A knowledge graph-based phosphate rock industry internet intelligent decision-making method and system

Abstract

The application relates to the technical field of decision management, and provides a phosphorus mine industrial internet intelligent decision method and system based on a knowledge graph. Production data in a phosphorus mine industry is collected through intelligent sensors, and decision information is generated by performing edge processing on the production data through an edge computing node; knowledge tuples are generated based on the production data and the decision information; entity devices and their relationship information are mapped into device nodes and edges of a Bayesian network, a knowledge state code and a hidden state at a set time are generated based on a confidence weight of the device nodes; the knowledge state and probability of the device nodes at a next time are predicted based on the hidden state at the set time; and a dynamically evolved knowledge graph is generated based on the hidden state at the set time and the knowledge state and probability at the next time, so that the phosphorus mine industrial production is decided through the knowledge graph. The dynamically evolved knowledge graph generated by the scheme can assist decision making, and the efficiency and accuracy of dynamic decision making in the phosphorus mine industrial production are improved.

Description

Technical Field

[0001] This application relates to the field of decision management technology, and more specifically, to an intelligent decision-making method and system for the phosphate mining industry based on knowledge graphs. Background Technology

[0002] Traditional industrial internet systems in mining typically employ a centralized data processing architecture, relying on a single data source or simple data integration methods, lacking a unified multi-source heterogeneous data processing framework. The separation of data acquisition and computation leads to insufficient real-time performance; knowledge graph construction largely depends on general-purpose NLP tools, resulting in errors in entity extraction due to domain terminology segmentation, and low accuracy in extracting state knowledge tuples. Decision-making mechanisms are primarily based on static rule engines, lacking dynamic optimization capabilities, and their weak edge data processing capabilities, relying on cloud computing, result in high latency. Therefore, existing technologies suffer from low accuracy in management decisions for the phosphate mining industry. Summary of the Invention

[0003] This application provides a knowledge graph-based intelligent decision-making method and system for the phosphate mining industry, which can at least partially solve the problem of low accuracy in the management decision-making process of the phosphate mining industry.

[0004] Other features and advantages of this application will become apparent from the following detailed description, or may be learned in part from practice of this application.

[0005] According to one aspect of this application, a knowledge graph-based intelligent decision-making method for the phosphate mining industry is provided, comprising: collecting production data in the phosphate mining industry through intelligent sensors; performing edge processing on the production data through edge computing nodes to generate decision information; generating knowledge tuples representing equipment states based on the production data and the decision information; mapping entity equipment and its relationship information to device nodes and edges in a Bayesian network; generating knowledge state encoding and hidden states at a set time based on the confidence weights of the device nodes and the knowledge tuples; predicting the knowledge state and probability of the device nodes at the next time based on the hidden states at the set time; and generating a dynamically evolving knowledge graph based on the hidden states at the set time, the knowledge state and probability at the next time, so as to make decisions on phosphate mining production through the knowledge graph.

[0006] In this application, based on the aforementioned scheme, the step of collecting production data in the phosphate mining industry through smart sensors and performing edge processing on the production data through edge computing nodes to generate decision information includes: collecting production data in the phosphate mining industry through smart sensors; pre-deploying edge computing nodes near the smart sensors and performing edge processing on the production data through the edge computing nodes to generate decision information.

[0007] In this application, based on the aforementioned scheme, the step of generating a knowledge tuple representing the equipment status based on the production data and the decision information includes: determining time series information and confidence weights based on the production data and the decision information; and generating a knowledge tuple representing the equipment status based on the time series information and the confidence weights.

[0008] In this application, based on the aforementioned scheme, the step of mapping entity devices and their relationship information to device nodes and edges of a Bayesian network, and generating knowledge state encoding and hidden state at a set time based on the confidence weights of the device nodes and the knowledge tuples, includes: generating device nodes of a Bayesian network based on the device information of the entity devices; generating edges of a Bayesian network based on the relationship information between the entity devices; and generating knowledge state encoding and hidden state at a set time based on the confidence weights of the device nodes and the knowledge tuples.

[0009] In this application, based on the aforementioned scheme, predicting the knowledge state and probability of the device node at the next time step based on the hidden state at the set time step includes: predicting the knowledge state of the device node at the next time step based on the hidden state at the set time step; and determining the probability that the device node is in the knowledge state at the next time step based on the knowledge state.

[0010] In this application, based on the aforementioned scheme, the step of generating a dynamically evolving knowledge graph based on the hidden state at the set time, the knowledge state at the next time, and the probability, and making decisions on phosphate mining industrial production through the knowledge graph, includes: determining the temporal evolution trajectory and confidence weight of equipment nodes based on the hidden state at the set time, the knowledge state at the next time, and the probability, and generating a dynamically evolving knowledge graph; and making decisions on phosphate mining industrial production through the knowledge graph.

[0011] In this application, based on the aforementioned scheme, the method further includes updating the knowledge graph based on a set time period.

[0012] According to one aspect of this application, a knowledge graph-based intelligent decision-making system for the phosphate mining industry is provided, comprising:

[0013] The acquisition module is used to collect production data in the phosphate mining industry through smart sensors, and to perform edge processing on the production data through edge computing nodes to generate decision information.

[0014] The knowledge module is used to generate knowledge tuples representing the equipment status based on the production data and the decision information;

[0015] The encoding module is used to map entity devices and their relationship information to device nodes and edges of a Bayesian network, and generate knowledge state encoding and hidden state at a set time based on the confidence weight of the device nodes and the knowledge tuples.

[0016] The probability module is used to predict the knowledge state and probability of the device node at the next time step based on the hidden state at the set time.

[0017] The decision-making module is used to generate a dynamically evolving knowledge graph based on the hidden state at the set time, the knowledge state at the next time, and the probability, so as to make decisions on phosphate mining production through the knowledge graph.

[0018] According to one aspect of this application, a computer-readable medium is provided having a computer program stored thereon, which, when executed by a processor, implements the knowledge graph-based intelligent decision-making method for the phosphate mining industry internet as described in the above embodiments.

[0019] According to one aspect of this application, an electronic device is provided, comprising: one or more processors; and a storage device for storing one or more programs, which, when executed by the one or more processors, cause the one or more processors to implement the knowledge graph-based intelligent decision-making method for the phosphate mining industry internet as described in the above embodiments.

[0020] According to one aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the knowledge graph-based intelligent decision-making method for the phosphate mining industry provided in the various optional implementations described above.

[0021] This application's technical solution collects production data in the phosphate mining industry using intelligent sensors, performs edge processing on the production data using edge computing nodes to generate decision information, generates knowledge tuples representing equipment states based on the production data and the decision information, maps physical equipment and its relationships to device nodes and edges in a Bayesian network, generates knowledge state encoding and hidden states at a set time based on the confidence weights of the device nodes and the knowledge tuples, predicts the knowledge state and probability of the device nodes at the next time based on the hidden states at the set time, and generates a dynamically evolving knowledge graph based on the hidden states at the set time, the knowledge state and probability at the next time, and uses the knowledge graph to make decisions about phosphate mining production. This solution constructs a complete intelligent decision-making chain for the phosphate mining industry. First, it collects production data and performs edge processing to generate decision information, ensuring real-time performance and accuracy. Then, it generates equipment state knowledge tuples to accurately characterize the equipment situation, maps them to a Bayesian network, and generates knowledge states to uncover potential information. Finally, it generates a dynamic knowledge graph to assist decision-making, allowing decision-makers to make decisions based on comprehensive dynamic information, thus improving the efficiency and accuracy of dynamic decision-making.

[0022] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0023] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0024] Figure 1 The flowchart illustrating a knowledge graph-based intelligent decision-making method for the phosphate mining industry is shown in one embodiment of this application.

[0025] Figure 2 The flowchart illustrating the generation of decision information is shown in one embodiment of this application.

[0026] Figure 3 The illustration shows a schematic diagram of a knowledge graph-based intelligent decision-making system for the phosphate mining industry internet, as shown in one embodiment of this application.

[0027] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown. Detailed Implementation

[0028] Exemplary embodiments will now be described more fully with reference to the accompanying drawings. However, these exemplary embodiments can be implemented in many forms and should not be construed as limited to the examples set forth herein; rather, these embodiments are provided to make this application more comprehensive and complete, and to fully convey the concept of the exemplary embodiments to those skilled in the art.

[0029] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this application. However, those skilled in the art will recognize that the technical solutions of this application can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this application.

[0030] The block diagrams shown in the accompanying drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0031] The flowcharts shown in the accompanying drawings are merely illustrative and do not necessarily include all content and operations / steps, nor do they necessarily have to be performed in the described order. For example, some operations / steps can be broken down, while others can be combined or partially combined; therefore, the actual execution order may change depending on the specific circumstances.

[0032] The implementation details of the technical solution of this application are described below:

[0033] Figure 1 A flowchart of a knowledge graph-based intelligent decision-making method for the phosphate mining industry internet, according to an embodiment of this application, is shown. (Refer to...) Figure 1 As shown, the knowledge graph-based intelligent decision-making method for the phosphate mining industry includes at least steps S110 to S150, which are detailed below:

[0034] S110 collects production data from the phosphate mining industry through smart sensors, and performs edge processing on the production data through edge computing nodes to generate decision information.

[0035] In the phosphate mining industry, the first step is to deploy smart sensors. Based on the characteristics of the phosphate mining process, various types of smart sensors are installed at key locations. For example, in the ore mining stage, vibration sensors are installed on key mechanical parts of the excavator to monitor the intensity and frequency of vibrations during operation. Vibration conditions reflect the excavator's operating status, and abnormal vibrations may indicate potential equipment malfunctions. In the ore transportation stage, pressure sensors are installed under the conveyor belts to monitor the material transport volume by sensing changes in the weight of the material on the belt. Simultaneously, temperature sensors are installed in the ore processing workshop to monitor the temperature of the processing equipment and prevent damage from overheating. After these sensors are deployed, data acquisition is activated. Through the sensor chips built into the smart sensors, a self-organizing network or sensor network is established between the smart sensors to continuously collect and process various industrial big data from the phosphate mining process. Specifically, this data covers information on multiple aspects, including machinery operation, material transportation, and equipment temperature.

[0036] like Figure 2 As shown, in one embodiment of this application, production data in the phosphate mining industry is collected by smart sensors, and the production data is processed by edge computing nodes to generate decision information, including:

[0037] S210 collects production data in the phosphate mining industry through smart sensors;

[0038] S220, edge computing nodes are pre-deployed near the smart sensors, and the production data is processed at the edge through the edge computing nodes to generate decision information.

[0039] After collecting data, smart sensors transmit this data accurately. They typically use wired or wireless communication methods. For sensors that are close to each other and require high data transmission stability, such as multiple sensors installed on the same device, wired transmission is used, transmitting data to nearby edge computing nodes via cables. This method ensures stable data transmission and low latency, ensuring that data is not lost or distorted during transmission. For sensors that are far away or whose installation locations are inconvenient for wiring, such as sensors located in remote corners of open-pit mines, wireless transmission is used. This can be achieved through Wi-Fi, Bluetooth, or specialized industrial wireless communication protocols, forming wireless sensor networks or wireless networks to send data to edge computing nodes. During transmission, to ensure data security and integrity, the sensors perform simple encoding and encryption to prevent data theft or tampering during transmission.

[0040] Edge computing nodes are pre-deployed near data sources or smart sensors, such as in the control room or server room near equipment at a phosphate mine production site. Once smart sensors transmit data to the edge computing nodes via wired or wireless means, the nodes quickly and accurately identify and receive data from different sensors. During data reception, the edge computing nodes perform preliminary verification, checking for correct format and completeness. If problems are found, such as incorrect data format or partial data loss, a retransmission request is sent to the corresponding sensor, requiring it to resend the data to ensure accuracy and reliability.

[0041] After receiving the data, the edge computing nodes preprocess the raw data to remove impurities and noise. For example, data collected by temperature sensors may contain outliers due to sensor errors or environmental interference. The edge computing nodes will remove these abnormal temperature data that exceed the normal range by setting reasonable threshold ranges. Simultaneously, for duplicate data collection, the device nodes will perform deduplication to avoid data redundancy. Furthermore, to improve the efficiency of subsequent processing, the edge computing nodes will also normalize the data, unifying data with different dimensions into the same range. For example, temperature and pressure data will be normalized to between 0 and 1, facilitating unified analysis and processing later.

[0042] The preprocessed data enters the core analysis phase. Edge computing nodes perform in-depth analysis of the data based on pre-defined analytical models and algorithms. For example, for mechanical vibration data, equipment nodes analyze vibration frequency, amplitude, and other characteristics, combined with historical data and empirical models, to determine whether the mechanical equipment is operating normally. If a significant deviation is found between the vibration characteristics and the normal state, the equipment nodes further analyze possible causes, such as wear and tear on equipment parts or loosening. Based on these analysis results, the edge computing nodes generate decision information. The decision information clearly indicates which equipment has problems, the severity of the problems, and recommended measures, such as immediate shutdown for inspection and maintenance, enhanced equipment monitoring, or simple adjustments and maintenance. This decision information is transmitted via the network to the management system or relevant management personnel's terminal devices in the phosphate mining industry, enabling them to make timely decisions and ensure the smooth operation of phosphate mining production.

[0043] The above process utilizes smart sensors to acquire data in real-time and accurately at each stage of phosphate mining production, such as equipment operating parameters and material status. This provides rich and detailed foundational information for subsequent analysis, ensuring a comprehensive reflection of the actual production situation. Deploying edge computing nodes near the smart sensors allows for real-time processing of the collected data. This approach avoids the latency and network burden associated with transmitting large amounts of raw data to a remote center for processing. It enables rapid generation of decision-making information based on data processing results, supporting real-time decision-making and improving the production system's response speed to unexpected situations.

[0044] S120, Based on the production data and the decision information, generate knowledge tuples representing the equipment status.

[0045] In one embodiment of this application, after obtaining production data and decision information, the production data is analyzed in depth to identify various key information related to the equipment, such as actual operating parameters like temperature, pressure, and speed during equipment operation. At the same time, combined with the target range and expected state set for equipment operation in the decision information, the equipment is taken as the core subject. The actual operating characteristic information extracted from the production data and the corresponding standard or target state information determined based on the decision information are organized and integrated according to a specific logical relationship to generate a knowledge tuple that can accurately and comprehensively represent the current actual state of the equipment and its association with the expected state.

[0046] In one embodiment of this application, a knowledge tuple representing the equipment state is generated based on the production data and the decision information, including:

[0047] Based on the production data and the decision information, determine the time series information and confidence weights;

[0048] Based on the time series information and confidence weights, a knowledge tuple representing the device state is generated.

[0049] In this embodiment, production data is based on the timestamps of data collection. Each sensor attaches a precise timestamp when collecting data, recording the specific moment the data was generated. These timestamps are extracted, and the production data is sorted chronologically to clarify the distribution of production data over time and determine its temporal sequence information. For decision information, since decisions are based on the analysis of production data within a certain time period, the temporal sequence of the decision information is associated with the temporal sequence of the production data for the corresponding analysis period. Based on the time range of the production data upon which the decision information is based, a corresponding time interval is determined for the decision information and integrated into the overall temporal sequence framework. For example, if a decision is based on the analysis results of production data from the past hour, then the temporal sequence of the decision information corresponds to the time range of that hour.

[0050] The extracted equipment status features are fused with previously determined time-series information and confidence weights. For each equipment status feature at a given time point, its temporal position in the production process is determined based on its corresponding time-series information; simultaneously, the reliability of the feature is quantified according to the confidence weight. For example, at a certain time point, the motor temperature feature of the crusher is 80℃. The time-series information corresponding to this time point indicates that this is a specific moment in the production process, and the confidence weight corresponding to this temperature data is 0.8, indicating that the reliability of this temperature data is relatively high.

[0051] For production data, the accuracy and reliability of the sensors themselves are considered. Different sensors will exhibit varying measurement errors and stability over long-term use. Based on indicators such as sensor calibration records and the accuracy of historical measurement data, a confidence weight is assigned to the data collected by each sensor, based on data quality. For example, a sensor that has undergone multiple calibrations and has historical measurement data errors will have a higher confidence weight for its collected data; conversely, a sensor that was not calibrated in a timely manner and has historical data errors will have a lower confidence weight for its collected data.

[0052] Based on the confidence weight of the sufficiency of the decision-making basis, the completeness and diversity of the production data on which the decision information is generated are analyzed. If a decision is based on a large amount of comprehensive production data (covering data from multiple devices and production processes), then the confidence weight of the decision information will be higher, because sufficient data support makes the decision more reliable. Conversely, if a decision is based only on a small amount of partial production data, its confidence weight will be lower. For example, a decision about equipment failure that is based only on data from the equipment temperature indicator will have a lower confidence weight than a decision based on data from multiple indicators such as temperature, vibration, and pressure.

[0053] Comprehensive weight calculation: The confidence weight based on data quality and the confidence weight based on the sufficiency of decision-making evidence are calculated together. A weighted average method can be used, assigning different weight coefficients to these two weights according to the actual situation, and then calculating the final confidence weight. For example, assuming the weight coefficient based on data quality is 0.6 and the weight coefficient based on the sufficiency of decision-making evidence is 0.4, then the final confidence weight = 0.6 × weight based on data quality + 0.4 × weight based on the sufficiency of decision-making evidence.

[0054] The extracted equipment status features are fused with previously determined time-series information and confidence weights. For each equipment status feature at a given time point, its temporal position in the production process is determined based on its corresponding time-series information; simultaneously, the reliability of the feature is quantified according to the confidence weight. For example, at a certain time point, the motor temperature feature of the crusher is 80℃. The time-series information corresponding to this time point indicates that this is a specific moment in the production process, and the confidence weight corresponding to this temperature data is 0.8, indicating that the reliability of this temperature data is relatively high.

[0055] Using equipment as the basic unit, the equipment state features, which integrate time-series information and confidence weights, are organized into knowledge tuples. The knowledge tuples take the following form: equipment identifier, time, state feature 1, confidence weight of state feature 1, state feature 2, confidence weight of state feature 2. For example, for the aforementioned crusher, the generated knowledge tuple can be represented as (crusher 001, 2024-01-01 10:00:00, motor temperature 80℃, 0.8, crushing chamber pressure 0.5MPa, 0.7, vibration amplitude 2mm / s, 0.9, operating mode: normal operation). Such knowledge tuples can comprehensively and accurately represent the equipment's state information at a specific point in time and the reliability of this information.

[0056] For example, in this embodiment, during the generation of state knowledge tuples, each tuple contains time-series information t and confidence weight c. For instance, a device state knowledge tuple can be represented as: (Device A, Temperature Abnormal, Above Threshold, 2025) -08 -2514:30, 0.85, Safety Rule 1); where t = 2025 -08 -2514:30, c=0.85 represents the confidence level of this knowledge.

[0057] Knowledge representation theory organizes equipment status information in a structured manner to facilitate storage, processing, and analysis by computer systems. By incorporating temporal information and confidence weights into knowledge tuples, these tuples not only contain static information about the equipment status but also dynamic information regarding the time and reliability of this information, thus providing a more comprehensive reflection of the actual equipment status. Generating knowledge tuples representing equipment status can provide a unified and standardized knowledge representation for phosphate mining production management systems. These knowledge tuples can be stored in a knowledge base for easy querying, retrieval, and analysis. Simultaneously, the temporal information and confidence weights in the knowledge tuples provide richer evidence for system decision-making, enabling the system to make more scientific and rational decisions based on the dynamic changes in equipment status and the reliability of information, thereby improving the efficiency and safety of phosphate mining production.

[0058] The above process, by combining production data and decision-making information, can extract time-series information on how equipment status changes over time, while simultaneously assessing the reliability of this information, i.e., the confidence weight. Time-series information helps understand the development trend of equipment status, while the confidence weight provides a basis for judging the authenticity of the equipment status, making the grasp of equipment status more accurate. Knowledge tuples generated based on the time-series information and confidence weights represent equipment status in a structured form, integrating multiple key elements of equipment status to facilitate subsequent knowledge reasoning and analysis, laying the foundation for building a knowledge graph.

[0059] S130, map the entity device and its relationship information to device nodes and edges of a Bayesian network, and generate the knowledge state encoding and hidden state at a set time based on the confidence weight of the device node and the knowledge tuple.

[0060] In one embodiment of this application, each physical device is transformed into a device node in a Bayesian network. Edges are constructed based on the actual connections, collaborations, and influence relationships between devices. For example, if device A provides raw materials to device B, a directed edge is drawn between the device nodes representing these two devices, thus fully representing the device relationship network. Next, based on factors such as past operating records, performance evaluations, and expert experience, a confidence weight is determined for each device node. This weight reflects the reliability of the device's state information. Finally, at a set time, by combining the confidence weights of all device nodes and their relationships, a knowledge state code summarizing the current device knowledge state is generated through specific logical reasoning and calculation methods. Simultaneously, hidden states that are not directly observed but can be inferred from the relationships between device nodes are uncovered.

[0061] In one embodiment of this application, entity devices and their relationship information are mapped to device nodes and edges in a Bayesian network. Based on the confidence weights of the device nodes and the knowledge tuples, a knowledge state encoding and hidden state at a set time are generated, including:

[0062] The device nodes of the Bayesian network are generated based on the device information of the physical devices, and the edges of the Bayesian network are generated based on the relationship information between the physical devices.

[0063] Based on the confidence weight of the device node and the knowledge tuple, a knowledge state code and hidden state are generated at a set time.

[0064] In this embodiment, relevant information about physical equipment is collected from the production management system of the phosphate mining industry. This information covers equipment type (such as crushers, conveyors, ball mills, etc.), basic equipment parameters (such as crusher power, conveyor bandwidth, ball mill speed, etc.), equipment operating status (normal operation, fault shutdown, maintenance, etc.), and historical fault records. For each physical piece of equipment, a corresponding Bayesian network device node is created. Taking the crusher as an example, it is included as a device node in the Bayesian network. Each device node is assigned a unique identifier to facilitate subsequent identification and operation within the network. Simultaneously, the collected equipment information is stored as attributes of the device node. For example, the attributes of the crusher device node may include equipment type: crusher, power: 200kW, operating status: normal operation, historical fault count: 3, etc.

[0065] A thorough analysis of the relationships between the physical equipment is conducted. These relationships include physical connections (e.g., the connection between a conveyor and a crusher via a material transport pipeline), functional dependencies (e.g., the operation of a ball mill depends on the normal feeding of a feeder), and control relationships (e.g., the start-stop control of various devices by a central control system). Based on the analyzed relationships, edges are created between the corresponding Bayesian network device nodes. If a physical connection exists between a conveyor and a crusher, a directed edge is drawn between the device nodes representing the conveyor and the crusher, with the direction of the edge indicating the direction of material flow or the direction of control signal transmission. In this way, the various device nodes are connected according to their relationships in actual production, forming a complete Bayesian network structure.

[0066] A set of knowledge state coding rules is established based on the confidence weights of device nodes. For example, the confidence weights are divided into several intervals, each corresponding to a specific code. Assume that a confidence weight between 0 and 0.3 is encoded as "00", between 0.3 and 0.6 as "01", between 0.6 and 0.9 as "10", and between 0.9 and 1 as "11". For each device node, a corresponding knowledge state code is generated according to its calculated confidence weight and the coding rules. For example, if a device node has a confidence weight of 0.85, its knowledge state code is "10".

[0067] Considering the mutual influence between device nodes and knowledge state encoding, a hidden state model is constructed. The hidden state can be understood as a comprehensive state of a device at a given time, reflecting not only the device's current state but also the influence of the states of surrounding devices. For example, in a production process consisting of multiple devices, the hidden state of a particular device may be influenced by the states of its upstream and downstream devices. Based on the knowledge state encoding and the edges (representations of dependencies) between device nodes, Bayesian inference is used to calculate the hidden state at a given time.

[0068] Specifically, based on the information in the device status knowledge tuple, combined with historical data and expert experience, the confidence weight of each device node is determined. A binary encoding method is used to encode the device status. For example, if there are three device statuses (good, fair, faulty), they are represented by 2-bit binary numbers, with the encoding rule: good - 00, fair - 01, faulty - 10. Based on the conditional probability distribution of the Bayesian network, the confidence weights of the device nodes, and the aforementioned knowledge tuple, the hidden state H at a given time is calculated. Let the set of parent device nodes in the Bayesian network be... The set of sub-device nodes is Then the device node X The hidden state is calculated as follows:

[0069]

[0070] in, Represents each device node Confidence weights For the parent device node Device node in status X The conditional probability of being in a hidden state. For device nodes X Sub-device node in status The conditional probability of being in a hidden state. , These represent the status of the parent device node and the status of the child device node, respectively.

[0071] For example, by using the known knowledge state encoding and combining it with the conditional probability distribution among device nodes in the Bayesian network, the probability of each device node being in different hidden states at a given time is calculated. Then, the hidden state with the highest probability is selected as the hidden state of that device node at the given time. For instance, assuming that the knowledge state encoding of a device node is 10, and Bayesian inference calculates that its probability of being in the normal operation hidden state is 0.7, its probability of being in the "minor fault" hidden state is 0.2, and its probability of being in the severe fault hidden state is 0.1, then the hidden state of this device node at the given time is normal operation.

[0072] Based on the quantified representation of the confidence weights of equipment nodes, continuous confidence weights are transformed into discrete codes for easier computer storage and processing. Utilizing the probabilistic reasoning capabilities of Bayesian networks, combined with the dependencies between equipment nodes and knowledge state encoding, the most likely comprehensive state of the equipment at a given time is inferred. After generating the knowledge state encoding and hidden states at the given time, richer and more accurate equipment status information can be provided to the phosphate mine industrial production management system. The knowledge state encoding concisely represents the confidence status of equipment nodes, facilitating rapid querying and comparison; the hidden states comprehensively consider the mutual influences between equipment, more accurately reflecting the true state of the equipment in the actual production environment. This information can provide strong support for production scheduling, fault early warning, and equipment maintenance, improving the stability and reliability of production.

[0073] S140, based on the hidden state at the set time, predict the knowledge state and probability of the device node at the next time.

[0074] In one embodiment of this application, after obtaining the hidden states at a set time, the potential trends and related information of device operation implied by these hidden states are analyzed in depth. By reviewing the evolution of device node states after the appearance of similar hidden states in historical data, and combining the real-time environmental factors and external interference conditions of the current device operation, a logical judgment method based on experience accumulation and pattern recognition is used to make multi-dimensional inferences about the knowledge state that each device node may present in the next time step. Furthermore, based on factors such as the frequency and stability of different state transitions in history, the probability of each inferred knowledge state occurring is comprehensively evaluated, thereby generating the knowledge state of the device node in the next time step and the corresponding probability prediction result.

[0075] In one embodiment of this application, predicting the knowledge state and probability of the device node at the next time step based on the hidden state at the set time step includes:

[0076] Based on the hidden state at the set time, predict the knowledge state of the device node at the next time.

[0077] Based on the knowledge state, determine the probability that the device node will be in the knowledge state at the next moment.

[0078] In this embodiment of the application, the knowledge state of the device node at the next time step is predicted based on the hidden state at the set time. for:

[0079]

[0080] in, Represents device node i exist t The set of parent device nodes at any given moment is defined by the causal relationships of the mining system, such as the temperature of device A affecting the vibration of device B; Represents device node i In the next moment The state; Indicates at the set time t The hidden state, W and b These represent the weight matrix and bias vector obtained through training with historical data, respectively. This indicates activation.

[0081] In this embodiment of the application, the knowledge state of the device node at the next moment is used as a basis. Determine the current moment t knowledge graph Under the condition, the next time slice t+Δt knowledge graph The probability of:

[0082]

[0083] in, i and n These represent the device node identifier and the total number of device nodes, respectively. Indicates a time interval.

[0084] Knowing the hidden state at the previous time step, the probability of each possible knowledge state occurring at the next time step can be calculated based on the probabilistic relationship between the hidden state and the knowledge state learned in the model. Determining the probability of a device node being in a certain knowledge state at the next time step provides a more accurate basis for decision-making. For example, if the probability of a certain knowledge state occurring is high, it indicates that the system is very likely to enter that state, and relevant personnel can formulate countermeasures in advance; if the probabilities of multiple knowledge states are relatively close, it indicates that there is significant uncertainty in the system state, requiring further monitoring and analysis to optimize production management and resource allocation.

[0085] The above process, utilizing the hidden states at a given time, analyzes the evolution patterns and potential influencing factors of equipment states to predict the knowledge state that a device node might be in at the next time step. This helps to understand the future operating conditions of the equipment in advance, providing a basis for adjusting production plans and conducting preventative maintenance. The probability of the device node being in the predicted knowledge state at the next time step is determined, and the reliability of the prediction results is quantitatively assessed. Probabilistic information allows decision-makers to clearly understand the likelihood of different predicted outcomes, thus enabling them to more rationally weigh various situations and reduce decision-making risks.

[0086] S150, based on the hidden state at the set time, the knowledge state and probability at the next time, a dynamically evolving knowledge graph is generated to make decisions on phosphate mining production through the knowledge graph.

[0087] In one embodiment of this application, after obtaining the hidden state at a set time and the knowledge state and probability at the next time, the existing static knowledge graph is used as a framework to accurately integrate the potential operating characteristics and associations of the equipment reflected by the hidden state at the set time into the attributes of the corresponding equipment nodes and edges in the graph, giving the graph an initial dynamic characteristic. Next, based on the knowledge state and its probability at the next time, the states of the equipment nodes and edges in the graph are predictively updated. For example, if a certain equipment node is highly likely to be in a fault warning state at the next time, this state and its related probability are marked on the equipment node in the graph. Finally, through the dynamic changes of equipment nodes and edges in the knowledge graph, the evolution trend and mutual influence of equipment states in each stage of phosphate mining production are clearly displayed. Decision-makers, based on the comprehensive information presented by the graph, can judge potential risks and optimization points in the production process, and thus make reasonable decisions to guide phosphate mining production.

[0088] In one embodiment of this application, a dynamically evolving knowledge graph is generated based on the hidden state at the set time, the knowledge state at the next time, and the probability, in order to make decisions on phosphate mining industrial production through the knowledge graph, including:

[0089] Based on the hidden state at the set time, the knowledge state and probability at the next time, the temporal evolution trajectory and confidence weight of the device node are determined, and a dynamically evolving knowledge graph is generated.

[0090] Decisions are made in phosphate mining production using the knowledge graph.

[0091] In this embodiment, based on the hidden state at the set time, the knowledge state at the next time, and the probability, the temporal evolution trajectory and confidence weight of the device node are determined, generating a dynamically evolving knowledge graph. The hidden state of each device node at the set time, the predicted knowledge state at the next time, and the corresponding probability are collected. This data is arranged chronologically, and an initial time-series dataset is constructed for each device node. For example, for a mining equipment node in phosphate mine production, its hidden state at different set times is recorded, such as normal operation, low-load operation, etc., the predicted knowledge state at the next time, such as normal operation with improved efficiency, minor fault, etc., and the corresponding probability value.

[0092] Temporal analysis algorithms, such as trajectory generation methods based on Markov chains, are employed. Starting from an initial time point, the possible state change paths of the device node at multiple future time points are gradually deduced based on the current hidden state and the predicted knowledge state and probability of the next time point. Specifically, based on the current state and transition probability, the next possible state is determined, and this new state is then used as the current state to continue deducing the state for the next time point, and so on, forming a complete temporal evolution trajectory. For example, if the hidden state of the device node at the current time point is "normal operation," and the probability of the predicted knowledge state "normal operation with improved efficiency" at the next time point is 0.7, while the probability of "minor fault" is 0.3, then "normal operation with improved efficiency" is preferentially selected as the state for the next time point according to the probability, and the state for subsequent time points is deduced to generate the trajectory.

[0093] Construct a probability-based confidence weight calculation model. Consider factors such as the probability of a knowledge state and the accuracy and stability of that state's occurrence in historical data. For example, a weighted average method can be used, multiplying the probability of a knowledge state by its historical accuracy, and adding a stability coefficient to reflect the degree of change of the state under different environments, thus obtaining the confidence weight of each knowledge state in the time-series evolution trajectory. Based on the calculated confidence weights, assign weights to each state in the time-series evolution trajectory. Simultaneously, adjust the weights appropriately based on actual production conditions and expert experience. For example, if a knowledge state has a high probability, but its occurrence in actual production is often accompanied by some unconsidered risk factors, then its confidence weight can be appropriately reduced.

[0094] The knowledge graph structure is designed, with each equipment node in phosphate mining production (such as mining equipment, transport vehicles, and processing stages) as vertices, and the relationships between these nodes (such as material flow and information transmission) as edges. The graph is initialized by using the hidden states of each equipment node at a given time as the initial attributes of the vertices. A dynamic update mechanism for the knowledge graph is established based on the determined temporal evolution trajectories of the equipment nodes and their confidence weights. At each time step, the state attributes of the vertices in the graph are updated according to the trajectory and weights. For example, when a equipment node enters the next knowledge state according to the trajectory, the state attribute corresponding to that vertex in the graph is updated to the new knowledge state, and the relevant attribute values ​​of the vertex (such as reliability and efficiency) are adjusted according to the confidence weights. Simultaneously, based on the changes in the state of the equipment nodes, the relationships represented by the edges (such as material flow speed and information transmission frequency) are updated, realizing the dynamic evolution of the knowledge graph.

[0095] This system can clearly display the status change trends of equipment nodes over a future period, helping production personnel understand potential equipment and process conditions in advance, and providing a basis for production planning and adjustments. For example, the trajectory can predict when mining equipment may need maintenance, avoiding production interruptions due to equipment failure. By comprehensively considering multiple factors to determine confidence weights, the reliability of each knowledge state can be more accurately assessed, providing a more reliable reference for decision-making. For example, when faced with multiple possible decision options, the most likely successful option can be selected based on the confidence weight. The dynamically evolving knowledge graph can reflect the status changes of the phosphate mining industry production system in real time, enabling production personnel to intuitively understand the operation of the entire production process, promptly identify potential problems and risks, and improve the efficiency and accuracy of production management.

[0096] In this embodiment, a data collection task is initiated at the end of each set time period. This process relies on various smart sensors pre-deployed in the phosphate mining industry production environment. For example, in the ore mining area, there are sensors specifically designed to monitor the operating parameters of mining equipment, such as sensors monitoring excavator engine speed and hydraulic system pressure; in the ore transportation process, there are sensors tracking conveyor belt speed and transport volume; and in the ore dressing workshop, there are sensors detecting parameters such as solution concentration and temperature. These sensors continuously collect real-time data during the production process at a set frequency and transmit the collected data to the data aggregation device node via wired or wireless communication networks. The data aggregation device node performs preliminary processing and caching of the received data to ensure data integrity and sequence, providing a reliable data source for subsequent processing.

[0097] After data is aggregated to the device nodes, it enters the preprocessing stage. The purpose of preprocessing is to clean and transform the raw data to improve data quality. Data cleaning primarily targets noisy data and outliers. For example, sensors may be subject to external interference during data acquisition, generating data that significantly deviates from the normal range. Pre-defined rules are used to identify these outliers, such as setting a reasonable numerical range. If data collected by a sensor exceeds this range, it is marked as an anomaly and removed or corrected. Data transformation involves unifying data from different formats and units into formats and units suitable for subsequent analysis. For example, temperature data collected by different sensors is uniformly converted to degrees Celsius, and length data is uniformly converted to meters, ensuring data consistency and comparability.

[0098] The preprocessed data is then transmitted to the knowledge extraction module. The main task of this module is to extract entities, relationships, and attributes related to the knowledge graph from the data. Taking device entities as an example, basic information such as the name, model, and location of various devices is identified from the data as entity attributes. Relationships between devices, such as the output of device A being the input of device B, are extracted. During the extraction process, matching and identification are performed according to predefined rules and patterns. For example, by analyzing the time series and logical relationships in the data, the sequential operation order and causal relationships between devices are determined, thereby accurately extracting knowledge.

[0099] The extracted knowledge may contain issues such as duplication and inconsistency; the knowledge fusion stage aims to address these problems. It involves comparing and merging the extracted entities and relationships. For example, for the same device, data from different sensors may generate different entity descriptions. By analyzing the entity's attribute information, it's determined whether they refer to the same device. If so, these descriptions are merged to form a complete and accurate entity representation. Similar fusion processing is performed on relationships to ensure consistency and accuracy within the knowledge graph. Simultaneously, the newly extracted knowledge is validated and adjusted based on existing knowledge graph structures and rules to ensure it conforms to the overall knowledge system.

[0100] After knowledge fusion is complete, the actual updating of the knowledge graph begins. Based on the newly extracted and fused knowledge, the existing knowledge graph is modified and expanded. For newly added entities or relationships, new device nodes and edges are added to the corresponding locations in the knowledge graph. For example, when a new mineral processing device is introduced, a device node representing that device is created in the graph, and corresponding edges are added based on its relationships with other devices. If the attributes of existing entities or relationships change, the attribute information of the corresponding device nodes or edges is updated. For example, if the operating efficiency attribute of a device changes, the new efficiency value is updated to the corresponding attribute of the device node. During the update process, the structural integrity and logical consistency of the knowledge graph are ensured, avoiding isolated device nodes or unreasonable edge connections.

[0101] After the knowledge graph is updated, it needs to be validated and optimized. The updated knowledge graph is checked using preset validation rules, such as verifying whether the connections between device nodes and edges conform to business logic and whether attribute values ​​are within reasonable ranges. If validation fails, previous steps are reviewed to identify and correct the problem. Simultaneously, the rules and algorithms for knowledge extraction, fusion, and updating are optimized based on actual usage and feedback. For example, if inaccurate extraction of certain types of knowledge is found, the extraction rules are adjusted; if frequent erroneous merging occurs during knowledge fusion, the fusion algorithm is optimized. Through continuous validation and optimization, the quality and usability of the knowledge graph are ensured, enabling it to better support production decisions in the phosphate mining industry.

[0102] This application's technical solution collects production data in the phosphate mining industry using intelligent sensors, performs edge processing on the production data using edge computing nodes to generate decision information, generates knowledge tuples representing equipment states based on the production data and the decision information, maps physical equipment and its relationships to device nodes and edges in a Bayesian network, generates knowledge state encoding and hidden states at a set time based on the confidence weights of the device nodes and the knowledge tuples, predicts the knowledge state and probability of the device nodes at the next time based on the hidden states at the set time, and generates a dynamically evolving knowledge graph based on the hidden states at the set time, the knowledge state and probability at the next time, and uses the knowledge graph to make decisions about phosphate mining production. This solution constructs a complete intelligent decision-making chain for the phosphate mining industry. First, it collects production data and performs edge processing to generate decision information, ensuring real-time performance and accuracy. Then, it generates equipment state knowledge tuples to accurately characterize the equipment situation, maps them to a Bayesian network, and generates knowledge states to uncover potential information. Finally, it generates a dynamic knowledge graph to assist decision-making, allowing decision-makers to make decisions based on comprehensive dynamic information, thus improving the efficiency and accuracy of dynamic decision-making.

[0103] The following describes embodiments of the knowledge graph-based intelligent decision-making system for the phosphate mining industry, which can be used to execute the knowledge graph-based intelligent decision-making method for the phosphate mining industry described in the above embodiments of this application. It is understood that the knowledge graph-based intelligent decision-making system for the phosphate mining industry can be a computer program (including program code) running on a computer device; for example, the knowledge graph-based intelligent decision-making system for the phosphate mining industry is an application software. This knowledge graph-based intelligent decision-making system for the phosphate mining industry can be used to execute the corresponding steps in the methods provided in the embodiments of this application. For details not disclosed in the embodiments of the knowledge graph-based intelligent decision-making system for the phosphate mining industry of this application, please refer to the embodiments of the knowledge graph-based intelligent decision-making method for the phosphate mining industry described above.

[0104] Figure 3 A block diagram of a knowledge graph-based intelligent decision-making system for the phosphate mining industry internet, according to an embodiment of this application, is shown.

[0105] Reference Figure 3 As shown, a knowledge graph-based intelligent decision-making system for the phosphate mining industry according to an embodiment of this application includes:

[0106] The acquisition module 310 is used to collect production data in the phosphate mining industry through intelligent sensors, and to perform edge processing on the production data through edge computing nodes to generate decision information.

[0107] Knowledge module 320 is used to generate knowledge tuples representing equipment status based on the production data and the decision information;

[0108] The encoding module 330 is used to map entity devices and their relationship information to device nodes and edges of a Bayesian network, and generate knowledge state encoding and hidden state at a set time based on the confidence weight of the device nodes and the knowledge tuple.

[0109] Probability module 340 is used to predict the knowledge state and probability of the device node at the next time step based on the hidden state at the set time.

[0110] The decision module 350 is used to generate a dynamically evolving knowledge graph based on the hidden state at the set time, the knowledge state at the next time, and the probability, so as to make decisions on phosphate mining production through the knowledge graph.

[0111] In this application, based on the aforementioned scheme, the step of collecting production data in the phosphate mining industry through smart sensors and performing edge processing on the production data through edge computing nodes to generate decision information includes: collecting production data in the phosphate mining industry through smart sensors; pre-deploying edge computing nodes near the smart sensors and performing edge processing on the production data through the edge computing nodes to generate decision information.

[0112] In this application, based on the aforementioned scheme, the step of generating a knowledge tuple representing the equipment status based on the production data and the decision information includes: determining time series information and confidence weights based on the production data and the decision information; and generating a knowledge tuple representing the equipment status based on the time series information and the confidence weights.

[0113] In this application, based on the aforementioned scheme, the step of mapping entity devices and their relationship information to device nodes and edges of a Bayesian network, and generating knowledge state encoding and hidden state at a set time based on the confidence weights of the device nodes and the knowledge tuples, includes: generating device nodes of a Bayesian network based on the device information of the entity devices; generating edges of a Bayesian network based on the relationship information between the entity devices; and generating knowledge state encoding and hidden state at a set time based on the confidence weights of the device nodes and the knowledge tuples.

[0114] In this application, based on the aforementioned scheme, predicting the knowledge state and probability of the device node at the next time step based on the hidden state at the set time step includes: predicting the knowledge state of the device node at the next time step based on the hidden state at the set time step; and determining the probability that the device node is in the knowledge state at the next time step based on the knowledge state.

[0115] In this application, based on the aforementioned scheme, the step of generating a dynamically evolving knowledge graph based on the hidden state at the set time, the knowledge state at the next time, and the probability, and making decisions on phosphate mining industrial production through the knowledge graph, includes: determining the temporal evolution trajectory and confidence weight of equipment nodes based on the hidden state at the set time, the knowledge state at the next time, and the probability, and generating a dynamically evolving knowledge graph; and making decisions on phosphate mining industrial production through the knowledge graph.

[0116] In this application, based on the aforementioned scheme, the method further includes updating the knowledge graph based on a set time period.

[0117] This application's technical solution collects production data in the phosphate mining industry using intelligent sensors, performs edge processing on the production data using edge computing nodes to generate decision information, generates knowledge tuples representing equipment states based on the production data and the decision information, maps physical equipment and its relationships to device nodes and edges in a Bayesian network, generates knowledge state encoding and hidden states at a set time based on the confidence weights of the device nodes and the knowledge tuples, predicts the knowledge state and probability of the device nodes at the next time based on the hidden states at the set time, and generates a dynamically evolving knowledge graph based on the hidden states at the set time, the knowledge state and probability at the next time, and uses the knowledge graph to make decisions about phosphate mining production. This solution constructs a complete intelligent decision-making chain for the phosphate mining industry. First, it collects production data and performs edge processing to generate decision information, ensuring real-time performance and accuracy. Then, it generates equipment state knowledge tuples to accurately characterize the equipment situation, maps them to a Bayesian network, and generates knowledge states to uncover potential information. Finally, it generates a dynamic knowledge graph to assist decision-making, allowing decision-makers to make decisions based on comprehensive dynamic information, thus improving the efficiency and accuracy of dynamic decision-making.

[0118] Figure 4 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown.

[0119] It should be noted that the computer system of the electronic device in this embodiment is only an example and should not impose any limitations on the function and scope of use of the embodiments of this application.

[0120] In this embodiment, the computer system includes a central processing unit 401, which can perform various appropriate actions and processes based on a program stored in the read-only memory 402 or a program loaded from the storage section 408 into the random access memory 403, such as executing the knowledge graph-based intelligent decision-making method for the phosphate mining industry described in the above embodiment. The random access memory 403 also stores various programs and data required for system operation. The central processing unit 401, the read-only memory 402, and the random access memory 403 are interconnected via a bus 404. An input / output interface 405 is also connected to the bus 404.

[0121] The following components are connected to the input / output interface 405: an input section 406 including a keyboard, mouse, etc.; an output section 407 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 408 including a hard disk, etc.; and a communication section 409 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 409 performs communication processing via a network such as the Internet. A drive 410 is also connected to the input / output interface 405 as needed. A removable medium 411, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., is installed on the drive 410 as needed so that computer programs read from it can be installed into the storage section 408 as needed.

[0122] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program including a computer program for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 409, and / or installed from removable medium 411. When the computer program is executed by central processing unit 401, it performs various functions defined in the system of this application.

[0123] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying a computer-readable computer program. The transmitted data signal can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The computer program contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0124] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0125] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0126] According to one aspect of this application, a computer program product or computer program is provided, comprising computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the methods provided in the various alternative implementations described above.

[0127] In another aspect, this application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to implement the knowledge graph-based intelligent decision-making method for the phosphate mining industry internet described in the above embodiments.

[0128] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0129] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the methods according to the embodiments of this application.

[0130] Other embodiments of this application will readily occur to those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein.

[0131] It should be understood that this application is not limited to the precise structure described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A knowledge graph-based intelligent decision-making method for the phosphate mining industry, characterized in that, include: Production data from the phosphate mining industry is collected by smart sensors, and the production data is processed at the edge through edge computing nodes to generate decision information. Based on the production data and the decision information, knowledge tuples representing the equipment status are generated; The entity devices and their relationship information are mapped to device nodes and edges in a Bayesian network. Based on the confidence weights of the device nodes and the knowledge tuples, the knowledge state encoding and hidden state at a set time are generated. Based on the hidden state at the set time, predict the knowledge state and probability of the device node at the next time. Based on the hidden state at the set time, the knowledge state and probability at the next time, a dynamically evolving knowledge graph is generated to make decisions on phosphate mining production. Specifically, based on the production data and the decision information, a knowledge tuple representing the equipment status is generated, including: Based on the production data and the decision information, time series information and confidence weights are determined; specifically, the production data is sorted in chronological order to determine the time series information of the production data; the confidence weights for data quality and the confidence weights based on the sufficiency of decision-making basis are comprehensively calculated to obtain the final confidence weights. Based on the time series information and confidence weights, knowledge tuples representing equipment status are generated. Specifically, for each equipment status feature extracted at a given time point, the time position of the equipment status feature in the production process is determined based on its corresponding time series information. The reliability of the equipment status features is quantified based on the confidence weights. Using the equipment as the basic unit, the equipment status features that integrate time series information and confidence weights are organized into knowledge tuples.

2. The knowledge graph-based intelligent decision-making method for the phosphate mining industry according to claim 1, characterized in that, Production data from the phosphate mining industry is collected through smart sensors, and edge computing nodes are used to process this production data to generate decision information, including: Production data in the phosphate mining industry is collected using smart sensors; Edge computing nodes are pre-deployed near smart sensors, and the production data is processed at the edge through these edge computing nodes to generate decision information.

3. The knowledge graph-based intelligent decision-making method for the phosphate mining industry according to claim 1, characterized in that, The entity devices and their relationship information are mapped to device nodes and edges in a Bayesian network. Based on the confidence weights of the device nodes and the knowledge tuples, a knowledge state encoding and hidden state at a set time are generated, including: The device nodes of the Bayesian network are generated based on the device information of the physical devices, and the edges of the Bayesian network are generated based on the relationship information between the physical devices. Based on the confidence weight of the device node and the knowledge tuple, a knowledge state code and hidden state are generated at a set time.

4. The knowledge graph-based intelligent decision-making method for the phosphate mining industry according to claim 1, characterized in that, Based on the hidden state at the set time, predict the knowledge state and probability of the device node at the next time step, including: Based on the hidden state at the set time, predict the knowledge state of the device node at the next time. Based on the knowledge state, determine the probability that the device node will be in the knowledge state at the next moment.

5. The knowledge graph-based intelligent decision-making method for the phosphate mining industry according to claim 1, characterized in that, Based on the hidden state at the set time, the knowledge state at the next time, and the probability, a dynamically evolving knowledge graph is generated to make decisions on phosphate mining production, including: Based on the hidden state at the set time, the knowledge state and probability at the next time, the temporal evolution trajectory and confidence weight of the device node are determined, and a dynamically evolving knowledge graph is generated. Decisions are made in phosphate mining production using the knowledge graph.

6. The knowledge graph-based intelligent decision-making method for the phosphate mining industry according to claim 1, characterized in that, Also includes: The knowledge graph is updated based on a set time period.

7. A knowledge graph-based intelligent decision-making system for the phosphate mining industry, characterized in that, include: The acquisition module is used to collect production data in the phosphate mining industry through smart sensors, and to perform edge processing on the production data through edge computing nodes to generate decision information. The knowledge module is used to generate knowledge tuples representing the equipment status based on the production data and the decision information; The encoding module is used to map entity devices and their relationship information to device nodes and edges of a Bayesian network, and generate knowledge state encoding and hidden state at a set time based on the confidence weight of the device nodes and the knowledge tuples. The probability module is used to predict the knowledge state and probability of the device node at the next time step based on the hidden state at the set time. The decision-making module is used to generate a dynamically evolving knowledge graph based on the hidden state at the set time, the knowledge state at the next time, and the probability, so as to make decisions on phosphate mining production through the knowledge graph; Specifically, based on the production data and the decision information, a knowledge tuple representing the equipment status is generated, including: Based on the production data and the decision information, time series information and confidence weights are determined; specifically, the production data is sorted in chronological order to determine the time series information of the production data; the confidence weights for data quality and the confidence weights based on the sufficiency of decision-making basis are comprehensively calculated to obtain the final confidence weights. Based on the time series information and confidence weights, knowledge tuples representing equipment status are generated. Specifically, for each equipment status feature extracted at a given time point, the time position of the equipment status feature in the production process is determined based on its corresponding time series information. The reliability of the equipment status features is quantified based on the confidence weights. Using the equipment as the basic unit, the equipment status features that integrate time series information and confidence weights are organized into knowledge tuples.

8. The knowledge graph-based intelligent decision-making system for the phosphate mining industry according to claim 7, characterized in that, Production data from the phosphate mining industry is collected through smart sensors, and edge computing nodes are used to process this production data to generate decision information, including: Production data in the phosphate mining industry is collected using smart sensors; Edge computing nodes are pre-deployed near smart sensors, and the production data is processed at the edge through these edge computing nodes to generate decision information.

Images