Internet of Things-Based Gas Station Equipment Monitoring System and Method

Abstract

The present invention relates to the technical field of gas station equipment monitoring, specifically a gas station equipment monitoring system and method based on the Internet of Things. Specifically, it includes: real-time collection of multi-dimensional operation parameters of equipment and dynamics of business process nodes, construction of a two-way coupling graph model of equipment and business, and extraction of high-order features of physical associations and business logic interactions between equipment using a graph convolutional network; adoption of a dynamic health scoring mechanism to predict equipment failure risks, combination of an attention mechanism to trace the risk propagation chain, and mapping of multi-dimensional data into interpretable fault diagnosis information through an abnormal semantic generation network; introduction of a double-layer reinforcement learning game mechanism, with the inner layer coordinating equipment maintenance and risk suppression strategies, and the outer layer overall planning business efficiency and safety goals, generating optimized actions such as resource scheduling and oil product allocation, forming a perception-diagnosis-decision closed loop. The present invention realizes full-time perception of equipment status, active fault warning, and cross-domain collaborative optimization, significantly improving the operation and maintenance efficiency and safety of gas stations.

Description

Technical Field

[0001] The present invention relates to the technical field of gas station equipment monitoring, and particularly to an Internet of Things-based gas station equipment monitoring system and method. Background Art

[0002] With the expansion of the operation scale of gas stations and the complication of service requirements, the equipment monitoring system needs to cope with challenges such as the perception of the collaborative operation status of multiple devices, the adaptation of business dynamic loads, and the prevention and control of risk propagation chains; traditional gas station equipment monitoring is based on manual inspections and local sensor alarms, making it difficult to meet the requirements of real-time data fusion analysis and cross-domain collaborative optimization.

[0003] A Chinese invention patent with the publication number CN110702852B discloses an intelligent monitoring system based on multiple oil and gas concentration sensors of the Internet of Things. The system consists of an environmental parameter acquisition platform for the gas station oil tank area based on the ZigBee network and a monitoring subsystem for the environmental oil and gas concentration sensors in the gas station oil tank area; effectively solving the problem that the existing environmental monitoring system for the gas station oil tank area does not accurately detect the environmental oil and gas concentration in the gas station oil tank area and give early warnings of sensor failures according to the non-linear, large lag, and complex changes of the environmental oil and gas concentration in the gas station oil tank area, thus improving the accuracy and robustness of predicting the oil and gas concentration in the gas station.

[0004] However, there are still problems in traditional gas station equipment monitoring such as information islands, lagging responses, and low data utilization rates. With the development of the Internet of Things, cloud computing, and big data analysis, a new system that can achieve full-scale perception of equipment status, dynamic behavior modeling, intelligent optimization scheduling, and active fault warning is needed to improve the intelligence and refinement level of gas station equipment management. Summary of the Invention

[0005] The purpose of the present invention is to propose an Internet of Things-based gas station equipment monitoring system and method for the problems existing in the background art.

[0006] The technical solution of the present invention: An Internet of Things-based gas station equipment monitoring method includes the following specific implementation steps:

[0007] S1. Multidimensional operating parameters and dynamic business process nodes of key equipment in the gas station;

[0008] S2. By constructing a set of equipment and business process nodes and associating state vectors to describe the operating characteristics, using the attention mechanism to dynamically adjust the coupling weights, and combining with the graph convolutional network to extract high-order interaction features, an equipment-business coupling model is established;

[0009] S3. The comprehensive health representation vector is constructed by integrating the device status, neighbor node influence and coupling weight through the attention mechanism. The lightweight evaluation function is used to quantify the device health score. The graph time series model is combined to predict future failure risks, and the cumulative impact weight is traced back to identify the risk chain path.

[0010] S4. Construct a multi-dimensional anomaly feature vector by extracting device propagation features, use semantic generation network to map anomaly semantic categories, combine the device-business coupling model with the risk chain to deduce the anomaly causal path, and finally generate anomaly semantic recognition results;

[0011] S5. By constructing an operation state vector that includes equipment health, business indicators, and safety risks, defining a multi-objective optimization function and a two-layer reinforcement game optimization strategy generation mechanism, and dynamically adjusting the strategy priority in combination with abnormal semantic interpretation, the optimal operation decision is generated;

[0012] S6. Integrate the optimal operational decision and abnormal analysis results, jointly execute automated resource scheduling, combine manual intervention and continuous monitoring and feedback to dynamically adjust the strategy.

[0013] Preferably, the construction process of the device-service coupling model is:

[0014] S21. Define device collection , and define a collection of business process nodes , build a set of all nodes ;

[0015] Among them, e i represents the i-th key equipment node in the gas station; b j represents the jth business node of the business process; m represents the total number of key equipment in the gas station; n represents the total number of business process steps;

[0016] S22. Define each device node e i and business node b j The associated state vector , :

[0017] ; ;

[0018] in, Represents the kth state feature of the i-th device node; represents the lth process feature of the jth service node; M represents the total number of state features in the i-th device node; L represents the total number of process features of the j-th service node;

[0019] S23. Define the device-service initial coupling association matrix A:

[0020] ;

[0021] Among them, A uv represents the preliminary connection weight between nodes u and v;

[0022] S24. Introduce the dynamic coupling weight tensor W(t) to adaptively weight and correct the initial correlation matrix:

[0023] ; ;

[0024] Among them, W(t) represents the dynamic coupling weight tensor at time t; , respectively represent the set of state vectors of devices and services; represents the element-wise product; represents the device-service coupling graph after dynamic weighting; represents the weight learning function;

[0025] S25. Apply the graph convolutional network GCN to perform joint representation learning on device and service nodes and extract high-order interaction features between nodes:

[0026] ; ;

[0027] Among them, , respectively represent the node representation of the (l + 1)-th layer and the node representation of the l-th layer; represents the trainable weight of the l-th layer; represents the ReLU activation function; D represents the degree matrix; represents the dynamic weighted connection weight between nodes i and j at time t; represents the total connection weight of node i;

[0028] S26. Output the device-service coupling model G(t): ;

[0029] Among them, N represents the node set; represents the connection relationship between nodes after dynamic weighting; represents the joint embedding representation of each node.

[0030] Preferably, the risk chain path identification process is as follows:

[0031] S31. For each device node , construct its comprehensive health characterization vector:

[0032] ;

[0033] Among them, Represents the current device state vector; Represents the neighbor nodes coupled with the current device node e i ; Represents the dependence intensity learned by the attention mechanism; W1 represents the weight matrix; Represents the non-linear activation function ReLU; Represents the comprehensive health representation vector of device e i ;

[0034] S32. Use the comprehensive health representation vector of device e i to quantitatively score the current health state of device e : i ; ;

[0035] Among them, represents the health score, that is, the health score of device e i at time t. The closer it is to 0, the more abnormal it is, ; w2 and b respectively represent the weights and bias terms obtained by model training; the here represents the vector transpose operation;

[0036] Set the health risk threshold θ. When , it is considered that there is a health risk and enter the prediction process;

[0037] S33. Select the health representation sequences of the past time instants and input them into the graph time series prediction model GGRU: ;

[0038] Further predict the future fault risk score: ;

[0039] Among them, represents the health score of predicting device e i at the future τ time step, that is, the health score at the future time t + τ; represents the health representation of predicting device e i at the future time t + τ; represents the graph gated recurrent unit model, which takes into account both structure perception and time series memory capabilities;

[0040] S34. Identify high-risk devices through the health score threshold, extract the key upstream nodes by backtracking the cumulative influence weights until the termination condition, and output the risk chain path.

[0041] Preferably, the implementation process of extracting the key upstream nodes by backtracking the cumulative influence weights until the termination condition and outputting the risk chain path is as follows:

[0042] A1. For each device e i , set the health risk threshold θ according to the health score obtained from future predictions . If the following condition is satisfied: low < θ < θ low , then determine that device e i is a high - risk node of potential failure;

[0043] A2. Extract the influence weight of each edge (ei, ej) in the device - service coupling model G(t) on the health status of the node at time t , and construct the cumulative influence weight matrix of each node within the time window [t−NumT + 1, t]:

[0044] ;

[0045] where represents the cumulative health impact degree of device e j on device e i within the past NumT time moments;

[0046] A3. For each high - risk device node e i , perform the following backtracking:

[0047] Starting from e i , retrieve its set of direct neighbor nodes in reverse (i.e., the nodes that affected it in the past);

[0048] Arrange them in descending order according to the cumulative influence weight , and select the nodes e j whose cumulative influence weight is greater than the set threshold γ;

[0049] Recursively backtrack the upstream influence nodes of e j , continuously construct the path, and the termination conditions are: the cumulative weight drops below the threshold; the length of the backtracked path reaches the maximum depth d max ; reaching the source service node;

[0050] A4. Finally, output the risk chain path Chain(e i ).

[0051] Preferably, the process of generating the abnormal semantic recognition result is as follows:

[0052] S41. Extract the propagation feature vector of device e i ; ;

[0053] S42. For each device e i, construct a set of multi-dimensional anomaly feature vectors by synthesizing the current status, predicted trends, changes in business metrics, and link propagation characteristics : ;

[0054] Among them, represents the health score of device e i at time t; represents the predicted health score of device e i at the future τ time steps; represents the magnitude of health change, ; represents the set of current key business metrics; represents the propagation feature vector of device e i ;

[0055] S43. Introduce the anomaly semantic generation network SGN, with the goal of mapping through the feature vector to the specific set of anomaly semantic categories C;

[0056] Define the classification probability distribution : ;

[0057] S44. According to the identified anomaly categories, combine the device-business coupling graph G(t) with the risk chain Chain(e i ), and deduce the logical path of anomaly formation;

[0058] S45. Output the anomaly semantic recognition result.

[0059] Preferably, the extraction process of the propagation feature vector i of device e is as follows:

[0060] B1. Extract the basic attributes of the link itself:

[0061] Chain length : The shortest propagation steps from e i to the end node;

[0062] Maximum propagation width : The layer with the largest number of child nodes at any level;

[0063] Node clustering coefficient change : The change in the node clustering coefficient along the chain;

[0064] B2. During the propagation process, extract the risk transfer weight w(e i , e j ) of each edge, that is, the connection between devices, and define device ei Link cumulative propagation weight : , extract the average single-hop propagation intensity : ;

[0065] Among them, (e u , e v ) represents that in the risk chain Chain(e i ), there is a potential risk transmission relationship between device e u and device e v ; w(e u , e v ) represents the propagation intensity when describing the risk transmission from device e u to device e v ;

[0066] B3. Define the propagation eigenvector of device e i as: .

[0067] Preferably, the abnormal semantic generation network SGN converts multi-modal data into abnormal categories with business semantics. The input feature vector includes the current health score, prediction score, trend difference, business key indicators, and risk chain propagation characteristics. The semantic features of each category are extracted through a multi-layer perceptron MLP, and the probability distribution of abnormal semantic categories is generated through the Softmax function.

[0068] Preferably, the optimal operation decision-making process is as follows:

[0069] S51. Construct the current gas station operation status vector S t : ;

[0070] Among them, represents the vector composed of the health scores of all devices e i at time t; represents the business operation index vector; represents the current operation cost vector; represents the safety risk chain score vector; represents the time context information;

[0071] S52. Define the action set A = {a1, a2,..., a t ,..., a K}, each action a t represents an operation strategy, and construct the optimization objective function: ;

[0072] Among them, Denote the expected optimization benefit of policy π; π represents the policy function, which takes the current state S as input t and outputs the optimal action a t ; T represents the total number of steps in the decision-making period, that is, the time-step length considered in this round of optimization plan; denotes the discount factor, 0 < γ < 1; denotes the operating cost function, and the operating overhead corresponding to executing action a t in state S t ; denotes the safety and service benefit function; denotes the safety-cost regulation weight factor;

[0073] S53. Construct a two-layer reinforcement game optimization strategy generation mechanism, a two-layer structure of inner-layer device game + outer-layer service game, and model the game tension in the strategy evolution process based on reinforcement learning to generate multi-objective optimization strategies;

[0074] S54. Introduce a semantic interpretation regulation unit, abnormal semantic interpretation results, adjust the focus of the optimization strategy, and correct the action set generated by reinforcement learning through soft constraints;

[0075] S55. Select the optimal strategy .

[0076] Preferably, the implementation process of the two-layer reinforcement game optimization strategy generation mechanism is as follows:

[0077] C1. Participating subjects and state-action modeling:

[0078] Inner-layer game subjects, that is, device agents: Each key device e i is modeled as a reinforcement learning agent Agent i ;

[0079] Outer-layer game subjects, that is, business service agents: The gas station business process is modeled as a service agent Agent b ;

[0080] C2. Inner-layer game mechanism, that is, device game: Health-chain risk control tension modeling: Model the problem as a repeated game reinforcement learning, and each Agent i learns the strategy and participates in the following local benefit game: ;

[0081] Among them, denotes the cumulative benefit of device agent e i during the entire game period; denotes the discount factor, 0 < γ < 1; denotes device e i executing action at After that, the change in its health score; Indicates the risk chain propagation impact of the current device failure on other devices; Indicates the current action cost; α, β, and θ represent dynamic weight coefficients;

[0082] C3. Outer game mechanism, i.e., business game: Global efficiency - risk control coordination strategy: Business layer Agent b Observe the health status and task requirements of all devices, and generate a scheduling strategy based on the global state , and its revenue function: , output policy guiding factors: α, β, θ, and inject the policy guiding factors: α, β, θ into the revenue functions of all device Agents i ;

[0083] Among them, Indicates the cumulative operating revenue of the business agent during the entire game cycle; Indicates the current business service quality indicator; Indicates the overall device risk level of the current system; Indicates the operating cost caused by the current scheduling strategy; , , Indicates the business strategy preference weight;

[0084] C4. At each moment t, the game process is as follows:

[0085] Business agent Agent b Observe the global state S t , adjust the weights α, β, θ; each Agent i Under the guidance of the weights, execute device actions ; All device actions affect the system state S t+1 , feedback to the business agent; use the multi-agent reinforcement learning framework MADDPG to iteratively update the strategy;

[0086] C5. Define the composite revenue function of the overall system: ;

[0087] Among them, Indicates the composite revenue indicator after the overall system strategy is executed; Indicates the weight control parameter of the business layer revenue in the total target; N represents the number of devices participating in the inner game;

[0088] By setting and using the Pareto front screening mechanism, select the optimal strategy combination of multiple objectives { , };

[0089] Among them, represents the device action combination; represents the scheduling strategy;

[0090] C6. The final output is an optimization strategy with multiple objectives: the device layer action strategy table, the business layer operation strategy, and the strategy interpretive structure.

[0091] The technical solution of the present invention: A gas station equipment monitoring system based on the Internet of Things, which is used to execute the above-mentioned gas station equipment monitoring method based on the Internet of Things, includes:

[0092] The Internet of Things collaborative perception module is used to collect the operation data of various devices in the gas station in real time and monitor the dynamics of business process nodes at the same time;

[0093] The device-business two-way modeling engine is used to model the association between the device state and the business process state as a two-way coupling graph structure, use the graph structure to show the physical and logical connections between devices, and define the interaction rules between device load and business flow;

[0094] The abnormal semantic analysis module is used to construct a device health joint inference and prediction mechanism under dynamic business load and an abnormal parsing mechanism based on the device-business semantic structure. The abnormal semantic analysis module includes: a device health joint inference unit and an abnormal semantic recognition and interpretation unit;

[0095] The device health joint inference unit dynamically infers the device health state through a joint inference algorithm, takes into account the current state of the device and future business load changes, accurately predicts the occurrence probability of potential faults, so as to achieve early warning and take measures in advance;

[0096] The abnormal semantic recognition and interpretation unit deeply analyzes the predicted anomalies, and provides accurate fault diagnosis information for maintenance personnel by understanding the reasons, influence scope and related business processes behind the device anomalies;

[0097] The dynamic operation optimization module is used to comprehensively optimize fault warning, business load and device state. Based on the objective function of minimizing fault risk, operation cost and customer waiting time, an optimal resource scheduling plan is generated through an optimization algorithm;

[0098] The intelligent alarm and resource scheduling module is used for real-time alarm and automatically executes the optimization plan, schedules device resources and arranges necessary maintenance work.

[0099] Compared with the prior art, the above technical solution of the present invention has the following beneficial technical effects:

[0100] A gas station equipment monitoring system and method based on the Internet of Things provided by the present invention collect real-time equipment operation parameters and dynamic data of business processes; construct a dynamic coupling graph structure, and combine a graph convolutional network to extract high-order interaction features of physical associations and business logics between devices, realizing in-depth collaborative analysis of equipment status and business loads; adopt a joint inference algorithm to predict equipment health risks, combine an attention mechanism to trace the risk propagation chain and generate interpretable fault diagnosis information, effectively solving the ambiguity problem of traditional numerical alarms; through a two-layer reinforcement learning game mechanism, balance the goals of equipment maintenance, risk suppression, and business efficiency, dynamically generate resource scheduling and oil product allocation strategies, significantly reducing the operation cost and the risk of fault chain diffusion; realize the automatic execution and closed-loop feedback of optimization strategies, forming a perception-diagnosis-decision integrated system;

[0101] Compared with traditional manual inspections and static monitoring, this solution breaks through the limitations of information silos, improves data utilization and operation and maintenance response speed, reduces equipment failure rates through predictive maintenance, and optimizes the overall operation safety and economy of gas stations. Brief Description of the Drawings

[0102] Figure 1 It is a system architecture diagram of a gas station equipment monitoring system based on the Internet of Things proposed by the present invention;

[0103] Figure 2 It is a method flow chart of a gas station equipment monitoring method based on the Internet of Things proposed by the present invention. Detailed Embodiments

[0104] Example 1, as Figure 1 shown, a gas station equipment monitoring system based on the Internet of Things proposed by the present invention includes: an Internet of Things collaborative perception module, a device-business two-way modeling engine, an abnormal semantic analysis module, a dynamic operation optimization module, and an intelligent alarm and resource scheduling module.

[0105] The Internet of Things collaborative perception module collects real-time operation data of various devices in the gas station, including but not limited to the status information of oil pumps, fuel dispensers, storage tanks, and payment terminal devices, and at the same time monitors the dynamics of business process nodes (including but not limited to refueling, payment, car washing), and this data flows to the device-business two-way modeling engine;

[0106] The device-business two-way modeling engine models the association between the device status and the business process status as a two-way coupling graph structure, uses the graph structure to show the physical and logical connections between devices, and defines the interaction rules of device load and business flow, so as to provide a reliable basis for subsequent reasoning and analysis;

[0107] Anomaly Semantic Analysis Module, which constructs a device health joint inference and prediction mechanism under dynamic business loads and an anomaly parsing mechanism based on the device-business semantic structure. The Anomaly Semantic Analysis Module includes: a Device Health Joint Inference Unit and an Anomaly Semantic Recognition and Explanation Unit;

[0108] The Device Health Joint Inference Unit dynamically infers the device health status through a joint inference algorithm. Considering the current status of the device and future changes in business loads, it accurately predicts the occurrence probability of potential failures, thereby achieving early warning and taking preventive measures in advance;

[0109] The Anomaly Semantic Recognition and Explanation Unit deeply analyzes the predicted anomalies. By understanding the reasons, scope of influence, and related business processes behind device anomalies, it provides accurate fault diagnosis information for maintenance personnel, avoiding simple numerical alarms;

[0110] The Dynamic Operation Optimization Module comprehensively optimizes fault warnings, business loads, and device statuses. Based on the objective function of minimizing fault risks, operation costs, and customer waiting times, it generates an optimal resource scheduling plan through an optimization algorithm, including but not limited to automatically switching to standby devices and adjusting the oil supply path;

[0111] The Intelligent Alarm and Resource Scheduling Module issues real-time alarms and automatically executes the optimization plan, scheduling device resources and arranging necessary maintenance work.

[0112] Embodiment 2, as Figure 2 shown, a method for monitoring gas station equipment based on the Internet of Things proposed by the present invention is applied to a system for monitoring gas station equipment based on the Internet of Things proposed in Embodiment 1. The specific implementation steps are as follows:

[0113] S1. The Internet of Things collaborative perception module, through deploying a multi-source sensor network, real-time collects multi-dimensional operation parameters of key equipment in the gas station (including but not limited to oil pumps, fuel dispensers, storage tanks, pipelines, payment terminals), including but not limited to flow rate, pressure, temperature, vibration, current, voltage, and payment processing status;

[0114] At the same time, business process nodes (including but not limited to customer queuing, payment completion, and oil product switching) are used as perception elements for synchronous monitoring;

[0115] The perception data stream is transmitted to the device-business bidirectional modeling engine.

[0116] S2. The device-business two-way modeling engine establishes a two-way coupling model between the device state vector and the business process state vector based on the collected perception data stream. It uses a graph structure to represent the physical associations of devices (such as the connection relationship between an oil pump and an oil gun) and the logical flow of business (such as the sequence of refueling → payment → car wash). At the same time, it defines the interaction rules between device loads and business dynamics, which are used to dynamically infer the trend of device operating pressure with business changes. The specific implementation process is as follows:

[0117] S21. Obtain the perception data stream and define the device set , and define the set of business process nodes , and construct the set of all nodes ;

[0118] Among them, e i represents the i-th key device node in the gas station, including but not limited to oil pumps, oil guns, and storage tanks; b j represents the j-th business node of the business process, including but not limited to the start of refueling, payment confirmation, and fuel type switching; m represents the total number of key devices in the gas station; n represents the total number of business process steps;

[0119] S22. Each device node e i and business node b j are respectively associated with a state vector, denoted as and , which are used to describe their current operating state or business state characteristics:

[0120] ;

[0121] ;

[0122] Among them, represents the k-th state characteristic of the i-th device node, including but not limited to flow rate, current, and vibration amplitude; represents the l-th process characteristic of the j-th business node, including but not limited to queue length, payment response time, and refueling completion rate; M represents the total number of state characteristics in the i-th device node; L represents the total number of process characteristics in the j-th business node;

[0123] S23. Define the initial coupling association matrix A of the device and business:

[0124] ;

[0125] Among them, A uv represents the preliminary connection weight between nodes u and v;

[0126] It should be noted that for physical associations: such as an oil pump driving an oil gun to supply oil; for logical associations: such as payment confirmation affecting the start and stop of fuel type switching;

[0127] S24. To reflect the real-time impact intensity between the changes in device status and business load, a dynamic coupling weight tensor W(t) is introduced to adaptively weighted correct the initial correlation matrix:

[0128] ;

[0129] ;

[0130] where W(t) represents the dynamic coupling weight tensor at time t; 、 respectively represent the set of state vectors of devices and services; represents element-wise product; represents the device-service coupling graph after dynamic weighting; represents the weight learning function (adaptive learning based on the attention mechanism in this embodiment);

[0131] S25. After obtaining the dynamic coupling graph , apply the graph convolutional network (GCN) to perform joint representation learning on device and service nodes to extract high-order interaction features between nodes:

[0132] ;

[0133] ;

[0134] where 、 respectively represent the node representation of the (l + 1)-th layer and the node representation of the l-th layer; represents the trainable weight of the l-th layer; represents the activation function (ReLU is used in this embodiment); D represents the degree matrix; represents the dynamic weighted connection weight between node i and node j at time t; represents the total connection weight of node i (including the sum of its weighted connection values with all other device nodes and service nodes);

[0135] S26. Output the device-service coupling model G(t): ;

[0136] where N represents the node set (device nodes + service nodes); represents the connection relationship between nodes after dynamic weighting (reflecting the correlation strength between devices and services that changes over time); represents the joint embedding representation of each node (the device-service interaction feature vector obtained through graph convolutional learning).

[0137] S3. Based on the constructed device-business coupling model, the device health joint inference unit dynamically infers the health status and failure probability of each device in the current and future periods using the joint inference algorithm. The inference process comprehensively considers the current perceived status of the device, the business load pressure, and the historical operation mode, and identifies potential failure risks in advance through predictive calculations. The specific implementation process is as follows:

[0138] S31. For each device node , considering factors such as its business location, interaction neighbors, and task participation in the coupling graph, construct its comprehensive health representation vector:

[0139] ;

[0140] Among them, represents the current device state vector (including but not limited to temperature, voltage, pressure); represents the neighbor nodes (including devices and services) coupled with the current device node e i ; represents the dependence strength learned by the attention mechanism, reflecting the influence of node j on the health status of e i ; W1 represents the weight matrix for linear transformation; represents the non-linear activation function (ReLU) to enhance the modeling ability; represents the comprehensive health representation vector of device e i ;

[0141] Accordingly, three factors of cross-domain influence (device-service collaboration), structural location (the role of the device in the task link), and coupling dynamics (coupling weights change at any time) are embedded in the representation to achieve health status perception driven by the service;

[0142] S32. Use the comprehensive health representation vector i of device e , input it into a lightweight neural evaluation function, and quantitatively score the current health status of device e i : ; Here, represents the vector transpose operation;

[0143] Among them, represents the health score, that is, the health score of device e i at time t. The closer it is to 0, the more abnormal it is, ; w2 and b respectively represent the weights and bias terms obtained from model training;

[0144] Set the health risk threshold θ. When < θ, it is considered that there is a health risk and enter the prediction process;

[0145] S33. Select the health representation sequences at the past moments and input them into the graph time series prediction model (i.e., the Gated Graph Recurrent Unit, GGRU, graph gated recurrent unit model): ;

[0146] ;

[0147] Further predict the future fault risk score: ;

[0148] Among them, represents the health score of the predicted device e i at the future τ time steps, that is, the health score at the future time t + τ; represents the health representation of the predicted device e i at the future time t + τ; represents the graph gated recurrent unit model, taking into account both structure awareness and time series memory capabilities;

[0149] S34. Use the nodes with a significant decrease in the predicted score (i.e., <θ low ) to trace back the influence path, specifically:

[0150] S3401. For each device e i , according to the health score obtained from the future prediction, set the health risk threshold θ low . If it satisfies: <θ low , then determine that the device e i is a potential high-risk fault node;

[0151] S3402. Extract the influence weight of each edge (ei, ej) in the device-business coupling model G(t) on the node health state at time t and construct the cumulative influence weight matrix of each node within the time window [t - NumT + 1, t]:

[0152] ;

[0153] Among them, represents the cumulative health influence degree of the device e j on the device e i within the past NumT moments;

[0154] S3403. For each high-risk device node e i , perform the following backtracking:

[0155] Start from e iStart from it and retrieve its set of direct neighbor nodes in reverse (i.e., the nodes that influenced it in the past);

[0156] Arrange them in descending order according to the cumulative influence weight and select the node e whose cumulative influence weight is greater than the set threshold γ j ;

[0157] Recursively backtrack the upstream influence nodes of e j and continuously construct the path until: the cumulative weight drops below the threshold; the length of the traced path reaches the maximum depth d max ; or reach the source service node (including but not limited to the oil tank area, fuel dispenser management server node);

[0158] Finally, output the risk chain path Chain(e i ).

[0159] S4. Based on the output results of the device health joint reasoning and fault trend prediction by the abnormal semantic recognition and interpretation unit (i.e., predicting the health score of device e i at the future τ time steps and its risk chain path Chain(e i ))), further mine and identify the potential abnormal semantics of the device, and deduce the causal chain of the abnormality, and finally generate an abnormality description with business interpretability to provide an intuitive basis for the operation and maintenance decision-making. The specific implementation process is as follows:

[0160] S41. Extract the propagation feature vector of device e i :

[0161] (1) Extract the basic attributes of the link itself:

[0162] Chain length : The shortest propagation steps from e i to the end node;

[0163] Maximum propagation width : The layer with the largest number of child nodes at any level;

[0164] Node clustering degree change : The change amount of the node clustering coefficient (Clustering Coefficient) along the chain;

[0165] (2) During the propagation process, extract the risk transfer weight w(e i , e j ) of each edge (the connection between devices) (i.e., representing the influence degree of the abnormal signal from one device to another device), and define the link cumulative propagation weight of device e i : : , extract the average single-hop propagation intensity : ;

[0166] Among them, (e u , e v ) indicates that in the risk chain Chain(e i ), there is a potential risk transmission relationship (i.e., one hop) between device e u and device e v ; w(e u , e v ) represents the propagation intensity when describing the risk transmission from device e u to device e v ;

[0167] (3) Define the propagation feature vector of device e i as: ;

[0168] S42. Construct an abnormal feature vector (abnormal multi-dimensional representation): For each device e i predicted to have potential high risk, construct a set of multi-dimensional abnormal feature vectors by integrating the current state, prediction trend, business metric changes, and link propagation characteristics: ;

[0169] Among them, represents the health score of device e i at time t; represents the predicted health score of device e i at the future τ time steps; represents the health change amplitude, ; represents the current set of key business metrics (including but not limited to traffic, temperature, oil pressure); represents the propagation feature vector of device e i ;

[0170] S43. Introduce an abnormal semantic generation network (Semantic Generation Network, SGN), the goal of which is to map through the feature vector to the specific set of abnormal semantic categories C;

[0171] Define the classification probability distribution : ;

[0172] It should be noted that the Anomaly Semantic Generation Network (SGN) is a multi-modal, interpretable, and self-supervised anomaly semantic classification network for intelligent monitoring of business processes. Its main function is to convert multi-modal data such as equipment health scores, risk chain features, and business metrics into anomaly categories with business semantics, such as oil and gas leakage trends, signal anomalies between devices, etc.; the input feature vectors of SGN include the current health score, predicted score, trend difference, business key metrics, and risk chain propagation features, extract various feature semantics through a multi-modal embedding module (i.e., through a multi-layer perceptron MLP), and generate the probability distribution of anomaly semantic categories through the Softmax function;

[0173] S44. According to the identified anomaly category, combined with the device-business coupling graph G(t) and the risk chain Chain(e i ), deduce the logical path of anomaly formation;

[0174] S45. Output the anomaly semantic recognition result in the form of a graph + text summary, including but not limited to:

[0175] Highlight the key nodes in the risk chain; mark the changes in anomaly indicators for each node; give a concise business explanation;

[0176] For example: "Fuel dispenser #5 detects a continuous decrease in oil pressure, resulting from fluctuations in the oil tank sensor + aging of the pipeline valve leading to abnormal flow control, predicting a potential oil and gas leakage hazard."

[0177] S5. The dynamic operation optimization module introduces a reinforcement learning + game regulation mechanism + cost-risk quantification function to jointly generate optimization actions, and dynamically converts the identified equipment health trends, risk chain propagation paths, and anomaly semantic explanations and other information into executable operation optimization strategies (including but not limited to scheduling strategies, maintenance arrangements, load reconfiguration, oil product allocation), realizing a closed-loop of perception-identification-reasoning-optimization, and improving the overall safety and economic efficiency of the gas station. The specific implementation process is as follows:

[0178] S51. Construct the current gas station operation state vector S t : ;

[0179] Among them, represents the vector composed of the health scores of all devices e i at time t; represents the business operation index vector (including but not limited to the remaining oil volume, the number of customers in line); represents the current operation cost vector (including but not limited to maintenance, energy, labor); represents the safety risk chain score vector (in this embodiment, the average value of the Chain propagation depth is adopted); Represent time context information (including but not limited to peak / low, weather, scheduling period);

[0180] S52. Define the action set A = {a1, a2, …, a t ,..., a K}, where each action a t represents an operation strategy, and construct the optimization objective function: ;

[0181] Among them, represents the expected optimization benefit of strategy π, comprehensively considering the minimization of operation cost and the maximization of safety benefit; π represents the strategy function, which takes the current state S t as input and outputs the optimal action a t ; T represents the total number of steps in the decision-making period, that is, the time step length considered in this round of optimization plan; represents the discount factor, which controls the present value of future benefits, 0 < γ < 1; represents the operation cost function, which is the operation overhead corresponding to state S t when action a t is executed; represents the safety and service benefit function, which is the reduced safety risk or increased customer satisfaction when action a t is executed; represents the safety-cost regulation weight factor, which adjusts the trade-off degree between safety and operation efficiency in the optimization process and can be dynamically adjusted;

[0182] S53. To ensure multi-objective constraints (safety + cost + efficiency), construct a two-layer reinforcement game optimization strategy generation mechanism, a two-layer structure of inner-layer device game + outer-layer business game, and model the game tension in the strategy evolution process based on reinforcement learning to generate multi-objective optimization strategies;

[0183] S54. Introduce a semantic interpretation regulation unit, adjust the focus of the optimization strategy according to the abnormal semantic interpretation result output by S4, and correct the action set generated by reinforcement learning through soft constraints;

[0184] Exemplarily, if the abnormality stems from "oil quality problem", priority is given to recommending blending or cleaning; if there is a "secondary chain abnormality" of a device on the risk chain, priority is given to closing the risk propagation node;

[0185] S55. Select and output the optimal strategy .

[0186] S6. The intelligent alarm and resource scheduling module receives the optimal strategy, the abnormal semantic recognition result, the risk chain path Chain(e i ), and the health score at the future time t + τ Link with the collected perception data stream, drive the resource scheduling device, automatically execute some decisions (including but not limited to switching devices and adjusting the fuel supply path), while guiding manual intervention for necessary repairs or maintenance operations, continuously monitoring the execution effect and dynamically adjusting the strategy to form a closed-loop optimization.

[0187] Embodiment 3. A gas station equipment monitoring method based on the Internet of Things proposed by the present invention further includes a double-layer reinforcement game optimization strategy generation mechanism, and its specific implementation steps are as follows:

[0188] A1. Participants and state-action modeling:

[0189] Inner-layer game participant (equipment agent): Each key equipment e i is modeled as a reinforcement learning agent Agent i , with the following characteristics: State input: equipment health status , risk propagation degree ρ i (t), business dependence degree B i (t); Action output: including but not limited to repair / unload / share load / tolerate operation; Objective function: maximize the recovery of health trend + minimize the local risk chain propagation;

[0190] Outer-layer game participant (business service agent): The gas station business process (including but not limited to refueling, dispensing, and customer queuing) is modeled as a service agent Agent b , which controls: the task / load allocation strategy among multiple devices; manpower scheduling and oil product resource allocation; Objective function: maximize the overall customer satisfaction / business throughput + minimize the probability of business interruption.

[0191] A2. Inner-layer game mechanism (equipment game): Health-chain risk control tension modeling:

[0192] Model the problem as a repeated game reinforcement learning, and each Agent i learns a strategy , and participates in the following local revenue game:

[0193] ;

[0194] Among them, represents the cumulative revenue of the equipment agent e i during the entire game cycle; represents the discount factor, 0 < γ < 1, and the behavior closer to the current moment is more important; represents the change in the health score of equipment e i after executing the action a t ; Indicates the risk chain propagation impact of the current device failure on other devices, measured as the chained decrease term of the adjacent device health; Indicates the current action cost, including but not limited to maintenance cost, downtime loss, resource occupancy; α, β, θ represent dynamic weight coefficients, namely the importance of health restoration, the importance of risk propagation suppression, and the impact degree of control behavior cost respectively;

[0195] Accordingly: A game strategy is introduced into the risk chain item, and the strategies between devices generate a non-independent coupling game structure. If a device delays processing, its adjacent nodes (in the risk propagation graph) may increase the burden, forming a local non-cooperative game: If Agent i ignores maintenance, Agent j (dependent device) will experience a decrease in health; The system uses the graph attention mechanism to estimate ChainLoss i (t), reflecting the adjacent influence degree.

[0196] A3. Outer game mechanism (business game): Global efficiency - risk control coordination strategy:

[0197] Business layer Agent b Observes the health status and task requirements of all devices, and generates a scheduling strategy based on the global state , and its revenue function:

[0198] ;

[0199] Among them, Represents the cumulative operating revenue of the business agent within the entire game cycle; Represents the current business service quality indicators, including but not limited to the average customer waiting time, fueling task completion rate, service delay indicator; Represents the current overall device risk level of the system, aggregating all device risk chain impact values; Represents the operating cost caused by the current scheduling strategy, including but not limited to personnel scheduling overhead, oil product resource allocation cost, service switching loss; , , Represents the business strategy preference weights, namely the QoS priority, the importance of safety risk control, and the cost sensitivity degree;

[0200] Output strategy guiding factors: α, β, θ (injected as control parameters into the revenue functions of all device Agents i );

[0201] Accordingly: Embed the device layer strategy as a dynamic variable into the business layer game: Agent bInstead of directly controlling the device, it affects the device's policy direction through dynamic weights (α, β, θ) to achieve soft game guidance.

[0202] A4. Double-layer game co-evolution mechanism:

[0203] At each moment t, the game process is as follows:

[0204] Business agent Agent b Observes the global state S t , and adjusts the weights α, β, θ;

[0205] Each Agent i Under the guidance of the weights, executes the device actions ;

[0206] All device actions affect the system state S t+1 , and are fed back to the business agent;

[0207] Uses the multi-agent reinforcement learning framework (MADDPG is adopted in this embodiment) to iteratively update the policy;

[0208] Accordingly: Construct a business policy to regulate the device policy → the device response generates system feedback → the system feedback optimizes the business policy, forming a human-device-business multi-agent collaborative game feedback closed loop.

[0209] A5. Game convergence and policy selection mechanism:

[0210] Define the composite revenue function of the overall system: ;

[0211] Among them, represents the composite revenue index after the overall system policy is executed; represents the weight control parameter of the business layer revenue in the total target; N represents the number of devices participating in the inner layer game;

[0212] Accordingly: By setting and using the Pareto front screening mechanism, judge whether there exists: a set of device action combinations , a set of scheduling policies , such that any unilateral adjustment will cause the overall revenue to decline, which is regarded as the optimal collaborative policy output.

[0213] A6. The final output is an optimization strategy with multiple objectives: the device layer action policy table, the business layer operation policy, and the policy interpretive structure.

[0214] The embodiments of the present invention have been described in detail above with reference to the accompanying drawings, but the present invention is not limited to this. Within the scope of knowledge possessed by those skilled in the art of the relevant technology, various changes can be made without departing from the purpose of the present invention.

Claims

1. An Internet of Things-based gas station equipment monitoring method, characterized in that, The specific implementation steps include the following: S1. Collect multi-dimensional operating parameters of key equipment in gas stations and dynamic data of business process nodes; S2. By constructing a set of equipment and business process nodes and associating state vectors to describe the operation characteristics, the attention mechanism is used to dynamically adjust the coupling weights, and the graph convolutional network is combined to extract high-order interaction features to establish a device-business coupling model; The construction process of the device-business coupling model is as follows: S21. Define a set of devices , and define a set of business process nodes , and construct a set of all nodes ; Among them, e i represents the i-th key equipment node in the gas station; b j represents the j-th business node of the business process; m represents the total number of key equipment in the gas station; n represents the total number of business process steps; S22. Define the state vector associated with each device node e i and service node b j , :​ ； ； Among them, represents the k-th state feature of the i-th device node; represents the l-th process feature of the j-th service node; M represents the total number of state features in the i-th device node; L represents the total number of process features in the j-th service node; S23. Define the device-service initial coupling association matrix A: ； Among them, A uv represents the preliminary connection weight between nodes u and v; S24, introduce the dynamic coupling weight tensor W(t) to perform adaptive weighted correction on the initial correlation matrix: ； ； Among them, \(W(t)\) represents the dynamic coupling weight tensor at time \(t\); and respectively represent the state vector sets of devices and services; represents the element-wise product; represents the device-service coupling graph after dynamic weighting; represents the weight learning function; S25. Apply graph convolutional network GCN to perform joint representation learning on devices and business nodes and extract high-order interaction features between nodes: ； ； Among them, , represent the representation of the nodes in the (l + 1)-th layer and the representation of the nodes in the l-th layer respectively; represents the trainable weights in the l-th layer; represents the ReLU activation function; D represents the degree matrix; represents the dynamic weighted connection weight between node i and node j at time t; represents the total connection weight of node i. S26. Output device - service coupling model G(t): ; Among them, N represents the node set; represents the connection relationship between nodes after dynamic weighting; represents the joint embedding representation of each node; S3. The comprehensive health representation vector is constructed by integrating the device status, neighbor node influence and coupling weight through the attention mechanism. The lightweight evaluation function is used to quantify the device health score. The graph time series model is combined to predict future failure risks, and the cumulative impact weight is traced back to identify the risk chain path. The risk chain path identification process is as follows: S31. For each device node , construct its comprehensive health characterization vector: ； Among them, represents the current device state vector; represents the neighbor nodes coupled with the current device node e i ; represents the dependence intensity learned by the attention mechanism; W1 represents the weight matrix; represents the non-linear activation function ReLU; represents the comprehensive health characterization vector of device e i ; S32. Use device e i comprehensive health characterization vector of device e i to quantitatively score the current health status of device e: ; Among them, represents the health score, that is, the health score of device e i at time t. The closer it is to 0, the more abnormal it indicates. ; w2 and b respectively represent the weights and bias terms obtained by model training; here represents the vector transpose operation. Set the health risk threshold θ. When < θ, it is considered that there is a health risk and enter the prediction process; S33. Select the health representation sequences at the past number of moments , and input them into the graph time series prediction model GGRU: ; Further predict the future failure risk score: ; Among them, represents the health score of the prediction device e i at the future τ time steps, that is, the health score at the future time t + τ; represents the predicted health representation of device e i at the future time t + τ; represents the graph gated recurrent unit model, which takes into account both structure perception and time series memory capabilities; S34. Identify high-risk equipment through health score thresholds, extract cumulative impact weights, trace back to key upstream nodes until termination conditions, and output the risk chain path; S4. Construct a multi-dimensional anomaly feature vector by extracting device propagation features, use semantic generation network to map anomaly semantic categories, combine the device-business coupling model with the risk chain to deduce the anomaly causal path, and finally generate anomaly semantic recognition results; The process of generating abnormal semantic recognition results is as follows: S41. Extracting device e i 's propagation eigenvector ; S42. For each device e predicted to have potential high risk i , construct a set of multi-dimensional anomaly feature vectors by integrating the current state, prediction trend, business metric changes, and link propagation characteristics : ; Among them, represents the health score of device e i at time t; represents the predicted health score of device e i at the future τ time steps; represents the magnitude of health change, ; represents the current set of key business indicators; represents the propagation eigenvector of device e i ; S43. Introduce an abnormal semantic generation network SGN, with the goal of mapping through a feature vector to a specific set C of abnormal semantic categories; Define the categorical probability distribution : ; S44. According to the identified exception category, in combination with the device-business coupling graph G(t) and the risk chain Chain(e i ), deduce the logical path for the formation of the exception; S45, outputting abnormal semantic recognition results; S5. By constructing an operation state vector that includes equipment health, business indicators, and safety risks, defining a multi-objective optimization function and a two-layer reinforcement game optimization strategy generation mechanism, and dynamically adjusting the strategy priority in combination with abnormal semantic interpretation, the optimal operation decision is generated; S6. Integrate the optimal operational decision and abnormal analysis results, jointly execute automated resource scheduling, combine manual intervention and continuous monitoring and feedback to dynamically adjust the strategy.

2. The method for monitoring gas station equipment based on the Internet of Things according to claim 1, characterized in that The implementation process of extracting the cumulative impact weight and tracing back to the key upstream nodes until the termination condition and outputting the risk chain path is as follows: A1. For each device e i , based on the health score obtained from future predictions , set the health risk threshold θ low . If it satisfies: < θ low , then determine that device e i is a high-risk node for potential failures; A2. Extract the influence weight of each edge (ei, ej) in the device-business coupling model G(t) on the node health state at time t , and construct the cumulative influence weight matrix of each node within the time window [t−NumT+1,t]: ； Among them, represents device e j the cumulative health impact degree on device e within the historical NumT moments i ; A3. For each high-risk device node e i , perform the following backtracking: Starting from e i Retrieve its set of direct neighbor nodes in reverse, i.e., the nodes that influenced it in the past; According to the cumulative influence weight Arrange them in descending order, and select the node e whose cumulative influence weight is greater than the set threshold γ j ; Recursive backtracking e j The upstream influence nodes of, continuously build paths, and the termination conditions are: Condition 1: The cumulative weight drops below the threshold; Condition 2: The length of the backtracking path reaches the maximum depth d max ; Condition 3: Reach the source service node; If any of the above conditions is met, the recursion is terminated; A4. Final output risk chain path Chain(e i ) 3. The method for monitoring gas station equipment based on the Internet of Things according to claim 2, wherein, Device e i propagation eigenvector The extraction process is as follows: B1. Extract the basic properties of the link itself: Chain length : The shortest propagation steps from e i to the end node; Maximum propagation width : The layer with the largest number of child nodes at any level; Change in node clustering coefficient : The change in the clustering coefficient of nodes along the chain; B2. During the propagation process, extract the risk transmission weight w(e i ,e j ) of each edge, i.e., the connection between devices, and define the link cumulative propagation weight of device e i : . Extract the average single-hop propagation intensity : ; Among them, (e u , e v ) indicates that in the risk chain Chain(e i ), there is a potential risk transmission relationship between device e u and device e v ; w(e u , e v ) represents the propagation intensity when describing the risk transmission from device e u to device e v ; B3. Define device e i The propagation eigenvector of is as follows:

4. The method for monitoring gas station equipment based on the Internet of Things according to claim 3, characterized in that, The anomaly semantic generation network (SGN) converts multimodal data into anomaly categories with business semantics. The input feature vector includes the current health score, predicted score, trend difference, business key indicators and risk chain propagation characteristics. The multi-layer perceptron (MLP) is used to extract various feature semantics, and the probability distribution of anomaly semantic categories is generated through the Softmax function.

5. The method for monitoring gas station equipment based on the Internet of Things according to claim 4, wherein The optimal operational decision-making process is as follows: S51. Construct the current gas station operation status vector S t : ; Among them, represents the vector composed of the health scores of all devices e i at time t; represents the business operation index vector; represents the current operating cost vector; represents the safety risk chain score vector; represents the time context information; S52. Define the action set A = {a1, a2, …, a t ,..., a K}, where each action a t represents an operation strategy, and construct the optimization objective function: ; Among them, represents the expected optimization return of policy π; π represents the policy function, which takes the current state S as input t and outputs the optimal action a t ; T represents the total number of steps in the decision-making period, that is, the time-step length considered in this round of optimization plan; represents the discount factor, 0 < γ < 1; represents the operating cost function, and the operating expense corresponding to the state S when the action a is executed t ; t ; represents the safety and service return function; represents the safety-cost regulation weight factor; S53. Construct a double-layer reinforcement game optimization strategy generation mechanism, with a double-layer structure of inner device game + outer business game, and generate multi-objective optimization strategies based on the game tension in the strategy evolution process of reinforcement learning modeling; S54, introduce semantic interpretation control unit, abnormal semantic interpretation results, adjust the optimization strategy focus, and correct the action set generated by reinforcement learning through soft constraints; S55. Select the optimal strategy .

6. The method for monitoring gas station equipment based on the Internet of Things according to claim 5, characterized in that, The implementation process of the double-layer enhanced game optimization strategy generation mechanism is as follows: C1. Participants and state-action modeling: Inner game player, i.e., the device agent: Each key device e i is modeled as a reinforcement learning agent Agent i ; The outer game player, i.e., the business service agent: The gas station business process is modeled as the service agent Agent b ; C2. Inner game mechanism, i.e., device game: Health-chain risk control tension modeling: Model the problem as a repeated game reinforcement learning, and each Agent i learns a strategy , and participates in the following local revenue game: ; Among them, represents the cumulative revenue of the device agent e i during the entire game cycle; represents the discount factor, where 0 < γ < 1; represents the device e i after executing the action a t the change in its health score; represents the impact of the current device failure on the risk chain propagation of other devices; represents the current action cost; α, β, and θ represent dynamic weight coefficients; C3. Outer game mechanism, i.e., business game: global efficiency - risk control collaborative strategy: business layer Agent b Observe the health status and task requirements of all devices, and generate a scheduling strategy based on the global state , and its revenue function: , output policy guiding factors: α, β, θ, and inject the policy guiding factors: α, β, θ into the revenue functions of all device Agents i ; Among them, represents the cumulative operating income of the business agent over the entire game cycle; represents the current business service quality indicator; represents the current overall equipment risk level of the system; represents the operating cost caused by the current scheduling strategy; , , represents the business strategy preference weight; C4. At each moment t, the game process is as follows: Business Intelligence Agent b Observe the global state S t , adjust the weights α, β, θ; each Agent i Under the guidance of the weights, perform device actions ; all device actions affect the system state S t+1 , feedback to the business intelligence agent; use the multi-agent reinforcement learning framework MADDPG to iteratively update the policy; C5. Define the composite revenue function for the overall system: ; Among them, represents the composite return index after the overall system strategy is executed; represents the weight control parameter of the business layer return in the total target; N represents the number of devices participating in the inner layer game; By setting the use of the Pareto front screening mechanism, a multi-objective optimization strategy combination { , } is selected; Among them, represents the device action combination; represents the scheduling policy; C6. The final output is a multi-objective optimization strategy: the device layer action strategy table, the business layer operation strategy, and the strategy interpretive structure.

7. An Internet-of-Things-based gas station equipment monitoring system, which is used to execute an Internet-of-Things-based gas station equipment monitoring method described in any one of claims 1 to 6, characterized in that, Including: The Internet of Things collaborative sensing module is used to collect the operation data of various devices in the gas station in real time and monitor the dynamics of business process nodes at the same time; The device-business bidirectional modeling engine is used to model the association between device states and business process states as a bidirectional coupling graph structure, display the physical and logical connections between devices using the graph structure, and define the interaction rules between device loads and business flows; The abnormal semantic analysis module is used to construct a device health joint inference and prediction mechanism under dynamic business loads and an abnormal parsing mechanism based on the device-business semantic structure. The abnormal semantic analysis module includes: a device health joint inference unit and an abnormal semantic recognition and interpretation unit; The device health joint inference unit dynamically infers the device health state through a joint inference algorithm, takes into account the current state of the device and future changes in business loads, accurately predicts the probability of potential failures, and thus realizes early warning and takes measures in advance; The abnormal semantic recognition and interpretation unit deeply analyzes the predicted abnormalities, provides accurate fault diagnosis information for maintenance personnel by understanding the reasons, scope of influence, and related business processes behind device abnormalities; The dynamic operation optimization module is used to comprehensively optimize fault warnings, business loads, and device states. Based on the objective function of minimizing fault risks, operation costs, and customer waiting times, it generates the optimal resource scheduling plan through an optimization algorithm; The intelligent alarm and resource scheduling module is used for real-time alarm and automatically executes the optimization plan, schedules device resources, and arranges maintenance work.

Images