether a field has changed, and the graph path identifier is used to determine whether the change affects the current cost item. In this embodiment, whether a cost item needs to be recalculated is jointly determined by dependent variables and graph paths, reducing omissions caused by simply recalculating by cost name.

[0038] In a preferred embodiment, reference Figure 4The historical residual constraint correction module extracts residual samples from the historical mineral processing scheme library based on ore property similarity, process topology similarity, cost item similarity, and data caliber consistency. These residual samples are categorized by power consumption, reagents, media, water consumption, labor allocation, and tailings disposal. The historical residual constraint correction module maps these residuals to the corresponding cost item nodes in the coupled graph, and uses material balance results, cost aggregation caliber fields, and resource price versions as correction boundaries. Residual samples exceeding these boundaries are removed or scaled. Specifically, the historical mineral processing scheme library stores historical raw ore properties, historical process topologies, historical cost item calibers, historical resource price versions, historical calculated costs, and historical data. Historical measured costs are used to form residual samples, which are formed by the difference between historical measured costs and historical calculated costs. Ore property similarity is calculated by the difference in grade, grain size, and optional mineral properties. Process topology similarity is obtained by matching node types, node order, and edge relationships. Cost item similarity is obtained by matching cost names, aggregation scope, and units of measurement. Data caliber consistency is obtained by matching statistical periods, price versions, and aggregation objects. Before entering the current scheme, residual samples must be mapped to the same cost item node. Residuals that cannot be mapped to a node do not participate in correction. Residuals exceeding the correction boundary are limited or eliminated according to the boundary. In this embodiment, historical data is only used as a source of constrained sub-item corrections and does not replace the process calculation results of the current scheme.

[0039] In this embodiment, the historical residual constraint correction is described by the following formula:

[0040] ; in, This represents the constrained residual correction value for the k-th type of cost item. This represents the set of historical residual samples that can be mapped to the k-th type of cost item. Indicates the first The weights of each historical residual sample, This represents the overall similarity between the h-th historical residual sample and the current scheme. This represents the residual value of the h-th historical residual sample on the k-th cost item. and These represent the lower and upper boundaries of the residuals, determined by the material balance results, the cost aggregation caliber field, and the resource price version, respectively. This indicates a limit operation. This represents the corrected cost item of type k. For example, if the historical residual sample set of a certain drug cost item contains two samples with weights of 0.6 and 0.4, similarities of 0.8 and 0.5, and residuals of 200 and 500 respectively, and the boundary is 100 to 300, then the weighted residual will not exceed 300 after being limited. The corrected drug cost is formed within the material and price constraints of the current plan.

[0041] Furthermore, the comprehensive similarity of historical residual samples is composed of ore property similarity, process topology similarity, cost item similarity, and data caliber consistency. When the system retrieves samples in the historical scheme database, it does not use a single ore name as a filtering condition, but instead incorporates the differences in raw ore grade, optional particle size range, process node set, node connection order, cost aggregation caliber, and price version relationship into the matching process. For samples with similar grades but large differences in process topology, the system only allows them to enter the bypass sample pool of the same type of cost item, and does not allow them to enter the direct correction sample pool. For samples with similar process topology but large differences in price version, the system recalculates the historical cost according to the current price reference identifier before calculating the residual. For samples with the same cost item name but different aggregation calibers, the system does not perform cross-caliber direct correction, but retains them as an abnormal prompt record. After the residual sample mapping is completed, the system writes the source, caliber, similarity, residual value, and amplitude limiting processing result of each sample into the correction record. This embodiment allows the sources of historical deviations, adaptation basis, and correction boundaries to be reviewed, reducing the interference of mixed historical sample calibers on the current calculation link.

[0042] In a preferred embodiment, the associated cost incremental recalculation module establishes a cost version chain before generating cost recalculation results. This cost version chain records the unified calculation fields before and after the change, the recalculation starting point, the recalculation path, the unique main aggregation edge participating in the aggregation, and the reference associated edges not participating in the aggregation. When the same cost item node is accessed in different recalculation paths, the path record closest to the changed field is retained according to the cost version chain, and the remaining path records are set as bypass verification records. During cost aggregation, only the retained path records and their corresponding cost item node values ​​are referenced. In specific implementation, the cost version chain uses the scheme identifier and field version identifier as the main indexes, and the unified calculation fields before the change retain the original field values ​​and original... The unit, original caliber, and original price version are all stored in the unified calculation field after the change. The new field value, new unit, new caliber, and new price version are saved. The recalculation starting point stores the process node identifier that triggered the change. The recalculation path stores the edge sequence from the starting node to the cost item node. The unique primary aggregation edge stores the cost aggregation relationship that participates in the aggregation. The reference association edge stores the association relationship used to locate the impact but not participate in the aggregation. When the processing volume change and the pharmaceutical system change access the pharmaceutical cost node at the same time, the cost version chain compares the distance between the two paths, retains the direct path from the pharmaceutical system field to the selection node and then to the pharmaceutical cost node, and uses the processing volume path as a bypass verification record. This embodiment ensures that the same cost item can still maintain a unique aggregation source under multiple path triggers.

[0043] Furthermore, the cost version chain works in conjunction with the trigger expression, the cost aggregation caliber field, and the conflict propagation boundary. The trigger expression determines whether recalculation is required, the cost aggregation caliber field determines whether it participates in the aggregation, and the conflict propagation boundary determines whether the recalculation starting point is within the subgraph range that needs correction. The cost version chain records the results of the above judgments and forms a traceable cost version. During scheme evaluation, the system can simultaneously save the original version, the partially recalculated version, and the residual correction version. The original version records the cost results when the scheme fields were not modified, the partially recalculated version records the calculation results of the cost items affected by the field modification, and the residual correction version records the correction results after the historical residuals entered. The versions are associated with each other through field version identifiers and graph path identifiers, and cross-version direct overwriting of cost item values ​​is not allowed. When the scheme evaluation module reads the cost results, it looks up the corresponding cost version chain according to the current scheme field version, and then reads the main aggregation edge and cost item value participating in the aggregation. This embodiment enables scheme evaluation to distinguish the different cost sources caused by field changes, partial recalculation, and residual correction.

[0044] In a preferred embodiment, the historical residual constraint correction module establishes a residual write-back table. This table records the source of historical residual samples, the type of sub-item residuals, the corresponding cost item node, the correction boundary processing method, and the corrected cost version identifier. When subsequent measured cost data is obtained, the measured cost data is split according to the cost item nodes, and the difference is calculated between this data and the sub-item correction value corresponding to the corrected cost version identifier. The difference calculation result updates the residual sample weights according to the consistency of historical residual sample source and data caliber, and the updated residual sample weights are used for the next scheme evaluation and rapid calculation. Specifically, the residual write-back table uses the corrected cost version identifier as the primary key to record the data corresponding to each cost item node. The residual sample source, sample similarity, sample weight, amplitude limiting method, rejection reason, scaling ratio, and correction value are included. When the actual power consumption, actual reagent consumption, actual medium consumption, actual water consumption, actual manual allocation, or actual tailings disposal data of a certain scheme are subsequently obtained, the system will break down the actual cost by cost item node and find the sub-item correction value used at that time according to the cost version chain. Then, the difference between the actual value and the correction value will be calculated. If the difference comes from the price version change, the weight of the historical residual sample will not be adjusted. If the difference comes from the measurement deviation under the same caliber, the corresponding weight will be updated according to the consistency of sample source and caliber. This embodiment enables the historical residual sample to be corrected by sub-item with the actual measurement feedback, avoiding the repeated use of mismatched historical deviations in subsequent schemes.

[0045] In this embodiment, the residual sample weight write-back is described by the following formula: ; in, This represents the updated weight of the h-th historical residual sample. This represents the measured difference of the h-th historical residual sample on the k-th cost item. Indicates the first The measured difference of historical residual samples on the k-th type of cost item This indicates the prevention of small positive numbers with a denominator of zero. This indicates exponentiation. This represents a normalized summation over samples of the same type. For example, if there are two residual samples for the same drug cost item, both with a weight of 0.5 before the update, and the measured differences between the samples are 50 and 300 respectively, then... Under the given values, the first sample has a smaller exponential decay and a higher weight after the update than the second sample. Subsequent rapid calculations tend to use residual samples that are closer to the actual test feedback.

[0046] In this embodiment, when reading the rapid calculation results, the scheme evaluation module does not only read the total cost value, but also simultaneously reads the process node status in the coupling graph, the path record in the cost version chain, the correction record in the residual write-back table, and the consistency flag. The scheme evaluation module breaks down the cost result of each scheme into basic process cost, partial recalculation cost, and restricted residual cost. The basic process cost comes from the material balance and cost aggregation formula, the partial recalculation cost comes from the parameter influence index table and trigger expression, and the restricted residual cost comes from the historical residual restricted correction module. For schemes with constraint conflicts and incomplete boundary recalculation, the scheme evaluation module marks them as not directly comparable. They will be entered into the scheme comparison after the conflict node completes field completion or caliber correction. For schemes that pass the consistency verification, the scheme evaluation module forms evaluation data according to the cost per ton of raw ore, cost per ton of concentrate, unit metal cost, tailings disposal cost, and the source of each item cost. The module associates the node path and version identifier with each evaluation data. This embodiment enables the scheme evaluation results to be traced back to specific process nodes, cost items, and correction sources, avoiding the inability to locate the source of cost differences by only displaying the total amount.

[0047] In a preferred embodiment, the system sets up a verification chain for abnormal data. The verification chain consists of semantic verification, graph verification, constraint verification, and cost verification. Semantic verification is used to determine whether there are conflicts in the input fields regarding units, price versions, statistical periods, and aggregation calibers. Graph verification is used to determine whether there are missing edges, duplicate master aggregation edges, or isolated nodes in process nodes and cost item nodes. Constraint verification is used to determine whether the forward recursive value is consistent with the reverse verification boundary. Cost verification is used to determine whether there are unmapped residuals, out-of-boundary residuals, or cross-caliber aggregations in the sub-costs. When a semantic conflict is found, the system retains the original field and generates a field to be converted. After the field to be converted completes the unit conversion and caliber confirmation, it enters the graph node attributes. When a graph conflict is found, the system prevents the corresponding cost item from entering the summary. When a constraint conflict is found, the system executes conflict propagation boundary limitation. When a cost conflict is found, the system transfers the corresponding residual sample to the bypass sample pool. This embodiment allows different types of data anomalies to enter different processing chains, avoiding the direct propagation of abnormal fields to the scheme evaluation results.

[0048] In one embodiment, the overall data flow of the system is executed in the following order: scheme fields, graph nodes, constraint propagation, cost recalculation, residual correction, and scheme evaluation. Scheme fields, as the smallest traceable data object, enter the scheme data structuring module; graph nodes, as the shared carriers of process and cost, enter the process cost coupling graph construction module; constraint propagation, as a shared process of material closure and product indicator verification, enters the bidirectional constraint propagation calculation module; cost recalculation, as a partial update process after field changes, enters the associated cost increment recalculation module; residual correction, as a constrained mapping process of historical measured deviations, enters the historical residual constrained correction module; and scheme evaluation, as a process of reading cost versions... The process of expressing the results of the chain and residual write-back table runs; in each scheme update, the system only allows the graph path corresponding to the current field version to participate in the calculation, and the cost item value under the previous field version is retained as the historical version and cannot be mixed and summarized with the current version; the modules transmit data through field identifiers, node identifiers, edge identifiers, cost item identifiers and version identifiers, so that each sub-item cost can be traced back to the original field, process node, collection caliber, recalculation path and residual source. This embodiment forms a complete and auditable link from input data to scheme evaluation results, and makes the rapid calculation of mineral processing costs consistent in terms of process consistency, cost collection consistency and historical data correction consistency.

Claims

1. A system for rapid calculation of mineral processing costs and evaluation of treatment schemes, characterized in that, It includes a solution data structuring module, a process cost coupling graph construction module, a two-way constraint propagation calculation module, an associated cost incremental recalculation module, and a historical residual constraint correction module; The data structuring module of the solution converts the properties of raw ore, process combinations, product indicators, resource prices, and cost items into unified calculation fields; The process cost coupling graph construction module generates a coupling graph containing mineral processing process nodes, material flow relationships, and cost aggregation relationships based on the unified calculation fields. The bidirectional constraint propagation calculation module performs forward derivation of material balance and reverse verification of product indicators within the coupling graph. The associated cost increment recalculation module recalculates the affected cost items according to the node association relationship; The historical residual constraint correction module performs residual correction on the sub-item costs within the range of constraint propagation results.

2. The mineral processing cost rapid calculation and scheme evaluation system according to claim 1, characterized in that, The process cost coupling graph construction module sets the raw ore input node, crushing node, grinding node, classification node, beneficiation node, concentration and dewatering node, concentrate output node, tailings output node and cost item node as heterogeneous nodes, and establishes directed associations according to ore quantity transfer edge, grade recovery edge, unit consumption edge, cost collection edge and indicator constraint edge. The heterogeneous nodes are configured with attributes such as processing capacity, grade, recovery rate, yield, unit consumption, and unit price. The directed association configuration of source node, target node, constraint type, and recalculation flag ensures that cost items and corresponding mineral processing process nodes maintain the same data reference link.

3. The mineral processing cost rapid calculation and scheme evaluation system according to claim 1, characterized in that, The bidirectional constraint propagation calculation module includes forward material recursion processing and reverse index verification processing; The forward material recursion process calculates the amount of ore entering, the amount of ore exiting, the amount of metal, and the amount of tailings at each process node according to the process topology order. The reverse index verification process uses the target concentrate grade, target recovery rate, and allowable tailings metal loss as boundaries to back-transmit and mark the recovery rate, enrichment ratio, and loss rate of the beneficiation nodes. When the forward recursion result is inconsistent with the reverse verification boundary, the constraint conflict location and the corresponding set of process nodes are generated, and subsequent cost calculations are restricted to only referencing data links that have passed the consistency verification.

4. The mineral processing cost rapid calculation and scheme evaluation system according to claim 1, characterized in that, The associated cost increment recalculation module establishes a parameter influence index table, which records the correspondence between grinding fineness, feed grade, throughput, reagent system, recovery rate, concentrate yield, tailings yield and power consumption cost, reagent cost, media cost, water consumption cost, labor cost, and tailings disposal cost. When any scheme field changes, the associated cost incremental recalculation module extracts the affected process nodes and affected cost items based on the parameter impact index table, and performs local recalculation along the downstream direction of the coupling graph, starting from the affected process node.

5. The mineral processing cost rapid calculation and scheme evaluation system according to claim 2, characterized in that, The process cost coupling graph construction module sets a cost aggregation caliber field for the cost item nodes. The cost aggregation caliber field includes aggregation by raw ore quantity, aggregation by concentrate quantity, aggregation by operation quantity, and aggregation by tailings quantity. When the same cost item node is connected to multiple mineral processing nodes at the same time, a unique primary aggregation edge and at least one reference association edge are generated based on the cost aggregation caliber field. The unique primary aggregation edge participates in cost aggregation, the reference association edge participates in parameter influence index construction, and duplicate aggregation relationships and missing aggregation relationships are marked with graph consistency before cost aggregation.

6. The mineral processing cost rapid calculation and scheme evaluation system according to claim 3, characterized in that, The bidirectional constraint propagation calculation module configures a conflict propagation boundary for the constraint conflict location. The conflict propagation boundary is jointly determined by the nearest upstream material closure node, the nearest downstream product indicator node, and the associated cost item node. For process nodes located within the conflict propagation boundary, the ore quantity, grade, and recovery rate attributes are recalculated using the current scheme fields; For process nodes located outside the conflict propagation boundary, retain the attribute values ​​that have passed the consistency check; The cost recalculation starting point is formed by intersecting the conflict propagation boundary with the set of affected process nodes of the associated cost increment recalculation module.

7. The mineral processing cost rapid calculation and scheme evaluation system according to claim 4, characterized in that, The associated cost increment recalculation module sets a trigger expression for the affected cost item. The trigger expression is composed of process node attributes, cost aggregation caliber field and resource price version. When the resource price version changes but the process node attributes remain unchanged, the recalculation scope is limited to cost item nodes with the same price reference identifier. When the process node attributes change but the resource price version remains unchanged, the recalculation scope is limited to cost item nodes connected along the mineral quantity transfer edge and the unit consumption edge. When both change simultaneously, the two recalculation ranges will be merged and duplicate cost item nodes will be removed.

8. The mineral processing cost rapid calculation and scheme evaluation system according to claim 1, characterized in that, The historical residual constraint correction module extracts residual samples from the historical mineral processing scheme library based on ore property similarity, process topology similarity, cost item similarity, and data caliber consistency. The residual samples are categorized by power consumption, reagents, media, water consumption, labor allocation, and tailings disposal. The historical residual constraint correction module maps the sub-item residuals to the corresponding cost item nodes in the coupled graph, and uses the material balance result, cost aggregation caliber field and resource price version as correction boundaries to remove or scale residual samples that exceed the correction boundaries.

9. The mineral processing cost rapid calculation and scheme evaluation system according to claim 7, characterized in that, Before generating cost recalculation results, the associated cost incremental recalculation module establishes a cost version chain. The cost version chain records the unified calculation fields before the change, the unified calculation fields after the change, the recalculation starting point, the recalculation path, the unique main aggregation edge participating in the aggregation, and the reference associated edges not participating in the aggregation. When the same cost item node is accessed in different recalculation paths, the path record closest to the changed field is retained according to the cost version chain, and the remaining path records are set as bypass verification records. When summarizing costs, only the retained path records and their corresponding cost item node values ​​are referenced.

10. The mineral processing cost rapid calculation and scheme evaluation system according to claim 8, characterized in that, The historical residual restricted correction module establishes a residual write-back table, which records the source of historical residual samples, the type of sub-item residual, the corresponding cost item node, the correction boundary processing method, and the cost version identifier after correction. When obtaining the measured cost data in the future, the measured cost data will be split according to the cost item nodes, and the difference will be calculated with the item correction value corresponding to the corrected cost version identifier. The difference calculation results are updated with residual sample weights according to the consistency of historical residual sample sources and data caliber, and the updated residual sample weights are made available for the next scheme evaluation and rapid calculation.

Images