Intelligent decision-making system for fault maintenance of multi-type loading and unloading equipment of ore wharf

Abstract

The invention relates to the technical field of fault maintenance, and particularly provides an intelligent decision-making system for fault maintenance of multi-type loading and unloading equipment of an ore wharf, which comprises an acquisition and transmission module, a cleaning and extraction module, a fault diagnosis module, a fault pre-control module, a graded early warning module, a resource scheduling module and a system maintenance module. The full-dimensional monitoring of the equipment state can sensitively capture tiny abnormities in the operation of the equipment, find early fault symptoms of bearing wear and hydraulic oil degradation in advance, change the problem of high omission ratio of traditional manual inspection, strive for sufficient processing time for equipment maintenance, enable the equipment maintenance to be changed from dependence on artificial experience to data support, and improve the equipment maintenance efficiency. And the diagnosis efficiency and the prediction accuracy are greatly improved. The problem of conflicts in traditional resource scheduling is effectively solved, reasonable configuration of maintenance resources is achieved, the fault processing period is shortened, the overall efficiency of operation and maintenance of wharf equipment is improved, accurate control over the equipment state is kept, and the method adapts to dynamic changes of operation scenes of multiple types of loading and unloading equipment of an ore wharf.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of fault detection, and in particular to a fault maintenance intelligent decision system for multi-type loading and unloading equipment of an ore terminal. BACKGROUND

[0002] Traditional equipment maintenance of an ore terminal mainly relies on artificial periodic inspection with the aid of simple tools such as visual inspection and handheld temperature measuring instruments. This method cannot collect equipment operation data in real time, and it is difficult to capture early signs of abnormal vibration and temperature rise. Due to the limited frequency of inspection, subtle changes in the equipment operation process are easily overlooked, which can lead to the development of small problems into major failures, causing equipment unplanned downtime and affecting the continuity and efficiency of the terminal loading and unloading operation. In addition, artificial inspection is limited by the working state and professional level of personnel, and the accuracy and completeness of data collection cannot be guaranteed.

[0003] Traditional fault diagnosis highly depends on the experience of maintenance personnel and lacks quantitative analysis models based on data. In the face of complex faults, maintenance personnel are prone to misdiagnosis or missed diagnosis by relying on subjective judgment methods such as listening to abnormal sounds and checking parameter tables. Due to the lack of scientific diagnostic methods and tools, for some faults with strong concealment and involving multiple components, it takes a long time to troubleshoot and is inefficient, which affects the timeliness of equipment maintenance.

[0004] Traditional maintenance resource scheduling relies on artificial records and telephone communication and lacks a dynamic adjustment mechanism. In actual operation, it is difficult to reasonably allocate maintenance personnel, spare parts, and tools according to the emergency level of the fault and the real-time state of the resources, often resulting in resource conflicts or mismatches, causing delays in the maintenance of critical equipment and significant production losses. Spare parts management also lacks flexibility and is often based on historical experience to stock up, which can easily cause inventory accumulation or emergency shortages, increasing storage costs and affecting maintenance progress. Artificial scheduling cannot quickly respond to unexpected situations and cannot meet the needs of efficient port operations.

[0005] The traditional system lacks an automatic updating mechanism, and the adjustment of fault diagnosis thresholds and maintenance strategy parameters relies on manual operation. When the equipment is upgraded, the working conditions change, or new equipment is introduced, the model cannot be updated in a timely manner, increasing the risk of false positives or false negatives. Manual adjustment not only consumes time and effort but also is prone to errors due to human negligence or lack of professional knowledge, making it difficult to adapt to dynamic changes in equipment operating conditions. The long-term use of outdated models and strategies cannot effectively deal with new fault patterns, reducing the effectiveness and reliability of equipment maintenance SUMMARY

[0006] To achieve the above-mentioned purposes, the present application adopts the following technical solutions:

[0007] In one aspect of the present application, a fault maintenance intelligent decision system for multi-type loading and unloading equipment of an ore terminal is provided, comprising:

[0008] A multi-dimensional data sensing unit is configured to distribute key operating parts of various types of loading and unloading equipment at an ore terminal, collect multi-dimensional physical quantity data reflecting the operating status of the equipment, and upload the multi-dimensional physical quantity data after preliminary processing.

[0009] The data preprocessing and feature engineering unit is communicatively connected to the multidimensional data sensing unit and is configured to receive the multidimensional physical quantity data, perform cleaning, noise reduction and standardization on it, and adaptively extract a set of key feature parameters that can characterize the health status of the equipment through a preset feature extraction algorithm.

[0010] The fault diagnosis and location unit is communicatively connected to the data preprocessing and feature engineering unit and is configured to intelligently identify the current fault type of the equipment and accurately locate the specific component or subsystem where the fault occurs based on the key feature parameter set and using a hybrid diagnostic model that integrates deep learning and knowledge graphs.

[0011] The fault trend prediction and intelligent early warning unit is communicatively connected to the fault diagnosis and location unit and the data preprocessing and feature engineering unit. It is configured to integrate historical equipment operation data, real-time key feature parameters, environmental factors and maintenance records to construct a multi-factor coupled fault evolution prediction model to predict the development trend of a specific fault mode and the probability of occurrence within a preset time period in the future. It also dynamically classifies the early warning level according to the probability of occurrence and the potential impact of the fault on production and operation.

[0012] The dynamic resource scheduling and maintenance decision-making unit is communicatively connected to the fault trend prediction and intelligent early warning unit and the fault diagnosis and location unit. It is configured to generate the optimal maintenance plan suggestion through a multi-objective optimization algorithm based on the fault diagnosis results, predicted trends, early warning levels, and real-time status and geographical distribution information of maintenance resources, and automatically trigger or assist in scheduling maintenance resources to realize dynamic dispatching and resource coordination of maintenance tasks.

[0013] The system management and maintenance center communicates with each of the above-mentioned units and is configured to provide a unified user interface, access control, system parameter configuration, operation log recording and auditing, and support iterative updates of models and algorithms as well as expansion of system functions.

[0014] In one optional implementation, the multi-dimensional physical quantity data includes:

[0015] Vibration acceleration, velocity, and displacement signals for rotating components such as gearboxes and bearings in large equipment such as ship unloaders and stacker-reclaimer bucket wheel excavators;

[0016] Temperature field distribution data for the stator, rotor, and bearing housing of the drive motor of a ship unloader and a stacker-reclaimer bucket wheel excavator;

[0017] Data on hydraulic pipeline pressure, flow rate, and hydraulic oil temperature for hydraulically driven equipment such as hydraulic loading arms and pitching mechanisms;

[0018] Data on viscosity, moisture content, particle size, and ferrography of gearbox lubricating oil for gear transmission equipment.

[0019] In one optional implementation, the data preprocessing and feature engineering unit is further configured to perform stationarity testing and processing on time-series data, and to perform adaptive noise filtering; to dynamically calibrate feature parameters in conjunction with equipment operating information; and to use unsupervised learning methods to mine abnormal pattern features hidden in the data.

[0020] In one optional implementation, the method for extracting feature parameters reflecting the equipment's operating status is as follows: The root mean square value of the vibration frequency feature is extracted, which can be expressed as:

[0021]

[0022] Where RMS is the root mean square value, xi is the i-th vibration sample value, and N is the number of sampling points, which reflects the magnitude of vibration energy. The larger the value, the more severe the equipment wear.

[0023] Temperature change features can be used to extract the rate of temperature change, which can be expressed as:

[0024]

[0025] Where V is the rate of temperature change, T(t) is the current temperature value, T(t-Δt) is the temperature value Δt before time Δt, and Δt is the time interval;

[0026] Pressure fluctuation coefficient can be extracted from pressure numerical characteristics using the following expression:

[0027]

[0028] Where Pmax is the maximum pressure and Pmin is the minimum pressure. This represents the average pressure.

[0029] The viscosity change rate, extracted from the state characteristics of oil, can be expressed by the following expression:

[0030]

[0031] Where ΔU is the viscosity change rate, U(t) is the current viscosity value, and U0 is the new oil reference viscosity value.

[0032] In one optional implementation, the method for diagnosing equipment faults is as follows: A fault diagnosis model is constructed, and an equipment health index is calculated, which can be expressed by the following expression:

[0033]

[0034] Where HI is the equipment health index, Fi is the i-th feature value, Fi,min and Fi,max are the boundary values ​​of the normal range of feature parameters, and Wi is the weight of each feature;

[0035] The device health index is set as follows: within the ab range is normal; within the bc range, attention is needed; within the cd range, it is abnormal; and above the d range, it is a fault.

[0036] In one optional implementation, the fault evolution prediction model in the fault trend prediction and intelligent early warning unit is configured as follows:

[0037] Deep learning prediction models based on attention mechanisms can automatically focus on the key features that have the greatest impact on the development of faults.

[0038] The dynamic early warning level classification includes at least: advisory early warning, general early warning, important early warning and emergency early warning, and pre-set emergency response suggestions are associated with different early warning levels;

[0039] The dynamic resource scheduling and maintenance decision-making unit is configured for:

[0040] Establish a maintenance resource capability matrix and update maintenance personnel skill proficiency, spare parts inventory turnover rate, and tool availability status in real time;

[0041] When generating the optimal maintenance plan, the urgency of the fault, maintenance cost, impact on production plan, and balanced utilization of resources are comprehensively considered.

[0042] In one optional implementation, the method for predicting failure trends and probabilities is as follows: Single-device failure prediction uses an LSTM-GARCH model combined with a long short-term memory network and a volatility model to capture the nonlinear trend of equipment performance degradation; multi-device correlation prediction uses the equipment topology as a graph, inputs the state characteristics of each device, predicts the failure propagation path, and constructs the equipment failure probability dependency relationship; probability prediction uses a Cox proportional hazards model to estimate the failure probability of the equipment in the future time t, which can be expressed by the following expression:

[0043] h(t,X)=h0(t)e βX

[0044] Where X is the feature vector, and β is the weight of each feature in the feature vector.

[0045] In one optional implementation, the warning level is divided into different levels: a fault probability less than P1 is low probability, a fault probability greater than P1 but less than P2 is medium probability, and a fault probability greater than P2 is high probability; the degree of impact is divided into mild impact, moderate impact, and moderate impact according to preset evaluation rules; a three-dimensional warning level matrix is ​​constructed, dividing the probability and degree of impact into three levels, forming nine risk combinations, corresponding to a four-level warning system of red, orange, yellow, and blue.

[0046] In another aspect, the present invention provides an electronic device comprising:

[0047] At least one memory stores computer-executable instructions non-transitory;

[0048] At least one processor, configured to run the computer-executable instructions,

[0049] The computer-executable instructions are executed by the processor according to the aforementioned intelligent decision-making system for fault repair of various types of loading and unloading equipment at ore terminals.

[0050] In another aspect, the present invention provides a computer-readable storage medium storing computer-executable instructions, which, when executed by at least one processor, implement the intelligent decision-making system for fault repair of various types of loading and unloading equipment at an ore terminal as described above.

[0051] The positive effects of this invention are as follows: 1. By deploying sensors in key components of ship unloaders and stacker-reclaimer equipment, including gearboxes and bearings, this invention collects real-time operational data such as vibration frequency, temperature changes, pressure values, and oil status, achieving comprehensive monitoring of equipment status. This method can keenly detect subtle anomalies in equipment operation, identify early signs of bearing wear and hydraulic oil deterioration, and overcome the problem of high missed detection rates in traditional manual inspections, thus allowing sufficient time for equipment maintenance.

[0052] 2. This invention utilizes a health index model and fault tree analysis technology to deeply analyze extracted equipment characteristic parameters, constructing a data-driven fault diagnosis and prediction system. This mechanism can quickly locate equipment fault locations, clarify the logical relationships between mechanical and electrical faults, and combine long short-term memory network algorithms to predict fault development trends, shifting equipment maintenance from reliance on manual experience to data support, significantly improving diagnostic efficiency and prediction accuracy.

[0053] 3. This invention constructs a tiered early warning system based on the probability and severity of failures, classifying early warnings into different levels to achieve differentiated responses to failures. Combined with the real-time status of maintenance resources, an intelligent algorithm automatically generates optimized solutions including maintenance methods and resource allocation. This model effectively solves the conflict problems in traditional resource scheduling, achieves rational allocation of maintenance resources, shortens the failure handling cycle, and improves the overall efficiency of terminal equipment operation and maintenance.

[0054] 4. This invention utilizes an online learning mechanism and a distributed training framework, enabling the system to update model parameters in real time based on new fault data and equipment operating conditions. When equipment is upgraded or operating conditions change, the model can automatically complete optimization iterations, ensuring the accuracy of fault diagnosis and prediction. This allows the system to maintain a precise grasp of equipment status and continuously adapt to the dynamic changes in the operating scenarios of ore terminal equipment. Attached Figure Description

[0055] The accompanying drawings are provided to further illustrate the invention and form part of the specification. They are used in conjunction with embodiments of the invention to explain the invention and do not constitute a limitation thereof. In the drawings:

[0056] Figure 1 This is a system flowchart provided in Embodiment 1 of the present invention;

[0057] Figure 2 This is a flowchart illustrating the acquisition of multi-dimensional physical quantity data provided in Embodiment 2 of the present invention;

[0058] Figure 3 This is a block diagram of the electronic device provided in Embodiment 3 of the present invention;

[0059] Figure 4 This is a block diagram of a computer-readable storage medium provided in Embodiment 4 of the present invention. Detailed Implementation

[0060] The technical solutions of the present invention will now be described with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0061] Hereinafter, the terms "first," "second," etc., are used for descriptive convenience only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Thus, a feature defined with "first," "second," etc., may explicitly or implicitly include one or more of that feature. In the description of this invention, unless otherwise stated, "a plurality of" means two or more.

[0062] In this invention, unless otherwise explicitly specified and limited, the term "connection" should be interpreted broadly. For example, "connection" can be a fixed mechanical connection, a detachable mechanical connection, or an integral part; or, "connection" can be a direct connection or an indirect connection through an intermediate medium. Furthermore, unless otherwise explicitly specified and limited, the term "coupling" should be interpreted broadly. For example, "coupling" can be a direct electrical connection, such as physical contact and electrical conduction between two components; it can also be understood as an electrical connection between different components in a circuit structure through physical lines capable of transmitting electrical signals, such as copper foil or wires on a printed circuit board (PCB), to transmit electrical signals; or, "coupling" can be an indirect electrical connection between two components through an intermediate medium; or, "coupling" can be an electrical connection between two components in a non-contact manner, such as an electrical connection between two components using capacitive coupling to transmit electrical signals.

[0063] In this embodiment of the invention, directional terms such as "up," "down," "left," and "right" may be defined relative to the orientation of the components shown in the accompanying drawings. It should be understood that these directional terms can be relative concepts, used for relative description and clarification, and can change accordingly depending on the orientation of the components in the accompanying drawings.

[0064] The embodiments of the present invention can be used for

[0065] Key technical features of the embodiments of the present invention are described below:

[0066] Example 1:

[0067] like Figure 1 As shown, this embodiment of the invention provides an intelligent decision-making system for fault repair of various types of loading and unloading equipment at an ore terminal, characterized in that it includes:

[0068] A multi-dimensional data sensing unit is configured to distribute key operating parts of various types of loading and unloading equipment at an ore terminal, collect multi-dimensional physical quantity data reflecting the operating status of the equipment, and upload the multi-dimensional physical quantity data after preliminary processing.

[0069] The data preprocessing and feature engineering unit is communicatively connected to the multidimensional data sensing unit and is configured to receive the multidimensional physical quantity data, perform cleaning, noise reduction and standardization on it, and adaptively extract a set of key feature parameters that can characterize the health status of the equipment through a preset feature extraction algorithm.

[0070] The fault diagnosis and location unit is communicatively connected to the data preprocessing and feature engineering unit and is configured to intelligently identify the current fault type of the equipment and accurately locate the specific component or subsystem where the fault occurs based on the key feature parameter set and using a hybrid diagnostic model that integrates deep learning and knowledge graphs.

[0071] The fault trend prediction and intelligent early warning unit is communicatively connected to the fault diagnosis and location unit and the data preprocessing and feature engineering unit. It is configured to integrate historical equipment operation data, real-time key feature parameters, environmental factors and maintenance records to construct a multi-factor coupled fault evolution prediction model to predict the development trend of a specific fault mode and the probability of occurrence within a preset time period in the future. It also dynamically classifies the early warning level according to the probability of occurrence and the potential impact of the fault on production and operation.

[0072] The dynamic resource scheduling and maintenance decision-making unit is communicatively connected to the fault trend prediction and intelligent early warning unit and the fault diagnosis and location unit. It is configured to generate the optimal maintenance plan suggestion through a multi-objective optimization algorithm based on the fault diagnosis results, predicted trends, early warning levels, and real-time status and geographical distribution information of maintenance resources, and automatically trigger or assist in scheduling maintenance resources to realize dynamic dispatching and resource coordination of maintenance tasks.

[0073] The system management and maintenance center communicates with each of the above-mentioned units and is configured to provide a unified user interface, access control, system parameter configuration, operation log recording and auditing, and support iterative updates of models and algorithms as well as expansion of system functions.

[0074] In the above embodiments, the data acquisition and transmission module is used to collect raw data reflecting the operating status of various types of loading and unloading equipment (including but not limited to ship unloaders, stacker-reclaimer bucket elevators, belt conveyors, and gantry cranes) deployed at the ore terminal, and upload the collected data to the data processing center.

[0075] The specific data collection includes: collecting vibration signals from the housings or shaft ends of rotating components such as gearboxes and bearings in ship unloaders, stacker-reclaimer bucket excavators, and gantry cranes using installed vibration sensors. These vibration signals include vibration frequency, vibration acceleration, and vibration velocity. Temperature change data is collected from the stator surface of motors, the outer ring of bearing housings, and the outer walls of hydraulic pumps and cylinders in hydraulic drive equipment, all of which are used in ship unloaders, stacker-reclaimer bucket excavators, and gantry cranes. Real-time pressure values ​​are collected from the hydraulic oil inlet and return pipelines and key hydraulic valve block outlets in hydraulic drive equipment (such as the pitching hydraulic system of a ship unloader and the luffing hydraulic system of a bucket excavator). Oil state parameters, including oil viscosity, moisture content, particulate contamination, and ferromagnetic abrasive particle concentration, are collected from the oil sump or circulation pipelines of the lubricating oil system in gear transmission equipment (such as the gearbox of the ship unloader's traveling mechanism and the slewing gearbox of a stacker-reclaimer bucket excavator).

[0076] Data transmission method: Industrial Ethernet, 5G / 4G wireless communication or IoT communication technologies such as LoRa are used to upload the collected raw data to the data processing center in real time or near real time.

[0077] Data Cleaning and Feature Extraction Module: This module connects to the data acquisition and transmission module and receives the raw data, performing preprocessing and feature engineering operations on it. Data Cleaning: This includes, but is not limited to, removing noisy data, filling in missing values, smoothing abnormal fluctuations, and standardizing or normalizing data to ensure data quality. Feature Extraction: For the cleaned data, feature parameters that effectively reflect the equipment's operating status and potential fault information are extracted. For example: time-domain analysis (extracting peak value, kurtosis, root mean square, etc.) and frequency-domain analysis (extracting characteristic frequencies, spectral energy distribution, etc.) of vibration signals; temperature data extracting temperature rise rate, temperature gradient, and overheating duration; pressure data extracting pressure fluctuation range, pressure peak value, and pressure settling time; and oil state parameters extracting their rate of change and deviation from historical baselines.

[0078] Intelligent Fault Diagnosis Module: This module is connected to the data cleaning and feature extraction module. It receives the feature parameters and constructs a fault diagnosis model that integrates multi-source information to achieve accurate identification and location of equipment faults. Diagnostic Model Construction: Employs models including but not limited to deep learning models (such as Convolutional Neural Networks (CNN), Recurrent Neural Networks (LSTM), Support Vector Machines (SVM), Random Forests, or ensemble learning models. Model training data comes from historical fault case data, simulated fault experimental data, and normal operation data. Fault Diagnosis and Location: The real-time extracted feature parameters are input into the trained fault diagnosis model. The model outputs the fault type (e.g., bearing outer ring spalling, gear tooth surface wear, hydraulic valve jamming, motor inter-turn short circuit, etc.), fault severity, and specific fault location (e.g., the oil inlet pipe of the A chamber of the tilting hydraulic cylinder of the unloader's boom, or the bearing on the central column of the stacker-reclaimer turbine).

[0079] Fault Trend Prediction and Probability Assessment Module: This module connects to both the intelligent fault diagnosis module and the data cleaning and feature extraction module. It combines diagnosed fault information, real-time and historical feature parameters, equipment runtime, maintenance records, and other multi-source information to establish a fault evolution trend prediction algorithm model. Algorithm Model: Employs models including, but not limited to, physical degradation models, time series analysis models (such as ARIMA, Prophet), or machine learning-based remaining useful life prediction models (such as LSTM, GRU). Prediction Output: Predicts the development trend curve of current potential faults or early minor faults, the probability value of fault occurrence within a preset time period (such as 24 hours, 7 days), and an assessment of the remaining useful life (RUL) of key equipment components.

[0080] Dynamic Hierarchical Early Warning Module: This module connects to the fault trend prediction and probability assessment module. It constructs a dynamic early warning model based on the predicted probability of a fault, the impact of the fault on terminal production (e.g., downtime, maintenance costs, safety risk level), and the urgency of current terminal production tasks. Early Warning Level Classification: Early warning levels are divided into at least three levels, for example: Level 1 (Minor Warning): Low probability of fault occurrence or minor impact, prompting increased monitoring; Level 2 (Moderate Warning): Medium probability of fault occurrence or moderate impact, prompting the development of a maintenance plan and preparation of maintenance resources; Level 3 (Severe Warning): High probability of fault occurrence or severe impact, prompting immediate shutdown for maintenance or implementation of emergency protective measures. Early Warning Strategy Generation: For different early warning levels, fault prevention and control strategies, including inspection suggestions, maintenance priorities, and temporary operating parameter adjustment suggestions, are automatically generated.

[0081] Maintenance Resource Optimization and Scheduling Module: This module connects to the intelligent fault diagnosis module, the dynamic hierarchical early warning module, and the dock resource management database. Based on fault diagnosis results (fault type, location, severity), dynamic hierarchical early warning information (early warning level, suggested maintenance time window), and the real-time status of maintenance resources (including maintenance personnel skill matrix, available spare parts inventory quantity and location, maintenance tools, and emergency repair vehicle status), it provides optimal maintenance plan suggestions and automatically optimizes maintenance resource scheduling. Maintenance Suggestion Generation: This includes recommended maintenance methods (such as reactive maintenance, preventative maintenance, and predictive maintenance), a bill of materials (BOM) for maintenance, approximate maintenance man-hours, and safety precautions. Resource Scheduling Optimization: Based on intelligent optimization methods such as genetic algorithms, ant colony optimization, or particle swarm optimization, it automatically generates maintenance personnel dispatch plans, spare parts requisition and delivery route planning, and maintenance tool allocation plans with the goals of minimizing maintenance costs, minimizing downtime, or maximizing resource utilization.

[0082] System Management and Self-Iteration Module: This module serves as the foundational support for the system, providing comprehensive management and continuous optimization of the entire intelligent decision-making system. User Management and Permission Assignment: Provides user account management for different roles (e.g., system administrator, equipment engineer, maintenance dispatcher, frontline operator), and assigns corresponding operation and data access permissions to each role. System Configuration and Parameter Management: Allows administrators to configure system parameters such as sensor sampling frequency, data transmission protocol, diagnostic model parameter thresholds, early warning level judgment rules, and resource scheduling constraints. Log Recording and Auditing: Provides detailed recording of system operation logs, user operation logs, fault diagnosis logs, early warning event logs, and maintenance dispatch logs for traceability and auditing.

[0083] The system features model self-iteration and knowledge updates. Periodically, or when triggered by specific conditions (such as accumulating a certain number of new fault cases), it retrains and optimizes the fault diagnosis and trend prediction models using new operational and maintenance feedback data. This enables the models to learn and iterate independently, continuously improving diagnostic and prediction accuracy. Simultaneously, it constructs an equipment fault maintenance knowledge graph and automatically updates the fault mode and maintenance strategy libraries.

[0084] The present invention has the following beneficial effects:

[0085] Comprehensive perception and accurate diagnosis: By deploying multiple sensors in key parts of various types of loading and unloading equipment, comprehensive perception of the equipment's operating status is achieved; combined with advanced machine learning algorithms to build a diagnostic model, the fault type can be accurately identified and the fault location can be located, overcoming the limitations of traditional manual diagnosis.

[0086] Early warning and trend prediction: It can not only diagnose existing faults, but also predict and assess the development trend and probability of potential faults, and provide dynamic graded early warning, providing a sufficient time window for proactive maintenance.

[0087] Intelligent decision-making and resource optimization: The system can automatically generate maintenance suggestions and optimize resource scheduling based on fault conditions, warning levels and resource status, thereby improving maintenance response speed and resource utilization efficiency and reducing operation and maintenance costs.

[0088] System self-iteration and knowledge accumulation: It has the ability to self-iterate models and update knowledge. As the system runs for a long time, its diagnostic and prediction accuracy and decision rationality can be continuously improved, and valuable equipment maintenance knowledge can be accumulated.

[0089] Ensuring production and improving efficiency: By reducing unplanned downtime and increasing the mean time between failures (MTBF) of equipment, the service life of equipment is extended, thereby ensuring continuous and stable production at the ore terminal and significantly improving economic efficiency and management level.

[0090] Example 2:

[0091] like Figure 2 As shown, based on Example 1, the multi-dimensional physical quantity data provided in this embodiment of the invention includes:

[0092] Vibration acceleration, velocity, and displacement signals for rotating components such as gearboxes and bearings in large equipment such as ship unloaders and stacker-reclaimer bucket wheel excavators;

[0093] Temperature field distribution data for the stator, rotor, and bearing housing of the drive motor of a ship unloader and a stacker-reclaimer bucket wheel excavator;

[0094] Data on hydraulic pipeline pressure, flow rate, and hydraulic oil temperature for hydraulically driven equipment such as hydraulic loading arms and pitching mechanisms;

[0095] Data on viscosity, moisture content, particle size, and ferrography of gearbox lubricating oil for gear transmission equipment.

[0096] Specifically, the data preprocessing and feature engineering unit is also configured to perform stationarity testing and processing on time-series data, and to perform adaptive noise filtering; to dynamically calibrate feature parameters in conjunction with equipment operating information; and to use unsupervised learning methods to mine abnormal pattern features hidden in the data.

[0097] Specifically, the method for extracting feature parameters that reflect the operating status of the equipment is as follows: the root mean square value of the vibration frequency feature is extracted, which can be expressed as:

[0098] Where RMS is the root mean square value, x iLet be the i-th vibration sample value, and N be the number of sampling points. The value reflects the magnitude of the vibration energy; the larger the value, the more severe the equipment wear.

[0099] Temperature change features can be used to extract the rate of temperature change, which can be expressed as:

[0100]

[0101] Where V is the rate of temperature change, T(t) is the current temperature value, T(t-Δt) is the temperature value Δt before time Δt, and Δt is the time interval;

[0102] Pressure fluctuation coefficient can be extracted from pressure numerical characteristics using the following expression:

[0103]

[0104] Where Pmax is the maximum pressure, Pmin is the minimum pressure, and P is the average pressure;

[0105] The viscosity change rate, extracted from the state characteristics of oil, can be expressed by the following expression:

[0106]

[0107] Where ΔU is the viscosity change rate, U(t) is the current viscosity value, and U0 is the new oil reference viscosity value.

[0108] Specifically, the method for diagnosing equipment faults is as follows: A fault diagnosis model is constructed, and the equipment health index is calculated, which can be expressed by the following expression:

[0109]

[0110] Where HI is the equipment health index, Fi is the i-th feature value, Fi,min and Fi,max are the boundary values ​​of the normal range of feature parameters, and Wi is the weight of each feature;

[0111] The device health index is set as follows: within the ab range is normal; within the bc range, attention is needed; within the cd range, it is abnormal; and above the d range, it is a fault.

[0112] Specifically, the fault evolution prediction model in the fault trend prediction and intelligent early warning unit is configured as follows:

[0113] Deep learning prediction models based on attention mechanisms can automatically focus on the key features that have the greatest impact on the development of faults.

[0114] The dynamic early warning level classification includes at least: advisory early warning, general early warning, important early warning and emergency early warning, and pre-set emergency response suggestions are associated with different early warning levels;

[0115] The dynamic resource scheduling and maintenance decision-making unit is configured for:

[0116] Establish a maintenance resource capability matrix and update maintenance personnel skill proficiency, spare parts inventory turnover rate, and tool availability status in real time;

[0117] When generating the optimal maintenance plan, the urgency of the fault, maintenance cost, impact on production plan, and balanced utilization of resources are comprehensively considered.

[0118] Specifically, the methods for predicting failure trends and probabilities are as follows: Single-device failure prediction uses an LSTM-GARCH model combined with a long short-term memory network and a volatility model to capture the nonlinear trend of equipment performance degradation; multi-device correlation prediction uses the equipment topology as a graph, inputs the state characteristics of each device, predicts the failure propagation path, and constructs the equipment failure probability dependency relationship; probability prediction uses a Cox proportional hazards model to estimate the failure probability of the equipment in the future time t, which can be expressed by the following expression:

[0119] h(t,X)=h0(t)e βX

[0120] Where X is the feature vector, and β is the weight of each feature in the feature vector.

[0121] Specifically, the warning levels are divided into different levels: a fault probability less than P1 is low probability, a fault probability greater than P1 but less than P2 is medium probability, and a fault probability greater than P2 is high probability; the degree of impact is divided into mild impact, moderate impact, and moderate impact according to preset evaluation rules; a three-dimensional warning level matrix is ​​constructed, dividing probability and degree of impact into three levels, forming nine risk combinations, corresponding to a four-level warning system of red, orange, yellow, and blue.

[0122] In the above embodiments, in a more specific application of the present invention, in the data collection of key parts of various types of loading and unloading equipment at ore terminals, a technical architecture of layered deployment, precise perception, and interference-resistant transmission is adopted, and industrial-grade sensing equipment and IoT transmission solutions are combined to achieve full-dimensional data collection.

[0123] For rotating components of ship unloaders and stacker-reclaimer bucket wheel excavators, including gearboxes and bearings, triaxial accelerometers such as the IEPE type are used to collect vibration frequencies. The sensors are magnetically or bolted to the surface of the component housing, and the sampling frequency is set to 10-20kHz to capture high-frequency fault characteristics such as bearing spalling. To adapt to the high dust and humidity environment of ports, the sensors have an IP67 protection rating, and the cables are laid using metal-shielded corrugated conduits. The grounding terminal is protected against electromagnetic interference through a surge protector.

[0124] Temperature acquisition covers the motor surface and bearing area. An infrared temperature sensor monitors the motor housing temperature non-contactly; bearing temperature is measured directly using an embedded PT100 resistance temperature detector (RTD). During installation, a temperature measurement hole is pre-drilled in the bearing housing and filled with thermally conductive silicone grease to ensure efficient heat transfer. Both types of sensors integrate temperature compensation circuits to eliminate the influence of environmental temperature drift. The sampling period is set to 1-5 seconds to accommodate slow temperature changes.

[0125] The hydraulic system pressure acquisition uses a piezoresistive pressure transmitter. Pressure taps are welded to the high-pressure side of the hydraulic oil lines, such as at the pump outlet or before the control valve. The transmitter has a built-in shock-resistant diaphragm that can withstand an overload of 1.5 times the rated pressure. Signal transmission uses a 4-20mA current loop, paired with twisted-pair shielded cable, achieving a transmission distance of over 100 meters, meeting the wiring requirements of large dock equipment.

[0126] The gearbox lubricating oil condition monitoring system integrates an online viscosity sensor and a particle counter. The sensor assembly is installed in the bypass section of the main lubricating oil return line. The viscosity sensor uses a vibration principle to monitor changes in oil viscosity in real time; the particle counter detects oil contamination levels using a laser light-blocking method. Both types of equipment are equipped with automatic cleaning valves to periodically remove impurities adhering to the sensor probes, ensuring data accuracy.

[0127] All sensors aggregate data through an industrial-grade edge gateway. The gateway has a built-in real-time operating system that supports Modbus and CANopen industrial protocols. After filtering, denoising, and protocol conversion of the collected data, it is uploaded to the data processing center via a fiber optic ring network or a 5G private network, controlling transmission latency. The system also features breakpoint resume capability, locally caching 72 hours of data in the event of network failure to ensure data integrity.

[0128] The core technical methods for data cleaning are as follows: For vibration signals, a Butterworth low-pass filter is used to eliminate high-frequency electromagnetic interference; for hydraulic pressure signals, a moving average filter is used to smooth pulse noise. For occasional data loss from temperature sensors, Kalman filtering combined with the equipment's thermal inertia model is used for interpolation and completion. Heterogeneous data collected from different sensors are normalized to the [-1,1] interval using Z-score to eliminate dimensional influences.

[0129] A dynamic threshold is set for the vibration RMS value. When the value exceeds the threshold, the Isolation Forest algorithm is triggered to further determine whether it is a true anomaly. The current value is predicted using an autoregressive integral moving average model. If the deviation between the measured value and the predicted value is greater than a preset range and continues for 3 sampling periods, it is marked as an anomaly and replaced with a historical trend fitted value. When the bearing temperature rises abnormally, the spectral changes of adjacent vibration signals are checked simultaneously. If the two are not correlated, the temperature sensor is determined to be faulty, triggering a hardware self-test process.

[0130] When locating the fault location, fault tree analysis is used to systematically decompose the causal relationships of equipment failures. First, the event with the greatest impact on production is selected as the top event, and boundary conditions are defined. Then, the causes of the failure are decomposed layer by layer from both mechanical transmission and electrical control dimensions, constructing a tree structure using AND and OR gates. For example, the top event of a motor not turning can be decomposed into two OR gate-based events: power supply failure and mechanical jamming. The latter is further decomposed into bearing seizure and gear tooth breakage events.

[0131] An algorithm model is established by combining multi-source information to predict fault trends and probabilities. The multi-source information includes: equipment status data such as vibration frequency RMS value, temperature change rate, and pressure fluctuation coefficient real-time monitoring parameters; environmental parameters such as port humidity, salt spray concentration, and ore moisture; historical maintenance data such as fault type, maintenance time, and replacement spare parts life; and process parameters such as loading and unloading volume and material characteristics.

[0132] Single-device failure prediction uses an LSTM-GARCH model combined with a long short-term memory network and a volatility model to capture the nonlinear trend of equipment performance degradation. Input the vibration RMS value and temperature change rate of the past 72 hours, and output the failure probability curve for the next 24-72 hours.

[0133] Multi-device association prediction uses the device topology as a graph, inputs the state characteristics of each device, predicts the fault propagation path, and constructs the probability dependency relationship of device faults.

[0134] The above methods can overcome the limitations of traditional single-parameter prediction, integrate multi-dimensional information such as equipment status, environment, and process; not only predict the probability of failure, but also output the uncertainty range; reduce unplanned downtime, lower preventive maintenance costs, and improve spare parts inventory turnover.

[0135] Based on the probability of failure and the degree of impact, the warning level is divided into different grades: a failure probability less than P1 is low probability, greater than P1 but less than P2 is medium probability, and greater than P2 is high probability. The dimensions for assessing the degree of impact are: production impact includes downtime and throughput loss; economic impact includes maintenance costs and delay compensation; and safety impact includes accident risk and environmental hazards. According to the preset assessment rules, it is divided into mild impact, moderate impact, and medium impact.

[0136] A three-dimensional early warning level matrix is ​​constructed, dividing probability and impact into three levels, which in turn form nine risk combinations, corresponding to a four-level early warning system of red, orange, yellow, and blue, as shown below:

[0137] A blue alert indicates a low probability of mild impact, a yellow alert indicates a low probability of moderate impact, and an orange alert indicates a low probability of severe impact. A yellow alert indicates a medium probability of mild impact, an orange alert indicates a medium probability of moderate impact, and a red alert indicates a medium probability of severe impact. An orange alert indicates a high probability of mild impact, a red alert indicates a high probability of moderate impact, and a red alert indicates a high probability of severe impact.

[0138] The blue alert employs a standard monitoring strategy, maintaining normal equipment operation and recording status data for trend analysis. The response measures include maintaining regular inspection frequency, focusing on sensor data fluctuations, not adjusting maintenance resource allocation, and stockpiling spare parts according to safety stock standards.

[0139] A yellow alert adopts a level-one pre-control strategy to suppress the development of the fault and prepare for handling in advance. Handling measures include increasing monitoring frequency, activating vibration or temperature trend analysis models, deploying a basic maintenance team to stand by, checking relevant spare parts inventory, and developing a preliminary maintenance plan.

[0140] Orange alerts employ a proactive maintenance strategy, intervening before malfunctions occur to avoid unplanned downtime. The measures include convening a special meeting to confirm the repair plan; deploying a professional repair team, requisitioning spare parts, and preparing specialized tools; and conducting preventative maintenance during off-peak hours in conjunction with the production plan.

[0141] A red alert triggers an emergency intervention strategy, namely, eliminating potential faults and ensuring production continuity and safety. The measures include rapidly activating backup equipment and forcibly shutting down the main equipment; forming a cross-departmental emergency repair team with all spare parts on site; implementing a closed-loop process of repair, verification, and post-incident review; conducting a full-load trial run after repair; and submitting a root cause analysis report.

[0142] When multiple red alerts are triggered simultaneously, they are prioritized as follows: faults threatening personnel safety or the environment; faults affecting key equipment on the main production line; and faults with long repair times and no backup equipment. After each pre-control strategy is executed, the system records whether the actual fault occurred, the deviation between repair costs and the plan, and the percentage reduction in production losses.

[0143] Through the aforementioned tiered early warning system, ore terminals can transform failure risks into quantifiable and manageable tiered strategies, achieving an upgrade from passive response to proactive prevention, and significantly improving equipment reliability and production efficiency.

[0144] This system summarizes four types of information: fault diagnosis results, fault prevention and control strategies, tiered early warning information, and real-time status of maintenance resources. Fault diagnosis clarifies the problem and its cause; standard handling procedures are matched according to the prevention and control strategies; the urgency of handling is determined by tiered early warnings; and the availability of personnel, spare parts, and tools is monitored through real-time resource status. This information is then comprehensively analyzed, taking into account fault urgency and resource availability, to generate structured recommendations that include maintenance methods, required resources, time planning, and risk warnings. The content covers maintenance steps, resource lists, time planning, and risk warnings.

[0145] In the resource scheduling phase, the system automatically detects resource conflicts, such as two red alerts vying for the same hydraulic wrench, using a mixed-integer programming model. This triggers a genetic algorithm to redistribute resources: prioritizing faults with significant production impact, utilizing backup resources, and optimizing task order. When new alerts or resource status changes occur during maintenance, edge computing nodes update the scheduling plan in real time and push it to the execution end via the 5G network. The dynamic adjustments include: maintenance team route planning, emergency spare parts allocation procedures, and tool allocation priorities, ensuring that resource configuration always maintains a state of minimum downtime loss and maximum utilization.

[0146] The intelligent decision-making system adopts a layered architecture to achieve comprehensive management and optimization. User management is based on a hybrid RBAC and ABAC model, using OAuth2.0 for multi-factor authentication and enterprise-wide account lifecycle management. Permission allocation dynamically grants menu access, data read / write, and model operation permissions based on user roles and data sensitivity. System configuration utilizes the Apollo configuration center, supporting static configuration of basic parameters and dynamic adjustment of business parameters, automatically triggering canary releases and version rollbacks upon changes. Log recording uses ELKStack to collect operation logs, runtime logs, and model training logs, enabling user behavior auditing, service anomaly tracking, and model performance analysis.

[0147] Model parameter updates rely on Flink or Spark Streaming frameworks for real-time processing of newly incoming device data. Incremental parameter updates are triggered by online learning algorithms, and version management and performance comparison are achieved using MLflow. Algorithm optimization employs a distributed training framework based on Kubernetes cluster resource scheduling, continuously iterating through an automated pipeline of data preprocessing, model selection, hyperparameter optimization, and performance evaluation. The system automatically retrains based on the latest data periodically or when model evaluation metrics fall below thresholds, using RayTune for Bayesian hyperparameter optimization to ensure continuous improvement in model decision-making capabilities.

[0148] Example 3

[0149] Figure 3 A block diagram of an exemplary electronic device suitable for implementing embodiments of the present invention is shown.

[0150] The electronic device may include a central processing unit / microprocessor / main control chip, etc. 4; and a storage medium 5, coupled to the central processing unit / microprocessor / main control chip, etc. 4, and storing computer-executable instructions therein for performing the steps of various methods of embodiments of the present invention when executed by the processor.

[0151] The central processing unit / microprocessor / main control chip, etc., can include, but are not limited to, one or more processors or microprocessors.

[0152] Storage medium 5 may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, computer storage media (e.g., hard disk, floppy disk, solid-state drive, removable disk, CD-ROM, DVD-ROM, Blu-ray disc, etc.).

[0153] In addition, the electronic device may also include (but is not limited to) a data bus 6, an input / output bus / external bus / device bus 7, a display 8, and input / output devices 9 (e.g., keyboard, mouse, speaker, etc.).

[0154] The central processing unit / microprocessor / main control chip, etc. 4 can communicate with external devices (8, 9, etc.) via I / O bus 7 through wired or wireless network (not shown).

[0155] The storage medium 5 may also store at least one computer-executable instruction for performing the steps of various functions and / or methods in the embodiments described herein when the central processing unit / microprocessor / main control chip, etc., 4 is running.

[0156] In one embodiment, the at least one computer-executable instruction may also be compiled into or comprise a software product, wherein one or more computer-executable instructions are executed by a processor to perform the steps of the various functions and / or methods in the embodiments described herein.

[0157] Example 4

[0158] Figure 4 A schematic diagram of a computer-readable storage medium according to an embodiment of the present invention is shown.

[0159] like Figure 4As shown, the non-transitory computer-readable storage medium 11 stores instructions, such as computer-readable instructions 10. When the computer-readable instructions 10 are executed by a processor, the various methods described above can be performed. The non-transitory computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Non-transitory non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. For example, the non-transitory computer-readable storage medium 11 can be connected to a computing device such as a computer, and then, when the computing device executes the computer-readable instructions 10 stored on the computer-readable storage medium 11, the various methods described above can be performed.

[0160] In the several embodiments provided by this invention, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0161] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0162] Furthermore, the functional units in the various embodiments of the present invention can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0163] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for executing all or part of the steps of the methods of the various embodiments of this invention through a computer device (which may be a personal computer, server, or network device, etc.). The aforementioned storage medium includes: USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, optical disks, and other media capable of storing program code.

[0164] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A smart decision-making system for fault repair of various types of loading and unloading equipment at an ore terminal, characterized in that, include: A multi-dimensional data sensing unit is configured to distribute key operating parts of various types of loading and unloading equipment at an ore terminal, collect multi-dimensional physical quantity data reflecting the operating status of the equipment, and upload the multi-dimensional physical quantity data after preliminary processing. The data preprocessing and feature engineering unit is communicatively connected to the multidimensional data sensing unit and is configured to receive the multidimensional physical quantity data, perform cleaning, noise reduction and standardization on it, and adaptively extract a set of key feature parameters that can characterize the health status of the equipment through a preset feature extraction algorithm. The fault diagnosis and location unit is communicatively connected to the data preprocessing and feature engineering unit and is configured to intelligently identify the current fault type of the equipment and accurately locate the specific component or subsystem where the fault occurs based on the key feature parameter set and using a hybrid diagnostic model that integrates deep learning and knowledge graphs. The fault trend prediction and intelligent early warning unit is communicatively connected to the fault diagnosis and location unit and the data preprocessing and feature engineering unit. It is configured to integrate historical equipment operation data, real-time key feature parameters, environmental factors and maintenance records to construct a multi-factor coupled fault evolution prediction model to predict the development trend of a specific fault mode and the probability of occurrence within a preset time period in the future. It also dynamically classifies the early warning level according to the probability of occurrence and the potential impact of the fault on production and operation. The dynamic resource scheduling and maintenance decision-making unit is communicatively connected to the fault trend prediction and intelligent early warning unit and the fault diagnosis and location unit. It is configured to generate the optimal maintenance plan suggestion through a multi-objective optimization algorithm based on the fault diagnosis results, predicted trends, early warning levels, and real-time status and geographical distribution information of maintenance resources, and automatically trigger or assist in scheduling maintenance resources to realize dynamic dispatching and resource coordination of maintenance tasks. The system management and maintenance center communicates with each of the above-mentioned units and is configured to provide a unified user interface, access control, system parameter configuration, operation log recording and auditing, and support iterative updates of models and algorithms as well as expansion of system functions.

2. The intelligent decision-making system for fault repair of various types of loading and unloading equipment at an ore terminal as described in claim 1, characterized in that, The multi-dimensional physical quantity data includes: Vibration acceleration, velocity, and displacement signals for rotating components such as gearboxes and bearings in large equipment such as ship unloaders and stacker-reclaimer bucket wheel excavators; Temperature field distribution data for the stator, rotor, and bearing housing of the drive motor of a ship unloader and a stacker-reclaimer bucket wheel excavator; Data on hydraulic pipeline pressure, flow rate, and hydraulic oil temperature for hydraulically driven equipment such as hydraulic loading arms and pitching mechanisms; Data on viscosity, moisture content, particle size, and ferrography of gearbox lubricating oil for gear transmission equipment.

3. The intelligent decision-making system for fault repair of various types of loading and unloading equipment at an ore terminal as described in claim 1, characterized in that, The data preprocessing and feature engineering unit is also configured to perform stationarity testing and processing on time-series data, and to perform adaptive noise filtering; to dynamically calibrate feature parameters in conjunction with equipment operating information; and to use unsupervised learning methods to mine abnormal pattern features hidden in the data.

4. The intelligent decision-making system for fault repair of various types of loading and unloading equipment at an ore terminal as described in claim 1, characterized in that, The data preprocessing and feature engineering unit includes extracting feature parameters that reflect the equipment's operating status, such as: extracting the root mean square value of vibration frequency features, which can be expressed as: Where RMS is the root mean square value, xi is the i-th vibration sample value, and N is the number of sampling points, which reflects the magnitude of vibration energy. The larger the value, the more severe the equipment wear. Temperature change features can be used to extract the rate of temperature change, which can be expressed as: Where V is the rate of temperature change, T(t) is the current temperature value, T(t-Δt) is the temperature value Δt before time Δt, and Δt is the time interval; Pressure fluctuation coefficient can be extracted from pressure numerical characteristics using the following expression: Where Pmax is the maximum pressure and Pmin is the minimum pressure. This represents the average pressure. The viscosity change rate, extracted from the state characteristics of oil, can be expressed by the following expression: Where ΔU is the viscosity change rate, U(t) is the current viscosity value, and U0 is the new oil reference viscosity value.

5. The intelligent decision-making system for fault repair of various types of loading and unloading equipment at an ore terminal as described in claim 1, characterized in that, The fault diagnosis and location unit includes methods for diagnosing equipment faults, such as constructing a fault diagnosis model and calculating the equipment health index, which can be expressed as: Where HI is the equipment health index, Fi is the i-th feature value, Fi,min and Fi,max are the boundary values ​​of the normal range of feature parameters, and Wi is the weight of each feature; The device health index is set as follows: within the range of ab, it is normal; within the range of bc, attention is required; within the range of cd, it is abnormal; and above the range of d, it is a fault. Here, a, b, c, and d represent the range of the device health index.

6. The intelligent decision-making system for fault repair of various types of loading and unloading equipment at an ore terminal as described in claim 5, characterized in that, The fault evolution prediction model in the fault trend prediction and intelligent early warning unit is configured as follows: Deep learning prediction models based on attention mechanisms can automatically focus on the key features that have the greatest impact on the development of faults. The dynamic early warning level classification includes at least: advisory early warning, general early warning, important early warning and emergency early warning, and pre-set emergency response suggestions are associated with different early warning levels; The dynamic resource scheduling and maintenance decision-making unit is configured for: Establish a maintenance resource capability matrix and update maintenance personnel skill proficiency, spare parts inventory turnover rate, and tool availability status in real time; When generating the optimal maintenance plan, the urgency of the fault, maintenance cost, impact on production plan, and balanced utilization of resources are comprehensively considered.

7. The intelligent decision-making system for fault repair of various types of loading and unloading equipment at an ore terminal as described in claim 6, characterized in that, The methods for predicting failure trends and probabilities are as follows: Single-device failure prediction uses an LSTM-GARCH model combined with a long short-term memory network and a volatility model to capture the nonlinear trend of equipment performance degradation; multi-device correlation prediction uses the equipment topology as a graph, inputs the state characteristics of each device, predicts the failure propagation path, and constructs the equipment failure probability dependency relationship; probability prediction uses a Cox proportional hazards model to estimate the failure probability of the equipment in the next time period t, which can be expressed by the following expression: h(t,X)=h0(t)e βX Where X is the feature vector, and β is the weight of each feature in the feature vector.

8. The intelligent decision-making system for fault repair of various types of loading and unloading equipment at an ore terminal as described in claim 1, characterized in that, The warning levels are divided into different levels: a fault probability less than P1 is low probability, a fault probability greater than P1 but less than P2 is medium probability, and a fault probability greater than P2 is high probability. The degree of impact is divided into mild impact, moderate impact, and severe impact according to preset evaluation rules. A three-dimensional warning level matrix is ​​constructed, which divides the probability and degree of impact into three levels, forming nine risk combinations, corresponding to a four-level warning system of red, orange, yellow, and blue, where P1 and P2 are the level probability values ​​of the fault occurrence.

9. An electronic device, comprising: At least one memory stores computer-executable instructions non-transitory; At least one processor, configured to run the computer-executable instructions, The computer-executable instructions are executed by the processor to implement the intelligent decision-making system for fault repair of multi-type loading and unloading equipment at ore terminals according to any one of claims 1-8.

10. A computer-readable storage medium, wherein, The computer-readable storage medium stores computer-executable instructions, which, when executed by at least one processor, implement the intelligent decision-making system for fault repair of multi-type loading and unloading equipment at an ore terminal according to any one of claims 1-8.

Images