A Knowledge Graph-Based Decision-Making Method and System for Mine Energy Efficiency Optimization

Abstract

This invention discloses a knowledge graph-based decision-making method and system for mine energy efficiency optimization, specifically relating to the field of mine production energy efficiency management technology. It addresses the technical problem of conflicts between equipment-level energy-saving operations and system-level energy efficiency targets under existing hierarchical optimization models. By constructing an energy consumption transmission topology network integrating equipment-level, process-level, and enterprise-level entities, it achieves multi-level collaborative analysis of energy efficiency data. Based on topology path detection, it identifies energy flow vortex phenomena caused by equipment energy-saving operations, accurately locating conflict nodes leading to system energy consumption imbalances. It quantifies the degree of conflict impact through three-dimensional spatial energy density gradient deviation values ​​and calculates energy consumption transmission coefficients using historical data regression models. Finally, based on the transmission coefficients, it generates equipment-level control parameter optimization schemes that simultaneously meet the comprehensive energy consumption target per ton of ore and the energy consumption limit of the process chain, achieving a leap from local control to global collaboration in mine energy efficiency.

Description

Technical Field

[0001] This invention relates to the field of energy efficiency management technology in mine production, and more specifically, to a mine energy efficiency optimization decision-making method and system based on knowledge graphs. Background Technology

[0002] In the field of energy efficiency management in mining production, existing technologies generally adopt a hierarchical independent optimization strategy. Equipment-level control systems (such as PLCs) adjust the operating parameters of individual equipment in real time to achieve local energy saving, while enterprise-level management systems (such as ERP) set long-term goals based on macro-level energy efficiency indicators. The two types of systems are disconnected in terms of data hierarchy, time scale, and decision-making objectives: equipment-level optimization focuses on second-level real-time data and minimizing the energy consumption of a single machine, while enterprise-level management relies on monthly statistical indicators and comprehensive energy consumption control per ton of ore. This hierarchical model results in a lack of coordination mechanism between the bottom-level equipment control commands and the top-level energy efficiency strategy.

[0003] Existing technologies cause conflicts in energy efficiency targets at multiple scales due to hierarchical fragmentation: energy-saving operations at the equipment level may disrupt the energy efficiency balance at the system level. For example, energy-saving strategies for a single piece of equipment (such as reducing ventilation frequency) may trigger a chain reaction of increased energy consumption in downstream processes (such as a surge in cooling load due to increased underground temperature), causing local optimization to deviate from the global objective. It is impossible to quantify the transmission impact of equipment actions on system energy consumption, resulting in the failure of overall energy efficiency optimization decisions in the mine. Summary of the Invention

[0004] In order to overcome the above-mentioned deficiencies of the prior art, the present invention provides a mine energy efficiency optimization decision-making method and system based on knowledge graphs to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: Knowledge graph-based decision-making methods for mine energy efficiency optimization include: S1. Obtain real-time operating parameters of equipment and enterprise-level energy efficiency indicators in the mining production system; S2. Construct an industrial knowledge graph to connect equipment-level entities, process-level entities, and enterprise-level entities through energy consumption transmission relationships to form a cross-level topology network; S3. Based on the industrial knowledge graph, topology path analysis identifies conflict nodes by detecting the energy flow vortex phenomenon caused by equipment-level energy-saving operations on the topology path. The energy flow vortex phenomenon is manifested as the formation of a closed loop in the local energy consumption conduction path. S4. Monitor the spatial energy density imbalance caused by conflict nodes through the three-dimensional topological relationship of the industrial knowledge graph, and calculate the energy distribution gradient deviation value. S5. Based on the energy distribution gradient deviation, calculate the energy transmission coefficient of the conflict node using a historical data regression model; S6. Adjust the equipment-level control parameters based on the energy consumption transmission coefficient to generate a global optimization decision scheme.

[0006] Furthermore, obtain real-time operating parameters at the equipment level and enterprise-level energy efficiency indicators in the mining production system, including: Real-time operating parameters at the equipment level are collected from the equipment control system. These parameters include equipment power consumption data, equipment operating frequency data, and equipment temperature data. Enterprise-level energy efficiency indicators are periodically obtained from the enterprise resource management system. These indicators include the comprehensive energy consumption target per ton of ore and the energy consumption limit per unit of the process chain.

[0007] Furthermore, an industrial knowledge graph is constructed, linking equipment-level entities, process-level entities, and enterprise-level entities through energy consumption transmission relationships to form a cross-level topology network, including: A device-level entity is created based on the device-level real-time operating parameters. The device-level entity includes a device identifier, a device type attribute, and a real-time power consumption attribute. Based on the division of mining production processes, process-level entities are created. Each process-level entity includes a process identifier, a process type attribute, and an energy consumption baseline attribute. Create an enterprise-level entity and configure enterprise-level energy efficiency indicator attributes; Based on the physical connection and control logic of the equipment in the mining production system, establish the subordinate relationship between equipment-level entities and process-level entities; Based on the upstream and downstream sequence of mining production processes, establish the energy consumption transmission relationship between process-level entities; Enterprise-level entities and process-level entities are linked through global energy consumption constraints to form a cross-level topology network.

[0008] Furthermore, the energy consumption transmission relationship includes the energy consumption impact chain of equipment operation on downstream processes.

[0009] Furthermore, based on topology path analysis using industrial knowledge graphs, conflict nodes are identified by detecting energy flow vortices caused by equipment-level energy-saving operations on the topology path, including: Locate the source device-level entity that implements device-level energy-saving operations on the topological path of the industrial knowledge graph; Track the propagation path of energy efficiency disturbances caused by energy-saving operations along the direction of energy consumption transmission, and record the energy consumption deviation values ​​of process-level entities in the propagation path; Detect whether there is a closed-loop transmission path in the topology that starts from the source device level entity and returns to the source device level entity through the process level entity; When the cumulative energy consumption deviation of the closed-loop transmission path exceeds the energy consumption baseline attribute of the process-level entity and forms a self-excited oscillation trend, all entity nodes included in the closed-loop transmission path are marked as conflict nodes.

[0010] Furthermore, by monitoring the spatial energy density imbalance caused by conflicting nodes through the three-dimensional topological relationships of the industrial knowledge graph, the energy distribution gradient deviation is calculated, including: Based on the three-dimensional topological relationship of industrial knowledge graph, the spatial coordinates of conflict nodes in the three-dimensional space of the mine are obtained; Centered on the conflict node, the measured energy density values ​​of adjacent process-level entities within a preset range are collected along the spatial transmission direction of the three-dimensional topological relationship. Calculate the three-dimensional spatial distance between the conflict node and each adjacent process-level entity based on the spatial location coordinates; Based on the measured values ​​of three-dimensional spatial distance and energy density, a spatial energy density distribution field of the local region where the conflict node is located is constructed. The magnitude of the gradient vector of the spatial energy density distribution field is calculated as the energy distribution gradient deviation value.

[0011] Furthermore, based on the energy distribution gradient deviation, the energy transmission coefficient of the conflict nodes is calculated using a historical data regression model, including: Obtain the historical energy distribution gradient deviation sequence of conflict nodes and the corresponding historical energy consumption conduction data; Based on the equipment type and process type attributes of the conflict nodes, a historical regression sample set of the same type of nodes is selected from the historical energy efficiency database. A multiple linear regression model was trained using historical energy distribution gradient deviation as the independent variable and historical energy consumption transmission data as the dependent variable. Input the current energy distribution gradient deviation into the trained multiple linear regression model, and output the energy transmission coefficient of the conflict node.

[0012] Furthermore, based on the energy consumption transmission coefficient, the equipment-level control parameters are adjusted to generate a global optimization decision scheme, including: The correction amount of equipment-level control parameters is calculated based on the energy consumption transmission coefficient; The correction amount is compared with the target value of comprehensive energy consumption per ton of mine in the enterprise-level energy efficiency index to screen the set of effective correction amounts that meet the target value of comprehensive energy consumption per ton of mine. Verify the feasibility of the effective correction set within the unit energy consumption limit value of the process chain based on the energy consumption baseline attributes of the process-level entities. The equipment-level control parameter corrections that simultaneously meet the comprehensive energy consumption target value per ton of ore and the unit energy consumption limit value of the process chain are combined to generate a global optimization decision scheme.

[0013] Furthermore, the equipment-level control parameters include the equipment operating frequency setpoint and the equipment power threshold.

[0014] On the other hand, the present invention provides a mine energy efficiency optimization decision system based on knowledge graphs, comprising: The parameter acquisition module is used to acquire real-time operating parameters of equipment and enterprise-level energy efficiency indicators in the mining production system. The topology modeling module is used to build an industrial knowledge graph, which connects equipment-level entities, process-level entities, and enterprise-level entities through energy consumption transmission relationships to form a cross-level topology network. The vortex detection module is used for topology path analysis based on industrial knowledge graphs. It identifies conflict nodes by detecting energy flow vortex phenomena caused by energy-saving operations at the device level on the topology path. The gradient calculation module is used to monitor the spatial energy density imbalance caused by conflict nodes through the three-dimensional topological relationship of the industrial knowledge graph, and to calculate the energy distribution gradient deviation value. The conduction modeling module is used to calculate the energy conduction coefficient of conflict nodes based on the energy distribution gradient deviation value and a regression model of historical data. The collaborative optimization module is used to adjust equipment-level control parameters based on the energy consumption transmission coefficient and generate a global optimization decision scheme.

[0015] Compared with the prior art, the present invention has the following beneficial effects: 1. By constructing an energy consumption transmission topology network of equipment-level, process-level, and enterprise-level entities, dynamic collaborative optimization of multi-level energy efficiency targets in mines can be achieved; based on the cross-level association mechanism of industrial knowledge graph, real-time operating parameters of equipment and comprehensive energy consumption indicators per ton of mine of enterprises are integrated and analyzed in a unified topology space, breaking through the data barriers of traditional hierarchical optimization mode; by accurately identifying the energy flow vortex phenomenon caused by energy-saving operation of equipment on the topology path, conflict nodes that cause self-excited oscillation of system energy consumption can be located in real time, fundamentally solving the technical contradiction between local optimization and global energy efficiency targets.

[0016] 2. A three-dimensional spatial energy density gradient deviation analysis is introduced to quantify the degree of spatial energy imbalance caused by conflict nodes; an energy consumption transmission coefficient calculation mechanism is established by combining historical data regression models to accurately characterize the transmission effect of equipment-level operations on process chain energy consumption; finally, through multi-objective constraint collaborative decision-making, an optimized equipment control parameter scheme that simultaneously meets the comprehensive energy consumption target per ton of ore and the unit energy consumption limit of the process chain is generated; this technical path achieves: first, establishing a quantitative transmission model of equipment actions and system energy efficiency to eliminate blind spots in hierarchical decision-making; second, predicting energy consumption imbalance trends through spatial energy field gradient analysis to improve the foresight of optimization; and third, constructing a triple constraint verification mechanism at the enterprise level, process level, and equipment level to ensure the global feasibility of the optimization scheme in terms of time scale and spatial dimension. Attached Figure Description

[0017] Figure 1 This is a flowchart of the knowledge graph-based mine energy efficiency optimization decision-making method of the present invention. Figure 2 This is a schematic diagram of the structure of the knowledge graph-based mine energy efficiency optimization decision system of the present invention. Detailed Implementation

[0018] 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 embodiments of the present invention, and not all embodiments. 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. Example

[0019] Figure 1 This invention presents a knowledge graph-based decision-making method for mine energy efficiency optimization, comprising: S1. Obtain real-time operating parameters of equipment and enterprise-level energy efficiency indicators in the mining production system; S2. Construct an industrial knowledge graph to connect equipment-level entities, process-level entities, and enterprise-level entities through energy consumption transmission relationships to form a cross-level topology network; S3. Based on the industrial knowledge graph, topology path analysis identifies conflict nodes by detecting the energy flow vortex phenomenon caused by equipment-level energy-saving operations on the topology path. The energy flow vortex phenomenon is manifested as the formation of a closed loop in the local energy consumption conduction path. S4. Monitor the spatial energy density imbalance caused by conflict nodes through the three-dimensional topological relationship of the industrial knowledge graph, and calculate the energy distribution gradient deviation value. S5. Based on the energy distribution gradient deviation, calculate the energy transmission coefficient of the conflict node using a historical data regression model; S6. Adjust the equipment-level control parameters based on the energy consumption transmission coefficient to generate a global optimization decision scheme.

[0020] S1. Obtain real-time operating parameters of equipment and enterprise-level energy efficiency indicators in the mining production system. The specific implementation is as follows: A data acquisition interface is deployed in the equipment control system of the mining production system. This interface establishes a physical connection with the programmable logic controllers (PLCs) of the crusher, ball mill, and conveyor belt equipment via an industrial Ethernet network. The data acquisition interface reads real-time operating parameters from the PLC registers at a frequency of once per second. A cyclic redundancy check (CRC) mechanism ensures data integrity during the reading process; if a check fails, the reading request is automatically re-initiated until successful. Equipment power consumption data in the real-time operating parameters is obtained from smart meters connected to the equipment's power distribution cabinet. The instantaneous power value, in kilowatt-hours, is written to a specific register address in the PLC. Equipment operating frequency data is read from the inverter's operating status register, representing the actual rotational frequency of the motor drive shaft in Hertz (Hz). Equipment temperature data is collected by temperature sensors embedded in the equipment bearings. The raw voltage value is converted to Celsius by multiplying the voltage sample value by a range factor and adding a calibration offset. For example, the typical calibration offset for a PT100 temperature sensor is 0, and the range factor is 250. The converted temperature value is stored in a designated register of the PLC. The data acquisition interface adds millisecond-level timestamps to the above three types of data and packages them into a specific format message, which is then transmitted to the central processing server in real time via the message queue telemetry transmission protocol.

[0021] The central processing server accesses the energy efficiency index database of the enterprise resource management system (ERP) at a fixed time each day via a database connection interface. This access uses a service account with read-only permissions for authentication. The comprehensive energy consumption target per ton of ore in the ERP is stored in a dedicated database table. This value represents the baseline energy consumption set by the enterprise, in kilowatt-hours per ton. For example, an iron ore mine might set an annual target of 5.8 kilowatt-hours per ton. The energy consumption limit per unit of the process chain is stored in another dedicated database table, which stores the limit value by process chain identifier. For example, the limit for the crushing process chain is no more than 3.2 kilowatt-hours of electricity per ton of ore processed. The central processing server executes a database query command to retrieve the daily comprehensive energy consumption target per ton of ore and the latest effective energy consumption limit per unit of the process chain. The retrieved data is stored in a memory cache and simultaneously written to the historical energy efficiency index database of the persistent storage system, with a date identifier automatically added during storage.

[0022] A mapping relationship is established between the storage structure of equipment-level real-time operating parameters and enterprise-level energy efficiency indicators, specifically through the association between equipment identifiers and process chain identifiers. For example, the power consumption data of a specific crusher in a crushing process chain is associated with the unit energy consumption limit value of the process chain through the process chain identifier to which the equipment belongs. When transmitting equipment-level real-time operating parameters, the data acquisition interface includes a unique equipment identifier field in the message. The central processing server uses this identifier to query the equipment process association table to obtain the corresponding process chain identifier. All collected data undergoes validity verification. When the equipment temperature data exceeds the equipment material tolerance threshold, the data point is automatically discarded and an alarm log is generated. The equipment material tolerance threshold is set according to the technical parameters provided by the equipment manufacturer; for example, the temperature threshold for ball mill bearings is set to 120 degrees Celsius.

[0023] The central processing server performs data cleaning on the collected real-time data. Cleaning rules include: replacing negative device power consumption data with the moving average of the previous few valid sampling points; truncating device operating frequency data when it exceeds a certain percentage of the rated frequency (the rated frequency value is derived from the device's technical documentation); and discarding device temperature data points outside the set temperature range, for example, setting the lower limit to -10 degrees Celsius and the upper limit to 200 degrees Celsius. The cleaned data is stored in a dedicated data table in a time-series database, which is sharded by device number. Enterprise-level energy efficiency indicators are stored in a dedicated table in a relational database, containing core fields such as date, indicator type, and value. During storage, a data integrity check code is automatically generated, calculated using a specific hash algorithm.

[0024] To ensure real-time data transmission, a priority transmission channel is established between the equipment control system and the central processing server. This channel prioritizes the transmission of equipment operating frequency data and equipment temperature data. The enterprise resource management system and the central processing server employ a breakpoint resume mechanism, automatically recording the breakpoint location when the network is interrupted. All data acquisition operations are logged in detail, including fields such as operation timestamp, data source, data volume, and processing status. Log files are archived and stored daily.

[0025] S2. Construct an industrial knowledge graph, linking equipment-level entities, process-level entities, and enterprise-level entities through energy consumption transmission relationships to form a cross-level topology network. The specific implementation is as follows: Equipment-level entities are created based on real-time operating parameters collected from the equipment control system. The specific creation process is as follows: The central processing server extracts the latest real-time operating parameter records from the time-series database. Each record contains an equipment identifier field. A corresponding equipment-level entity node is created for each unique equipment identifier. The equipment identifier is generated using a unified mining coding rule; for example, the identifier format for a cone crusher is "PJ-Production Line Number-Equipment Serial Number," where the production line number is a 3-digit number and the equipment serial number is a 2-digit number. The equipment type attribute is set according to predefined classification standards in the equipment archive. The classification standards are based on equipment function; for example, jaw crushers and cone crushers are classified as crushing equipment, and ball mills are classified as grinding equipment. The real-time power consumption attribute is directly obtained from the equipment power consumption data in the real-time operating parameters. This attribute value is updated in real-time with the data acquisition interface, and the update cycle is consistent with the data acquisition frequency, for example, once per second. Equipment-level entities are stored in the form of graph database nodes. The node attributes include three fields: equipment identifier, equipment type attribute, and real-time power consumption attribute.

[0026] Work-level entities are created based on the division of mining production processes. The specific creation process is as follows: The process configuration table of the production management system is parsed, which records the correspondence between process chains and process steps. An independent work-level entity node is created for each process step. The process identifier is generated by combining the process chain number and the process step number; for example, the first process in the crushing process chain is identified as "Crushing Chain 001 - Coarse Crushing Step". The process type attribute is set according to the technological function classification, specifically according to the ore processing stage, such as coarse crushing stage, medium crushing stage, fine crushing stage, grinding stage, and beneficiation stage. The energy consumption baseline attribute is determined by statistically analyzing the average energy consumption value of the process under historical normal production conditions. The specific calculation method is: selecting data from the same time period of the most recent specific number of days, excluding data from equipment maintenance days and days with production abnormalities, and then calculating the arithmetic mean; for example, selecting data from the most recent 30 production days, with daily data collection time from 9:00 to 18:00. The work-level entity node stores three fields: process identifier, process type attribute, and energy consumption baseline attribute.

[0027] Create an enterprise-level entity node and configure its enterprise-level energy efficiency index attributes. The enterprise-level entity is a single node, and its enterprise-level energy efficiency index attribute values ​​are derived from the comprehensive energy consumption target per ton of ore obtained in step S1 and the unit energy consumption limit value of the process chain. This node attribute is updated daily, with the update time consistent with the data synchronization time of the enterprise resource management system, for example, updating at 00:10 daily. Establish a global association between the enterprise-level entity and the process-level entities. Specifically, this association is achieved by creating a global constraint edge between the enterprise-level entity node and the first process node of each process chain.

[0028] Establish subordinate relationships between equipment-level entities and process-level entities. The physical connection relationship is determined by the equipment's installation coordinates in the mine's 3D space, which are derived from the mine's digital modeling system. The specific association method is as follows: query the equipment's location coordinates using its identifier, match these coordinates with the spatial range of the process area, and establish a subordinate relationship when the equipment coordinates are within the spatial range of a specific process area. For example, the equipment coordinate point (102.35, 35.78, -15.2) is located within the boundary box of the crushing process area numbered A-3. The control logic relationship is determined by the equipment control signal transmission path, which is derived from the distributed control system configuration table. When the control signal output of equipment A is connected to the control signal input of equipment B, a control logic relationship is established between equipment A and equipment B. For example, the crusher's operating status signal is connected to the input of the feeder controller. These two relationships are stored in the graph database as physical connection edges and control logic relationship edges, respectively.

[0029] Establish energy consumption transmission relationships between process-level entities. The upstream and downstream sequence relationships are determined based on the production process flow chart, which is derived from the mine production planning system. Specifically, the process connection direction in the flow chart is analyzed. When the output material of process A directly enters process B, an energy consumption transmission relationship edge is established from process A to process B. The quantification of the energy consumption impact chain is achieved through equipment action impact coefficients, which are obtained based on historical operating data analysis. The specific calculation process involves collecting downstream process energy consumption change data during specific equipment parameter adjustment periods. A mathematical relationship between parameter changes and energy consumption fluctuations is established using regression analysis. For example, for every 1 Hz decrease in crusher frequency, the average energy consumption of the downstream grinding process increases by 0.35%. This percentage is stored as an impact coefficient in the relationship edge attributes. The impact coefficient is updated monthly.

[0030] Construct a cross-level topology network. Enterprise-level entity nodes are connected to the first process node of each process chain via global constraint edges. For example, an enterprise-level node is connected to the first process node of crushing process chain 001. Within each process chain, energy consumption transmission chains are established based on material flow direction, such as a transmission path from coarse crushing process node to medium crushing process node to fine crushing process node. Equipment-level entity nodes are connected to their corresponding process-level entity nodes via subordinate association edges. For example, the crusher node numbered PJ-001-01 is associated with the coarse crushing process node. The final topology network contains three types of entity nodes and four types of relationship edges: equipment-level entity nodes carry equipment identifiers, equipment type attributes, and real-time power consumption attributes; process-level entity nodes carry process identifiers, process type attributes, and energy consumption baseline attributes; enterprise-level entity nodes carry enterprise-level energy efficiency index attributes; physical connection edges represent equipment spatial location associations; control association edges represent equipment control logic; energy consumption transmission edges carry influence coefficient attributes; and global constraint edges represent enterprise-level energy consumption index constraints. The topology network is stored in a graph database, and network structure updates are triggered when new equipment is added or processes are adjusted.

[0031] S3. Based on industrial knowledge graph-based topology path analysis, conflict nodes are identified by detecting energy flow vortex phenomena caused by equipment-level energy-saving operations on the topology path. Specifically, this is implemented as follows: When performing topology path analysis within the industrial knowledge graph's topology network, the source device-level entity implementing the energy-saving operation is first located. The specific location method involves real-time monitoring of the device-level entity's real-time power consumption attribute values. An energy-saving operation is determined to have occurred when the power consumption decrease exceeds a set threshold within a specific time period. This set threshold is dynamically adjusted based on the device's historical operating data, for example, taking a specific multiple of the standard deviation of the device's normal fluctuation range over a specific historical time period. The source device-level entity is identified by comparing the current power consumption value with a historical benchmark value. The historical benchmark value is the average power consumption value of the device over a specific number of days under the same production load conditions. The location results record the complete device identifier of the source device-level entity and the complete process identifier of its associated process-level entity.

[0032] The propagation path of energy efficiency disturbances caused by energy-saving operations is traced along the energy consumption transmission relationship. The tracing process begins at the process-level entity node to which the source device-level entity belongs, and traverses downstream process-level entity nodes according to the direction of the energy consumption transmission relationship edges in the topology network. Upon reaching each process-level entity node, the real-time energy consumption deviation value of that node is calculated. The formula for calculating the real-time energy consumption deviation value is the current real-time power consumption attribute value minus the node's energy consumption baseline attribute value, divided by the percentage of the energy consumption baseline attribute value. For example, if the energy consumption baseline attribute value of a process node is a specific kilowatt-hour value, and the current real-time power consumption attribute value is another specific kilowatt-hour value, the specific percentage difference is calculated. The propagation path is recorded as an ordered sequence of nodes, where each node corresponds to the complete process identifier and real-time energy consumption deviation value of the process-level entity.

[0033] The system detects whether a closed-loop propagation path exists within the topology. The detection process includes: searching the recorded propagation path node sequence for the process-level entity node to which the source device-level entity belongs. When the node is found in the sequence, the path closure is further verified: starting from this node, the upstream node is traced along the reverse path of the energy consumption propagation relationship. If the path can return to the process-level entity node to which the source device-level entity belongs through continuous propagation relationship edges, then a closed-loop propagation path is determined to be formed. For example, there exists a propagation path of "crushing process node → grinding process node → beneficiation process node → crushing process node" where each node has continuous energy consumption propagation relationship edges. The closed-loop path record contains a complete sequence of process identifiers for all process-level entity nodes on the complete path.

[0034] Determine if the conflict node marking conditions are met. The cumulative energy consumption deviation is calculated as the algebraic sum of the absolute values ​​of the real-time energy consumption deviations of all process-level entity nodes on the closed-loop transmission path. The self-oscillation trend is determined by: within a specific time window, detecting whether the real-time energy consumption deviation of each node on the closed-loop path shows a continuous increasing trend. When the deviation increment of multiple consecutive samples exceeds a specific increase threshold, a self-oscillation trend is determined to have formed. For example, the detection window is set to a specific number of minutes, and the deviation increment of a specific number of consecutive samples exceeds a specific percentage value. When the cumulative energy consumption deviation exceeds a specific percentage threshold of the energy consumption baseline attribute value of the process-level entity and a self-oscillation trend is formed, all process-level entity nodes and their associated equipment-level entity nodes included in the closed-loop transmission path are marked as conflict nodes. Conflict nodes record their complete equipment identifier or complete process identifier and conflict marking attributes.

[0035] After the marking operation is completed, the status attribute of the conflict node is updated to "pending processing". The path analysis process is executed cyclically at fixed time intervals, and local path analysis is triggered immediately when a new energy-saving operation event is detected. The path traversal adopts a depth-first search algorithm, and the maximum search depth is set to a specific number of levels to control computational complexity. For example, the search stops after a maximum of penetrating a certain number of process nodes. When a cyclic path is detected, the search in that direction is automatically terminated to avoid infinite loops. All analysis results are stored in a dedicated storage area and a version tag is created.

[0036] S4. Monitor the spatial energy density imbalance caused by conflict nodes through the three-dimensional topological relationships of the industrial knowledge graph, and calculate the energy distribution gradient deviation value. The specific implementation is as follows: The spatial coordinates of conflict nodes in the 3D space of the mine are obtained based on the 3D topological relationships of an industrial knowledge graph. Specifically, for equipment-level conflict nodes, the equipment location database is queried using the equipment identifier. This database stores the coordinates of the equipment base center point obtained from total station surveying. For process-level conflict nodes, the process area spatial model library is queried using the process identifier. This model library records the calculated coordinates of the geometric center of the process area. The spatial coordinates are represented using a 3D numerical representation of the mine's global coordinate system, with the origin being the mine's reference point. The unit is meters, and the coordinate accuracy is controlled at a specific centimeter level. For example, the coordinate value format of a cone crusher equipment node includes three components: east-west offset, north-south offset, and altitude. The coordinate data is updated synchronously at specific intervals, triggering re-surveying when equipment is moved or the process area is modified.

[0037] Centered on the conflict node, the system collects measured energy density values ​​of adjacent process-level entities within a preset range along the spatial transmission direction of the three-dimensional topological relationship. The preset range is defined as a spherical region centered on the coordinates of the conflict node. The radius of the sphere is dynamically set according to the process type attribute; for example, a specific radius is used for the crushing process region, and another specific radius is used for the grinding process region. The spatial transmission direction is determined based on the material flow direction between processes recorded in the topological relationship, such as along the centerline of the ore conveyor belt. The calculation process for the measured energy density is as follows: First, obtain the real-time power consumption attribute values ​​of all equipment-level entities associated with adjacent process-level entities and calculate the sum of these attribute values; second, query the spatial volume value corresponding to the process-level entity, which is calculated based on the length, width, and height dimensions of the process region; finally, divide the sum of power consumption by the spatial volume value to obtain the measured energy density value, expressed in kilowatt-hours per cubic meter. For example, the spatial volume of a grinding process region is the product of its length, width, and height dimensions. The collection range covers all directly associated process-level entity nodes within the sphere.

[0038] The three-dimensional spatial distance between conflicting nodes and adjacent process-level entities is calculated based on their spatial coordinates. The distance calculation uses the Euclidean distance algorithm: the difference between the coordinates of the conflicting node and its adjacent nodes along the X-axis is taken, and the square of this difference is calculated. Similarly, the squares of the differences along the Y-axis and Z-axis are calculated. The squares in all three directions are summed to obtain the total value. Finally, the square root of the sum is taken to obtain the distance value. For example, the spatial coordinates of a conflicting node contain three components: X-axis, Y-axis, and Z-axis coordinates. The spatial coordinates of adjacent nodes also contain these three components. The distance calculation process is as follows: first, the difference between the X-axis coordinates of the two nodes is calculated, and this difference is multiplied by itself to obtain the square of the X-axis difference. Similarly, the squares of the differences in the Y-axis and Z-axis coordinates are calculated. Then, the three squares are summed to obtain the total value. Finally, the square root of the sum is taken to obtain the distance value. The calculation results are in meters, retaining a specific decimal place precision. The calculation process iterates through all acquired adjacent process-level entity nodes, generating a list of distance values.

[0039] Based on measured values ​​of 3D spatial distance and energy density, a spatial energy density distribution field is constructed for the local area where the conflict node is located. The construction method employs an inverse distance-weighted spatial interpolation algorithm: the conflict node's location is set as a reference point, and surrounding adjacent nodes are set as sampling points. The weight of each sampling point is equal to the reciprocal of a specific power of its distance to the reference point. For example, a sampling point a certain distance from the reference point has a weight coefficient equal to the reciprocal of the square of its distance value. The spatial energy density distribution field is discretized into a 3D mesh structure, with the mesh cell size set to a cube with a specific meter dimension, such as a cube mesh with a specific meter side length. The energy density value of each mesh vertex is calculated by the weighted average of the measured energy density values ​​of sampling points within a specific range around it, using the aforementioned inverse distance-weighted value as the weight coefficient. The distribution field covers the process area where the conflict node is located and adjacent buffer zones within a specific meter dimension.

[0040] The magnitude of the gradient vector of the spatial energy density distribution field is calculated as the energy distribution gradient deviation value. The gradient vector is calculated using the central difference method: a reference point is located in the distribution field grid, and the difference between the energy density values ​​of adjacent grid points in the positive X direction and the adjacent grid points in the negative X direction is calculated. This difference is divided by twice the grid spacing to obtain the X-direction gradient component. The Y-direction and Z-direction gradient components are calculated using the same method. The magnitude of the gradient vector is equal to the square root of the sum of the squares of the X-direction, Y-direction, and Z-direction gradient components. This magnitude is output as the energy distribution gradient deviation value, reflecting the spatial rate of change of the energy distribution.

[0041] The spatial energy density distribution field is reconstructed at specific time intervals. Reconstruction is triggered by the arrival of the time period or the detection of a sudden energy density event. An early warning signal is generated when the energy distribution gradient deviation exceeds a set threshold. This threshold is set based on the statistical distribution of gradient values ​​under historical normal operating conditions, for example, by taking the mean plus a specific multiple of the standard deviation. Detailed logs are recorded for all calculation processes, including timestamps, conflict node identifiers, and gradient deviation values.

[0042] S5. Based on the energy distribution gradient deviation, the energy transmission coefficient of the conflict node is calculated using a historical data regression model. The specific implementation is as follows: Obtain the historical energy distribution gradient deviation value sequence and corresponding historical energy consumption transmission data for conflict nodes. Specifically, extract historical records within the most recent specific time range from the historical operation database using the complete equipment identifier or complete process identifier of the conflict node. The historical energy distribution gradient deviation value sequence refers to the numerical sequence of energy distribution gradient deviation values ​​generated by the node each time within a specific number of days in the past, arranged in chronological order. Historical energy consumption transmission data refers to the actual change in energy consumption of downstream processes recorded at the same time. This change is calculated by comparing the percentage change in real-time power consumption attribute values ​​of downstream processes within a specific time window before and after the implementation of energy-saving operations. For example, for a certain crusher conflict node, extract all energy distribution gradient deviation values ​​and their corresponding downstream grinding process energy consumption change records within a specific production day in the past. Data records from equipment downtime and periods of production abnormality are excluded during data extraction.

[0043] Based on the equipment type and process type attribute values ​​of conflicting nodes, a historical regression sample set of nodes of the same type is selected from the historical energy efficiency database. The selection rule is: select all nodes with completely identical equipment type and process type attribute values. The historical energy efficiency database stores complete data recorded by all equipment-level and process-level entities during historical operation. The sample set construction process is as follows: first, determine the specific content of the equipment type and process type attribute values ​​of the conflicting nodes; second, retrieve all node records with the same equipment type and process type attribute values ​​from the database; finally, extract the historical energy distribution gradient deviation value sequence and the corresponding historical energy consumption transmission data of these nodes to form the sample set. For example, when the equipment type attribute of a conflicting node is crushing equipment and the process type attribute is coarse crushing process, select all historical data of nodes with the equipment type attribute of crushing equipment and the process type attribute of coarse crushing process. The sample set size requires no less than a certain number of valid sample points to ensure statistical reliability.

[0044] A multiple linear regression model is trained using historical energy distribution gradient deviation as the independent variable and historical energy consumption transmission data as the dependent variable. The training process includes a data preprocessing stage and a model optimization stage. In the data preprocessing stage, the sample data is standardized: the mean and standard deviation of the independent variables are calculated, and each independent variable value is subtracted from the mean and divided by the standard deviation; the dependent variable data is processed in the same way. In the model optimization stage, the least squares method is used for parameter estimation: the model constant term parameters and slope coefficient parameters are initialized; the parameter values ​​are adjusted through iterative calculations to minimize the sum of squared residuals between the predicted and actual values. The iteration stopping condition is set to the change in the sum of squared residuals between two consecutive iterations being less than a specific threshold or the number of iterations reaching a specific upper limit. After training, the model determination coefficient is calculated; when the determination coefficient is greater than a specific threshold, the model is considered effective. When storing model parameters, they are associated with the corresponding equipment type attribute values ​​and process type attribute values.

[0045] The current energy distribution gradient deviation is input into a trained multiple linear regression model, which outputs the energy transmission coefficient of conflict nodes. The input processing involves standardizing the current energy distribution gradient deviation in the same way as the training data: subtracting the mean of the training sample independent variables and dividing by the standard deviation of the training sample independent variables. The model calculation process involves multiplying the standardized current value by the model slope coefficient and adding the model constant term to obtain the standardized predicted value. The result transformation process involves destandardizing the standardized predicted value: multiplying it by the standard deviation of the training sample dependent variable and adding the mean of the training sample dependent variable, finally outputting the energy transmission coefficient value in actual units. For example, when the current energy distribution gradient deviation is a specific value, the model calculates and outputs an energy transmission coefficient of a specific percentage value. The calculation result is appended with a timestamp and node identifier and stored in the database. The energy transmission coefficient represents the percentage change rate of energy consumption per unit time in downstream processes.

[0046] The multiple linear regression model is updated and maintained at specific intervals. Model retraining is triggered when the cumulative number of newly added historical data records reaches a specific threshold. Prediction accuracy is monitored in real time during model application; when the absolute value of the prediction error exceeds the allowable range for a specific number of consecutive times, a model failure alarm is generated and an emergency retraining process is initiated. All computational tasks are executed on a distributed computing cluster, and detailed operation logs are recorded for key computational steps.

[0047] S6. Adjust the equipment-level control parameters based on the energy consumption transmission coefficient to generate a global optimization decision scheme, specifically implemented as follows: The correction amount for equipment-level control parameters is calculated based on the energy consumption transmission coefficient. The specific calculation process is as follows: The energy consumption transmission coefficient value of the conflict node is obtained, representing the percentage change in energy consumption of downstream processes caused by a unit energy-saving operation. The correction amount for the equipment operating frequency setpoint is calculated by multiplying the energy consumption transmission coefficient by the frequency influence factor and then by a negative one value. The frequency influence factor is obtained from the equipment characteristic database based on the equipment type attribute value and represents the sensitivity of equipment operating frequency changes to energy consumption transmission. The correction amount for the equipment power threshold is calculated by multiplying the energy consumption transmission coefficient by the power adjustment coefficient, which is derived from historical equipment operating data. For example, the frequency influence factor for a crusher may be a specific value, and the power adjustment coefficient may be another specific value. The correction amount calculation considers the current operating state of the equipment. When the equipment is under high load, a load compensation coefficient is introduced, calculated based on the real-time load rate of the equipment. All calculation parameters are derived from the equipment characteristic database, which stores the operating characteristic parameters of various types of equipment.

[0048] The correction amount is compared with the target value of comprehensive energy consumption per ton of ore in the enterprise-level energy efficiency index. The constraint comparison method is as follows: predict the comprehensive energy consumption per ton of ore after implementing the correction amount. This predicted value is equal to the current comprehensive energy consumption per ton of ore plus the expected change in energy consumption resulting from the correction amount. The expected change in energy consumption is calculated by multiplying the energy consumption transmission coefficient by the correction amount. The constraint condition is set that the predicted value must be less than or equal to the target value of comprehensive energy consumption per ton of ore. The screening process traverses all possible combinations of correction amounts, retaining the combinations that satisfy the constraint condition to form a set of effective correction amounts. For example, when the target value of comprehensive energy consumption per ton of ore is a specific kilowatt-hour per ton, all combinations of correction amounts that ensure the predicted value does not exceed the target value are screened out. The comparison process uses an iterative optimization algorithm, setting a maximum number of iterations to a specific value to prevent infinite loops. Each iteration generates a specific number of candidate combinations of correction amounts.

[0049] The feasibility of the effective set of corrections within the unit energy consumption limit of the process chain is verified based on the energy consumption baseline attributes of process-level entities. The verification method is as follows: for each combination of corrections in the effective set, its propagation impact on the process chain is calculated. First, the starting process-level entity for implementing the corrections is determined; second, the energy consumption change is calculated process by process along the energy consumption transmission direction, where the energy consumption change equals the energy consumption change of the upstream process multiplied by the energy consumption transmission coefficient between processes; finally, the energy consumption changes of each process are summed to obtain the total energy consumption change of the process chain. The feasibility judgment condition is: the unit energy consumption value of the corrected process chain equals the current unit energy consumption value of the process chain plus the total energy consumption change, and this value does not exceed the unit energy consumption limit of the process chain. For example, if the unit energy consumption limit of a crushing to grinding process chain is a specific kilowatt-hour per ton, only the combination of corrections that ensures the corrected value does not exceed this limit is retained after verification. The verification process considers the synergistic effect between processes; when a conflict is detected, a coordination factor is introduced for adjustment. The coordination factor is calculated based on the material balance relationship between processes.

[0050] A global optimization decision scheme is generated by combining equipment-level control parameter corrections that simultaneously meet the target value for comprehensive energy consumption per ton of ore and the unit energy consumption limit value of the process chain. The scheme generation process is as follows: First, the verified correction combinations are comprehensively scored. The scoring criteria include three dimensions: energy consumption reduction magnitude, implementation cost coefficient, and stability index, with each dimension assigned a specific weight value. The energy consumption reduction magnitude is calculated based on the expected energy savings generated by the correction, the implementation cost coefficient is calculated based on the resource consumption required to adjust the equipment parameters, and the stability index is calculated based on the success rate of historical adjustment records. Second, a specific number of combinations with the highest comprehensive scores are selected as candidate schemes, and the number of candidate schemes is set to a specific value. Finally, the candidate schemes are checked for safety boundaries to ensure that the equipment operating frequency setting value is within the range of the minimum and maximum allowable frequency setting values ​​of the equipment, and that the equipment power threshold is within the range of a specific percentage lower limit and a specific percentage upper limit of the equipment's rated power. The global optimization decision scheme includes fields such as equipment identifier, control parameter type, correction value, and implementation time window. For example, the generated scheme specifies that the crusher with equipment identifier PJ-001-01 will adjust its operating frequency from the current value to the new setting value within a specific time period. A feasibility analysis report is attached to the output of the plan, which includes energy consumption prediction data and safety verification results.

[0051] The global optimization decision scheme is updated at specific time intervals, and a scheme regeneration is triggered when a change in operating conditions or the addition of a conflict node is detected. Before implementation, the scheme is simulated and verified. The implementation effect is predicted using a digital twin system, which builds a virtual model based on the current production status and simulates the operating state after parameter adjustments. The scheme is withdrawn when the predicted energy consumption reduction rate falls below a certain threshold percentage. All schemes are stored in the optimization decision database, and implementation results are fed back to the historical database for model optimization. Detailed operation logs are recorded during scheme execution, including fields such as scheme identifier, execution time, and actual energy savings.

[0052] This embodiment constructs an energy consumption transmission topology network linking process equipment to achieve closed-loop tracking of energy efficiency disturbances caused by energy-saving operations in a multi-process system. Specifically, based on dynamic monitoring of energy density gradient deviation in three-dimensional space, it overcomes the limitations of traditional single-point energy efficiency optimization and accurately locates self-excited oscillation conflict nodes caused by cross-process transmission. A historical data-driven energy consumption transmission coefficient regression model is introduced to establish a personalized transmission relationship quantification mechanism for different equipment-process combinations, addressing the lack of universality of empirical parameters in existing technologies. In the global optimization stage, a multi-objective constraint collaboration mechanism is applied to triple-couple and verify equipment-level parameter corrections with enterprise-level comprehensive energy consumption targets per ton of ore and energy consumption limits per unit of the process chain, ensuring the feasibility of the optimization scheme in both transmission path and spatial distribution dimensions. This overcomes the technical contradiction between local energy saving and global energy efficiency imbalance under complex mining conditions, achieving a leap from single-machine energy saving to system energy efficiency through a closed-loop logic of transmission topology modeling, gradient deviation analysis, regression coefficient calculation, and multi-constraint decision-making. Example

[0053] Figure 2 A schematic diagram of the knowledge graph-based mine energy efficiency optimization decision-making system of the present invention is given. The knowledge graph-based mine energy efficiency optimization decision-making system includes: The parameter acquisition module is used to acquire real-time operating parameters of equipment and enterprise-level energy efficiency indicators in the mining production system. The topology modeling module is used to build an industrial knowledge graph, which connects equipment-level entities, process-level entities, and enterprise-level entities through energy consumption transmission relationships to form a cross-level topology network. The vortex detection module is used for topology path analysis based on industrial knowledge graphs. It identifies conflict nodes by detecting energy flow vortex phenomena caused by energy-saving operations at the device level on the topology path. The gradient calculation module is used to monitor the spatial energy density imbalance caused by conflict nodes through the three-dimensional topological relationship of the industrial knowledge graph, and to calculate the energy distribution gradient deviation value. The conduction modeling module is used to calculate the energy conduction coefficient of conflict nodes based on the energy distribution gradient deviation value and a regression model of historical data. The collaborative optimization module is used to adjust equipment-level control parameters based on the energy consumption transmission coefficient and generate a global optimization decision scheme.

[0054] The calculations involved in the embodiments are all dimensionless numerical calculations, and the preset parameters and thresholds in the calculations are set by those skilled in the art according to the actual situation.

[0055] It should be noted that this invention can be deployed on the device itself to realize embedded applications, or it can run on a PC or other terminal with a user interface, thereby meeting various hardware environments and usage requirements.

[0056] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wireless or wired transmission; wired transmission methods include optical fiber, twisted pair, coaxial cable, etc.; wireless transmission includes infrared, microwave, etc. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center containing one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0057] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and modules described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0058] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of modules is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple modules or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or modules may be electrical, mechanical, or other forms.

[0059] The modules described as separate components may or may not be physically separate. The components shown as modules may or may not be physical modules; they may be located in one place or distributed across multiple network modules. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0060] In addition, the functional modules in the various embodiments of this application can be integrated into one processing module, or each module can exist physically separately, or two or more modules can be integrated into one module.

[0061] If the aforementioned functions are implemented as software functional modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a portion of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0062] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

[0063] In conclusion, the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A knowledge graph-based decision-making method for mine energy efficiency optimization, characterized in that, include: S1. Obtain real-time operating parameters of equipment and enterprise-level energy efficiency indicators in the mining production system; S2. Construct an industrial knowledge graph to connect equipment-level entities, process-level entities, and enterprise-level entities through energy consumption transmission relationships to form a cross-level topology network; S3. Based on the industrial knowledge graph, topology path analysis identifies conflict nodes by detecting the energy flow vortex phenomenon caused by equipment-level energy-saving operations on the topology path. The energy flow vortex phenomenon is manifested as the formation of a closed loop in the local energy consumption conduction path. S4. Monitor the spatial energy density imbalance caused by conflict nodes through the three-dimensional topological relationship of the industrial knowledge graph, and calculate the energy distribution gradient deviation value. S5. Based on the energy distribution gradient deviation, calculate the energy transmission coefficient of the conflict node using a historical data regression model; S6. Adjust the equipment-level control parameters based on the energy consumption transmission coefficient to generate a global optimization decision scheme.

2. The mine energy efficiency optimization decision-making method based on knowledge graphs according to claim 1, characterized in that, Obtain real-time operating parameters of equipment and enterprise-level energy efficiency indicators in the mining production system, including: Real-time operating parameters at the equipment level are collected from the equipment control system. These parameters include equipment power consumption data, equipment operating frequency data, and equipment temperature data. Enterprise-level energy efficiency indicators are periodically obtained from the enterprise resource management system. These indicators include the comprehensive energy consumption target per ton of ore and the energy consumption limit per unit of the process chain.

3. The mine energy efficiency optimization decision-making method based on knowledge graphs according to claim 2, characterized in that, Construct an industrial knowledge graph that connects equipment-level entities, process-level entities, and enterprise-level entities through energy consumption transmission relationships to form a cross-level topology network, including: A device-level entity is created based on the device-level real-time operating parameters. The device-level entity includes a device identifier, a device type attribute, and a real-time power consumption attribute. Based on the division of mining production processes, process-level entities are created. Each process-level entity includes a process identifier, a process type attribute, and an energy consumption baseline attribute. Create an enterprise-level entity and configure enterprise-level energy efficiency indicator attributes; Based on the physical connection and control logic of the equipment in the mining production system, establish the subordinate relationship between equipment-level entities and process-level entities; Based on the upstream and downstream sequence of mining production processes, establish the energy consumption transmission relationship between process-level entities; Enterprise-level entities and process-level entities are linked through global energy consumption constraints to form a cross-level topology network.

4. The mine energy efficiency optimization decision-making method based on knowledge graphs according to claim 3, characterized in that, Energy transmission relationships include the energy consumption impact chain of equipment operation on downstream processes.

5. The mine energy efficiency optimization decision-making method based on knowledge graphs according to claim 3, characterized in that, Based on industrial knowledge graph-based topology path analysis, conflict nodes are identified by detecting energy flow vortices caused by equipment-level energy-saving operations on the topology path, including: Locate the source device-level entity that implements device-level energy-saving operations on the topological path of the industrial knowledge graph; Track the propagation path of energy efficiency disturbances caused by energy-saving operations along the direction of energy consumption transmission, and record the energy consumption deviation values ​​of process-level entities in the propagation path; Detect whether there is a closed-loop transmission path in the topology that starts from the source device level entity and returns to the source device level entity through the process level entity; When the cumulative energy consumption deviation of the closed-loop transmission path exceeds the energy consumption baseline attribute of the process-level entity and forms a self-excited oscillation trend, all entity nodes included in the closed-loop transmission path are marked as conflict nodes.

6. The mine energy efficiency optimization decision-making method based on knowledge graphs according to claim 5, characterized in that, By monitoring the spatial energy density imbalance caused by conflict nodes through the three-dimensional topological relationships of an industrial knowledge graph, the deviation of the energy distribution gradient is calculated, including: Based on the three-dimensional topological relationship of industrial knowledge graph, the spatial coordinates of conflict nodes in the three-dimensional space of the mine are obtained; Centered on the conflict node, the measured energy density values ​​of adjacent process-level entities within a preset range are collected along the spatial transmission direction of the three-dimensional topological relationship. Calculate the three-dimensional spatial distance between the conflict node and each adjacent process-level entity based on the spatial location coordinates; Based on the measured values ​​of three-dimensional spatial distance and energy density, a spatial energy density distribution field of the local region where the conflict node is located is constructed. The magnitude of the gradient vector of the spatial energy density distribution field is calculated as the energy distribution gradient deviation value.

7. The mine energy efficiency optimization decision-making method based on knowledge graphs according to claim 6, characterized in that, Based on the energy distribution gradient deviation, the energy transmission coefficient of the conflict nodes is calculated using a regression model based on historical data, including: Obtain the historical energy distribution gradient deviation sequence of conflict nodes and the corresponding historical energy consumption conduction data; Based on the equipment type and process type attributes of the conflict nodes, a historical regression sample set of the same type of nodes is selected from the historical energy efficiency database. A multiple linear regression model was trained using historical energy distribution gradient deviation as the independent variable and historical energy consumption transmission data as the dependent variable. Input the current energy distribution gradient deviation into the trained multiple linear regression model, and output the energy transmission coefficient of the conflict node.

8. The mine energy efficiency optimization decision-making method based on knowledge graphs according to claim 7, characterized in that, Based on the energy consumption transmission coefficient, adjust the equipment-level control parameters to generate a global optimization decision scheme, including: The correction amount of equipment-level control parameters is calculated based on the energy consumption transmission coefficient; The correction amount is compared with the target value of comprehensive energy consumption per ton of mine in the enterprise-level energy efficiency index to screen the set of effective correction amounts that meet the target value of comprehensive energy consumption per ton of mine. Verify the feasibility of the effective correction set within the unit energy consumption limit value of the process chain based on the energy consumption baseline attributes of the process-level entities. The equipment-level control parameter corrections that simultaneously meet the comprehensive energy consumption target value per ton of ore and the unit energy consumption limit value of the process chain are combined to generate a global optimization decision scheme.

9. The mine energy efficiency optimization decision-making method based on knowledge graphs according to claim 8, characterized in that, Equipment-level control parameters include the equipment operating frequency setpoint and the equipment power threshold.

10. A knowledge graph-based mine energy efficiency optimization decision-making system, used to implement the knowledge graph-based mine energy efficiency optimization decision-making method according to any one of claims 1-9, characterized in that, include: The parameter acquisition module is used to acquire real-time operating parameters of equipment and enterprise-level energy efficiency indicators in the mining production system. The topology modeling module is used to build an industrial knowledge graph, which connects equipment-level entities, process-level entities, and enterprise-level entities through energy consumption transmission relationships to form a cross-level topology network. The vortex detection module is used for topology path analysis based on industrial knowledge graphs. It identifies conflict nodes by detecting energy flow vortex phenomena caused by energy-saving operations at the device level on the topology path. The gradient calculation module is used to monitor the spatial energy density imbalance caused by conflict nodes through the three-dimensional topological relationship of the industrial knowledge graph, and to calculate the energy distribution gradient deviation value. The conduction modeling module is used to calculate the energy conduction coefficient of conflict nodes based on the energy distribution gradient deviation value and a regression model of historical data. The collaborative optimization module is used to adjust equipment-level control parameters based on the energy consumption transmission coefficient and generate a global optimization decision scheme.

Images