Fault identification and maintenance method and system for multi-stage geothermal well group system and equipment

Abstract

The disclosure provides a fault identification and maintenance method and system of a multistage geothermal well group system, and equipment, relating to the technical field of geothermal system maintenance. The method comprises: collecting device operation data and device loss usage data of the geothermal system in real time; determining the current fault point and the current fault type through the device operation data and coarse and fine granularity simulation models; determining the future fault prediction result through the device loss usage data; generating a system operation and maintenance strategy according to the device loss usage data, the current fault point, the current fault type and the future fault prediction result; controlling each device in real time through the system operation and maintenance strategy, and pushing the current fault point, the current fault type and the system operation and maintenance strategy to the operation and maintenance object. The scheme can realize the collaborative closed-loop control of accurate positioning, trend prediction and active maintenance strategy optimization of faults in the multistage geothermal well group system, improve the stability of system operation, and reduce operation cost.

Description

Technical Field

[0001] This disclosure relates to the field of geothermal system maintenance technology, and more specifically, to a fault identification and maintenance method, system, and equipment for a multi-level geothermal well group system. Background Technology

[0002] With the continuous advancement of renewable energy development, geothermal energy, as a stable, efficient, and green clean energy source, is increasingly widely used in regional energy systems, especially in multi-level geothermal well systems. By integrating and controlling geothermal wells at different depths and temperature levels, large-scale, multi-load heat supply can be achieved. However, multi-level geothermal systems are complex in structure, diverse in equipment, and subject to frequent dynamic fluctuations in operating conditions, posing significant challenges to system operation and maintenance. For example, a geothermal system includes multiple collaborative subsystems such as wellside extraction, power transmission, heat extraction, hydraulic distribution, and terminal piping. These components are strongly coupled, and malfunctions in a single piece of equipment can trigger system-level failures, thereby affecting the overall stability of energy supply.

[0003] Currently, most geothermal system fault detection methods still rely on single-point sensor monitoring or empirical models, typically focusing on the operational status analysis of individual devices and lacking comprehensive identification capabilities at the subsystem and system-level collaborative states. In this approach, if multiple devices are interconnected or have hidden fault propagation paths, traditional monitoring methods often fail to identify and locate them promptly, thus delaying fault response time. Furthermore, while some solutions establish equipment operation threshold discrimination mechanisms using statistical models, these mechanisms struggle to adapt to dynamic changes in system operating status when facing complex environmental disturbances and equipment degradation processes, posing a risk of false alarms and missed alarms.

[0004] Furthermore, mechanistic simulation models are used in some technical solutions to simulate the operating characteristics of specific equipment under ideal conditions. While these models provide some explanation at the equipment structure level, they are generally difficult to efficiently simulate the dynamic behavior of the entire system, especially when data integrity is limited or external disturbances are frequent, making it difficult to balance simulation accuracy and real-time performance. In particular, current operation and maintenance mechanisms are mostly reactive, lacking the ability to predict the future operating state of the system. This makes it difficult to proactively intervene and optimize maintenance plans before failures occur, thus limiting further improvements in the operational efficiency of geothermal energy systems.

[0005] Therefore, the relevant technologies still have certain limitations in terms of collaborative fault identification, system-level operational status determination, and forward-looking maintenance support for geothermal well cluster systems. There is an urgent need to propose more accurate, efficient, and dynamic system-level diagnostic and maintenance methods to better support the stable, safe, and efficient operation of geothermal energy systems.

[0006] It should be noted that the information disclosed in the background section above is only used to enhance the understanding of the background of this disclosure, and therefore may include information that does not constitute prior art known to those skilled in the art. Summary of the Invention

[0007] The purpose of this disclosure is to provide a fault identification and maintenance method, a fault identification and maintenance system, and electronic equipment for a multi-level geothermal well group system, thereby enabling the coordinated closed-loop control of accurate fault location, trend prediction, and proactive maintenance strategy optimization in the multi-level geothermal well group system, improving the stability of system operation, and reducing operating costs.

[0008] Other features and advantages of this disclosure will become apparent from the following detailed description, or may be learned in part from practice of this disclosure.

[0009] According to a first aspect of the present disclosure, a method for fault identification and maintenance of a multi-level geothermal well group system is provided, including:

[0010] Real-time acquisition of equipment operation data at each level of the multi-level geothermal well group system, and synchronous acquisition of equipment wear and tear data of each device;

[0011] Based on the equipment operation data, as well as the pre-established coarse-grained simulation model and fine-grained simulation model, the current fault point and current fault type of the multi-level geothermal well group system are determined.

[0012] The future failure prediction results of the multi-stage geothermal well group system are determined by the equipment wear and tear data.

[0013] Based on the equipment wear and tear data, the current fault point, the current fault type, and the future fault prediction results, a system operation and maintenance strategy is generated.

[0014] The system operation and maintenance strategy is used to control each device in the multi-level geothermal well group system in real time, and to push the current fault point, the current fault type and the system operation and maintenance strategy to the operation and maintenance object.

[0015] In some exemplary embodiments of this disclosure, based on the foregoing scheme, the coarse-grained simulation model uses each level of subsystem in the multi-level geothermal well group system as the basic unit to describe the energy and mass transfer relationships between the subsystems; the fine-grained simulation model uses each independent device in the subsystem as the basic unit to describe the detailed operating characteristics of the device.

[0016] The process of determining the current fault point and current fault type of the multi-stage geothermal well group system using the equipment operation data and pre-established coarse-grained and fine-grained simulation models includes:

[0017] The equipment operation data is input into the coarse-grained simulation model to perform collaborative simulation calculations on each level of subsystems in the multi-level geothermal well group system, and to determine the system simulation operation results.

[0018] By comparing the system simulation results with the equipment operation data, we can identify the subsystems to be analyzed that have abnormal equipment operation data.

[0019] The equipment operation data corresponding to the subsystem to be analyzed is input into the fine-grained simulation model so as to perform collaborative simulation calculations on each device in the subsystem to be analyzed through the fine-grained simulation model and determine the equipment simulation operation results.

[0020] By comparing the simulation results and the operating data of the equipment, the current fault point and the current fault type of the multi-level geothermal well group system are determined.

[0021] In some example embodiments of this disclosure, based on the foregoing scheme, the method further includes:

[0022] The equipment operation data is input into a pre-trained fault identification and classification model to determine the estimated fault point and estimated fault type of the multi-level geothermal well group system.

[0023] Based on the estimated fault point and the estimated fault type, data items to be compared are filtered from the system simulation results, the equipment simulation results, and the equipment operation data.

[0024] In some example embodiments of this disclosure, based on the foregoing scheme, the equipment operation data is input into a pre-trained fault identification and classification model to determine the estimated fault points and estimated fault types corresponding to the multi-level geothermal well group system, including:

[0025] A sensor time series is constructed based on the device operation data, and the sensor time series is time normalized to obtain the time normalization result.

[0026] The time normalization result is mapped to an angle in a polar coordinate system, and the cosine value of the angle difference is calculated to generate the Gram angle difference field matrix.

[0027] The convolutional neural network in the fault identification and classification model extracts features from the Gram angle difference field matrix, and inputs the extracted feature map into the classification layer to determine the estimated fault point and estimated fault type corresponding to the multi-level geothermal well group system.

[0028] In some example embodiments of this disclosure, based on the foregoing scheme, determining the future failure prediction result of the multi-stage geothermal well group system using the equipment wear and tear data includes:

[0029] The equipment wear data is input into a pre-trained equipment life prediction model based on a long short-term memory network to generate the remaining life of each piece of equipment in the multi-level geothermal well group system under different output load percentages.

[0030] The equipment wear data is input into a pre-trained future failure prediction model to estimate the future failure points and future failure types of the multi-level geothermal well group system.

[0031] The remaining service life of the equipment under different output load percentages, as well as the future failure points and future failure types of the multi-stage geothermal well group system, are used as the future failure prediction results.

[0032] In some example embodiments of this disclosure, based on the foregoing scheme, generating a system operation and maintenance strategy according to the equipment wear and tear data, the current fault point, the current fault type, and the future fault prediction results includes:

[0033] Based on the equipment wear and tear data, the current fault point, and the current fault type, determine the fault severity score for each device;

[0034] Based on the equipment wear and tear data, the current fault points, and the future fault prediction results, a fault occurrence rate score is determined for each piece of equipment.

[0035] Based on the coverage of the monitoring equipment and the diagnostic system corresponding to the multi-level geothermal well group system, the fault detection score of each device is determined.

[0036] The risk order values ​​are determined by the fault severity score, the fault occurrence rate score, and the fault detectability score.

[0037] The maintenance tasks corresponding to each device are sorted based on the risk priority values ​​to generate a system operation and maintenance strategy.

[0038] In some example embodiments of this disclosure, based on the foregoing scheme, the method further includes:

[0039] With the goals of meeting minimum energy supply requirements and maximizing equipment lifespan, and taking equipment capacity and equipment health status as constraints, a multi-objective optimization model is established.

[0040] The multi-objective optimization model is solved using a non-dominated sorting genetic algorithm to determine the target equipment operation strategy;

[0041] The system operation and maintenance strategy is optimized based on the target device operation strategy to obtain the optimized system operation and maintenance strategy;

[0042] The optimized system operation and maintenance strategy is used to control each device in the multi-level geothermal well group system in real time.

[0043] In some example embodiments of this disclosure, based on the foregoing scheme, the multi-objective optimization model is as follows:

[0044]

[0045] in, The objective function representing the energy supply-demand imbalance is... This represents the total runtime of the system, and it seeks the longest stable operating cycle while satisfying power supply constraints. This represents the system heat supply at time t. This represents the heat demand at time t. The objective function representing system damage is... This represents the instantaneous damage rate at time t. This represents the geothermal water flow rate at time t. , These represent the minimum and maximum safe flow limits for the system's pump equipment, respectively. Indicates the rate of change of flow rate. This represents the upper limit of the rate of change of flow rate, and is the safe adjustment rate of the equipment. , Let represent the inlet water temperature and the outlet water temperature at time t, respectively. , These represent the minimum allowable inlet water temperature and the maximum allowable outlet water temperature, respectively.

[0046] According to a second aspect of the present disclosure, a fault identification and maintenance device for a multi-level geothermal well group system is provided, comprising:

[0047] The equipment data acquisition module is used to collect equipment operation data of each level of equipment in the multi-level geothermal well group system in real time, and to simultaneously acquire equipment wear and tear data of each of the equipment.

[0048] The equipment fault identification module is used to determine the current fault point and current fault type of the multi-level geothermal well group system by using the equipment operation data and pre-established coarse-grained simulation models and fine-grained simulation models.

[0049] The equipment failure prediction module is used to determine the future failure prediction results of the multi-level geothermal well group system based on the equipment wear and tear data.

[0050] The equipment operation and maintenance strategy generation module is used to generate system operation and maintenance strategies based on the equipment wear and tear data, the current fault point, the current fault type, and the future fault prediction results.

[0051] The equipment maintenance module is used to control each device in the multi-level geothermal well group system in real time through the system operation and maintenance strategy, and to push the current fault point, the current fault type and the system operation and maintenance strategy to the operation and maintenance object.

[0052] According to a third aspect of the present disclosure, an electronic device is provided, comprising: a processor; and a memory storing computer-readable instructions, wherein the computer-readable instructions, when executed by the processor, implement the fault identification and maintenance method for a multi-level geothermal well group system as described in the first aspect.

[0053] According to a fourth aspect of the present disclosure, a computer-readable storage medium is provided having a computer program stored thereon, which, when executed by a processor, implements the fault identification and maintenance method for a multi-level geothermal well group system as described in the first aspect.

[0054] The technical solutions provided in this disclosure may have the following beneficial effects:

[0055] The fault identification and maintenance method for a multi-stage geothermal well group system in the example embodiments of this disclosure, on the one hand, by collecting operating data and wear and tear data of equipment at each level in the multi-stage geothermal system, and performing multi-level modeling and analysis of the system state based on coarse-grained and fine-grained simulation models, can quickly identify the key subsystems corresponding to the abnormal propagation path when the system exhibits performance deviations. Furthermore, by combining equipment operating data, it can infer potential faulty equipment and their fault types, effectively improving the accuracy of fault identification and avoiding the lack of responsiveness of related technologies when dealing with systemic failures caused by equipment coupling, reducing false alarm rates, and thus improving fault repair efficiency. On the other hand, by fusing equipment wear and tear data with historical equipment operation and maintenance data... Fault trend prediction modeling based on multi-dimensional factors such as maintenance behavior and operation records enables maintenance strategies to shift from passive response to proactive control. This significantly compensates for the limitations of traditional maintenance methods, which rely on manual experience, have delayed planning, or insufficient coverage. It helps to arrange maintenance windows and resource allocation in advance, improves the planning and efficiency of system maintenance, and ensures the stability of system operation. Furthermore, by integrating fault identification, loss modeling, and prediction results, a system operation and maintenance strategy can be further formed. This strategy can be used to push operation control and maintenance tasks to various equipment in the geothermal system. Under the premise of meeting the system's continuous energy supply requirements, it can dynamically coordinate equipment start-up and shutdown and load distribution, thereby suppressing unplanned degradation of key equipment, improving the overall operational stability of the system, and reducing the system's operation and maintenance costs.

[0056] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and are not intended to limit this disclosure. Attached Figure Description

[0057] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this disclosure and, together with the description, serve to explain the principles of this disclosure. It is obvious that the drawings described below are merely some embodiments of this disclosure, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort.

[0058] Figure 1 The illustration shows a flowchart of a fault identification and maintenance method for a multi-stage geothermal well group system according to some embodiments of the present disclosure.

[0059] Figure 2 The schematic diagram illustrates a process flow diagram for simulating the current fault point and the current fault type according to some embodiments of the present disclosure.

[0060] Figure 3 The illustration shows a flowchart of predictive classification to determine the estimated fault point and the estimated fault type according to some embodiments of the present disclosure.

[0061] Figure 4 The illustration shows a flowchart of generating system operation and maintenance strategies based on risk sequence numbers according to some embodiments of the present disclosure.

[0062] Figure 5 The illustration schematically shows a process framework diagram for fault identification and maintenance of a multi-stage geothermal well group system according to some embodiments of the present disclosure.

[0063] Figure 6 The diagram illustrates a framework of a mechanistic model corresponding to a multi-stage geothermal well cluster system according to some embodiments of the present disclosure.

[0064] Figure 7 A schematic diagram of a fault identification and maintenance apparatus for a multi-stage geothermal well group system according to some embodiments of the present disclosure is shown.

[0065] Figure 8 The schematic diagram illustrates the structural schematic of a computer system of an electronic device according to some embodiments of the present disclosure.

[0066] Figure 9 A schematic diagram of a computer-readable storage medium according to some embodiments of the present disclosure is shown.

[0067] In the accompanying drawings, the same or corresponding reference numerals indicate the same or corresponding parts. Detailed Implementation

[0068] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numerals in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this specification. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this specification as detailed in the appended claims.

[0069] The terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to be limiting of this specification. The singular forms “a,” “the,” and “the” as used in this specification and the appended claims are also intended to include the plural forms unless the context clearly indicates otherwise. It should also be understood that the term “and / or” as used herein refers to and includes any and all possible combinations of one or more of the associated listed items.

[0070] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this disclosure. However, those skilled in the art will recognize that the technical solutions of this disclosure can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this disclosure.

[0071] Furthermore, the accompanying drawings are for illustrative purposes only and are not necessarily drawn to scale. The block diagrams shown in the drawings are merely functional entities and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in software, in one or more hardware modules or integrated circuits, or in different network and / or processor devices and / or microcontroller devices.

[0072] In this example embodiment, a fault identification and maintenance method for a multi-level geothermal well group system is first provided. This fault identification and maintenance method for a multi-level geothermal well group system can be applied to terminal equipment, such as maintenance terminals, host computers and other electronic devices, or it can be applied to servers. This example embodiment does not make any special limitations on this. The following description will take the server executing the method as an example. Figure 1 The illustration schematically depicts a fault identification and maintenance method flow for a multi-stage geothermal well cluster system according to some embodiments of the present disclosure. (Reference) Figure 1 As shown, the fault identification and maintenance method for this multi-stage geothermal well group system may include the following steps:

[0073] Step S110: Real-time acquisition of equipment operation data of each level of equipment in the multi-level geothermal well group system, and synchronous acquisition of equipment wear and usage data of each of the equipment.

[0074] Step S120: Using the equipment operation data and the pre-established coarse-grained simulation model and fine-grained simulation model, determine the current fault point and current fault type of the multi-level geothermal well group system.

[0075] Step S130: Determine the future failure prediction results of the multi-stage geothermal well group system based on the equipment wear and tear data;

[0076] Step S140: Generate a system operation and maintenance strategy based on the equipment wear and tear data, the current fault point, the current fault type, and the future fault prediction results;

[0077] Step S150: Real-time control of each device in the multi-level geothermal well group system is performed through the system operation and maintenance strategy, and the current fault point, the current fault type, and the system operation and maintenance strategy are pushed to the operation and maintenance object.

[0078] According to the fault identification and maintenance method for a multi-level geothermal well group system in this example embodiment, on the one hand, by collecting the operating data and wear and tear data of equipment at each level in the multi-level geothermal system, and performing multi-level modeling and analysis of the system state based on coarse-grained and fine-grained simulation models, the key subsystems corresponding to the abnormal propagation path can be quickly identified when performance deviations occur in the system. Furthermore, by combining the equipment operating data, potential faulty equipment and their fault types can be inferred, which can effectively improve the accuracy of fault identification, avoid the lack of responsiveness of related technologies when dealing with systemic failures caused by equipment coupling, reduce the false alarm rate, and thus improve fault repair efficiency. On the other hand, by integrating equipment wear and tear data with historical equipment operation and maintenance data, the system can effectively identify the key subsystems corresponding to the abnormal propagation path when performance deviations occur. Fault trend prediction modeling based on multi-dimensional factors such as behavior and operation records enables maintenance strategies to shift from passive response to proactive control. This significantly overcomes the limitations of traditional maintenance methods, which rely on manual experience, have delayed planning, or insufficient coverage. It helps to schedule maintenance windows and resource allocation in advance, improves the planning and efficiency of system maintenance, and ensures the stability of system operation. Furthermore, by integrating fault identification, loss modeling, and prediction results, a system operation and maintenance strategy can be further formed. This strategy can be used to push operation control and maintenance tasks to various devices in the geothermal system. Under the premise of meeting the system's continuous energy supply requirements, it can dynamically coordinate equipment start-up and shutdown and load distribution, thereby suppressing unplanned degradation of key equipment, improving the overall operational stability of the system, and reducing the system's operation and maintenance costs.

[0079] The following will further explain the fault identification and maintenance method for the multi-stage geothermal well group system in this example embodiment.

[0080] In step S110, real-time equipment operation data of each level of equipment in the multi-level geothermal well group system is collected, and equipment wear and tear data of each piece of equipment are acquired simultaneously.

[0081] In one example embodiment of this disclosure, equipment operation data refers to the sensing information that can dynamically reflect the operating status of various key equipment during the operation of a multi-level geothermal well group system. For example, equipment operation data may include, but is not limited to, multi-dimensional time-series parameters such as temperature, pressure, flow rate, current, voltage, rotational speed, vibration intensity, and start-stop status. Of course, it may also be other types of data that characterize the operating status of equipment, and this embodiment is not limited thereto.

[0082] High-speed acquisition, preliminary filtering, and normalization of equipment operation data can be achieved by deploying various industrial-grade sensors and combining them with edge acquisition modules. Sensors can be installed at various geothermal wells, heat pumps, water pumps, valves, heat exchangers, and pipeline nodes, and logically connected to the system edge data bus, thereby enabling synchronous acquisition of data from multiple devices and locations.

[0083] Equipment wear and tear data refers to a non-real-time data set reflecting the historical performance degradation process, maintenance behavior, and usage load characteristics of equipment. It is used to characterize the health trajectory of equipment over a medium- to long-term operating cycle. For example, equipment wear and tear data can include at least: the equipment's cumulative runtime, reflecting the overall degree of aging; historical operating load curves, used to identify high-frequency large fluctuations or prolonged overload operation; start-stop frequency and start-stop interval time series, used to infer the effects of mechanical shock and cyclic fatigue; planned and unplanned maintenance records, including maintenance type, maintenance time, and replaced parts information; fault history logs and alarm trigger records, used to locate hidden fault modes in the equipment; and operating condition change logs, including system switching strategies, operating strategy interventions, and energy-saving mode configurations. Equipment wear and tear data can be obtained from multi-source heterogeneous data platforms such as SCADA systems, DCS systems, equipment controllers, historical databases, and maintenance work order management systems, and integrated into models through structured interfaces. To ensure data quality, data preprocessing techniques such as sliding window aggregation, outlier removal, and time-aligned interpolation can be used in system implementation to generate data vectors with unified semantics. These preprocessing techniques are common techniques in this field and will not be elaborated here.

[0084] In step S120, the current fault point and current fault type of the multi-level geothermal well group system are determined by the equipment operation data and the pre-established coarse-grained simulation model and fine-grained simulation model.

[0085] In one example embodiment of this disclosure, a coarse-grained simulation model is used to model and simulate the coupling behavior between subsystems in a multi-level geothermal well group system from a system-level perspective. For example, the subsystems in a multi-level geothermal well group system may include a wellside extraction subsystem, a power transmission subsystem, a heat extraction subsystem, a hydraulic distribution subsystem, and a terminal transmission and distribution network system. The coarse-grained simulation model is built on the basis of mass conservation, energy balance, and hydraulic characteristics, and can reflect the flow, pressure, and temperature difference transmission laws between subsystems under different loads or operating conditions.

[0086] Equipment operating data can be input into a coarse-grained simulation model. By comparing the deviation between the model's predicted output and the actual observed data, subsystems exhibiting abnormal operating behavior can be identified as objects to be analyzed. Subsequently, for subsystems initially identified as abnormal, their subordinate key equipment can be further selected as input objects, and a fine-grained simulation model can be invoked for equipment-level simulation. The fine-grained simulation model is built based on the equipment's physical structure and operating mechanism, employing distributed parameter modeling or modular thermo-hydraulic modeling methods to simulate and recreate the equipment's response behavior. By comparing the equipment's simulated output values ​​with the actual collected values, combined with the equipment's operating rules, design specifications, and empirical threshold libraries, equipment nodes with significant deviations can be identified, and their corresponding current fault types can be determined based on the deviation characteristics.

[0087] In step S130, the future failure prediction results of the multi-stage geothermal well group system are determined by using the equipment loss data.

[0088] In one example embodiment of this disclosure, the future failure prediction result refers to a set of information generated based on equipment wear and tear data, used to describe the potential failure risk of the equipment during its future operating cycle. It aims to proactively model and predict equipment failures that have not yet actually occurred but have statistical or trend indicators. The future failure prediction result can be a device-level Remaining Useful Life (RUL) estimate, reflecting the time span during which the equipment can continue to operate stably under current load conditions; or it can be a system-level potential failure prediction item, indicating the possible failure points and corresponding failure type categories that may occur in the system within a specific future time period.

[0089] Future failure prediction results can be generated jointly by at least two parallel models. For example, for the remaining service life of equipment, a life prediction model based on Long Short-Term Memory (LSTM) networks can be used. Its inputs are time series features such as historical operating load sequences, maintenance codes, start-up and shutdown frequencies, and operating temperatures. Combined with physical consistency regularization constraints, the output is a predicted life function or life curve under different load levels, which is used to characterize the equipment degradation rate. For system-level potential failure prediction, a future failure type prediction model based on image-based time series classification can be used. By encoding multidimensional operating data into a Gram difference field matrix and inputting it into a convolutional neural network, the output is the classification probability distribution of possible failure types and the corresponding prediction time window.

[0090] In step S140, a system operation and maintenance strategy is generated based on the equipment wear and tear data, the current fault point, the current fault type, and the future fault prediction results.

[0091] In one example embodiment of this disclosure, the system operation and maintenance strategy refers to a set of execution strategies used to guide equipment scheduling, load allocation, start-up and shutdown control and preventive maintenance plans. Its core purpose is to maximize the service life of critical equipment, reduce the risk of sudden failures and optimize the allocation of operation and maintenance resources while ensuring the continuity of power supply.

[0092] Specifically, during strategy generation, based on the current fault point and fault type determination results, a list of equipment requiring immediate intervention can be identified. This list is then combined with the operational intensity and health score recorded in the equipment wear and tear data to calculate the fault severity level. Secondly, the warning information from future fault prediction results can be linked with the corresponding equipment operating curves for analysis, assessing the impact of potential future faults on system safety, power supply stability, and maintenance costs, thereby calculating the equipment fault occurrence rate level. Thirdly, based on the existing monitoring equipment and diagnostic capabilities of the multi-level geothermal well cluster system, the system's detectability level for various faults can be assessed, resulting in a fault detection score for the equipment. These three scoring items are mapped to Severity, Occurrence, and Detection, respectively. A Risk Priority Number (RPN) is calculated through multi-indicator fusion, and a maintenance priority queue is generated based on the RPN ranking.

[0093] Optionally, a multi-objective optimization model can be introduced, integrating multiple scheduling conditions such as system load demand, equipment health status, service life degradation rate, and start-stop frequency constraints. A solution space search is performed using a non-dominated sorting genetic algorithm (NSGA-II) or other swarm intelligence algorithms to output an operation scheduling scheme that simultaneously satisfies the dual objectives of "stable energy supply" and "extended equipment lifespan." Ultimately, the generated system operation and maintenance strategy includes not only fault handling measures, equipment start-stop plans, and load allocation instructions, but also maintenance window arrangements, maintenance method suggestions (such as on-site replacement, remote restart, or standby fault tolerance), and task execution order, effectively improving fault repair efficiency and ensuring system stability.

[0094] In step S150, the system operation and maintenance strategy is used to control each device in the multi-level geothermal well group system in real time, and the current fault point, the current fault type, and the system operation and maintenance strategy are pushed to the operation and maintenance object.

[0095] In one example embodiment of this disclosure, real-time control refers to synchronously sending optimized start / stop commands, load adjustment schemes, and operating state switching logic to the field equipment control system. Through equipment controllers, such as programmable logic controllers (PLCs) or edge execution units, corresponding scheduling commands are executed on the target equipment to achieve dynamic adjustment of the system's operating state. For example, based on the operating priorities set in the strategy, some redundant operating equipment can be switched, abnormal equipment can be reduced in load or shut down safely, and the load can be weighted and allocated to equipment with higher health levels to maintain a stable output of overall heating capacity.

[0096] Information push refers to the real-time delivery of identified fault points, corresponding fault types, and resulting maintenance strategies to the maintenance recipients in the form of structured text, graphical prompts, or operational suggestions, through the system's communication module or maintenance platform. These recipients could be on-site maintenance personnel, remote dispatchers, or backend maintenance information systems. To improve push effectiveness, content can be automatically linked to the device topology view for visualized, linked positioning, along with historical trend charts, diagnostic summary, and suggested response plans to assist maintenance personnel in rapid response and closed-loop management. Push methods can include local terminal alarms, maintenance mobile app notifications, SMS or email triggers, or integration into the unified maintenance platform's event management module. This supports automatic work order generation and task scheduling interface linkage, thereby achieving closed-loop management of the entire maintenance process from identification and diagnosis to execution.

[0097] The following is a detailed description of steps S110 to S150.

[0098] In one example embodiment of this disclosure, the coarse-grained simulation model uses subsystems at various levels within a multi-level geothermal well cluster system as basic units to describe the energy and mass transfer relationships between these subsystems. Subsystems may include, but are not limited to, wellside production subsystems, power transmission subsystems, heat extraction subsystems, hydraulic distribution subsystems, and transmission and distribution pipeline systems. Each subsystem consists of multiple devices that are physically coupled with each other in terms of flow rate, pressure, temperature difference, and load response.

[0099] Coarse-grained simulation models are primarily used for system-level operational state modeling. During the modeling process, differential control equations can be constructed based on the first law of thermodynamics, the hydraulic continuity equation, and the principle of energy conservation. Boundary conditions and system topology are introduced to calculate the expected response output of each subsystem under given operating conditions. For example, it can be used to simulate and calculate heat pump outlet temperature, water pump flow rate, water collector pressure distribution, and main pipe flow gradient. To improve the model's adaptability and computational efficiency, a modular networking approach can be used to construct simulation blocks for different subsystems, employing a unified node-connector description language to achieve overall system topology assembly. Alternatively, a state-space model or a semi-physical modeling method based on constitutive relations can be used, incorporating empirical parameter correction terms to mitigate the effects of sensor drift and boundary disturbances. This example implementation does not impose specific limitations on the construction method of the coarse-grained simulation model.

[0100] Fine-grained simulation models use individual devices within a subsystem as basic units to describe the detailed operating characteristics of those devices. Unlike coarse-grained simulation models, fine-grained models use devices as the core modeling unit, such as individual pumps, heat exchangers, fans, valves, and heat pump units, constructing microscopic physical behavior models based on their internal structural characteristics and operating laws. Taking a heat exchanger as an example, it can be modeled as a dual-channel coupled system driven by the temperature difference of the heat transfer surface, and its outlet temperature can be solved using dynamic differential control equations. For a pump, a multi-parameter characteristic curve of head-flow-speed can be introduced, while simultaneously superimposing wear attenuation coefficients and efficiency variation factors. Fine-grained simulation models can use refined time steps for simulation calculations, enabling the simulation of device response characteristics under rapid fluctuations or abnormal operating conditions. In terms of implementation, analytical simulation models based on energy conservation, finite element simulation models, or neural network approximation models can be used. These are all common simulation modeling methods in this field and will not be elaborated upon here. The specific approach can be determined by the modeling accuracy requirements, operational computational costs, and data availability.

[0101] It can be done Figure 2 The steps described herein enable the determination of the current fault point and type in a multi-stage geothermal well cluster system using equipment operation data and pre-established coarse-grained and fine-grained simulation models. (Refer to...) Figure 2 As shown, it can specifically include:

[0102] Step S210: Input the equipment operation data into the coarse-grained simulation model to perform collaborative simulation calculations on each level of subsystems in the multi-level geothermal well group system through the coarse-grained simulation model, and determine the system simulation operation results;

[0103] Step S220: Compare the system simulation results with the equipment operation data to identify the subsystems to be analyzed where the equipment operation data is abnormal;

[0104] Step S230: Input the equipment operation data corresponding to the subsystem to be analyzed into the fine-grained simulation model, so as to perform collaborative simulation calculation on each device in the subsystem to be analyzed through the fine-grained simulation model and determine the equipment simulation operation results;

[0105] Step S240: Compare the simulation results of the equipment with the equipment operation data to determine the current fault point and the current fault type of the multi-level geothermal well group system.

[0106] The system simulation results refer to a set of time-series system-level state prediction values ​​obtained by performing collaborative simulation calculations on each level of subsystems of the entire multi-level geothermal well group system based on a coarse-grained simulation model after inputting the current equipment operating data. The system simulation results reflect the normal operating state that the multi-level geothermal well group system should exhibit under given input conditions. Specifically, they can include predicted values ​​of key performance parameters of each subsystem, such as predicted heat pump outlet temperature, predicted water pump flow rate, water collector pressure balance value, and pipe network temperature drop distribution curve. The system simulation results are a quantitative expression of the system's "expected" behavior and serve as a fundamental reference for subsequent comparison with actual operating data, identification of system operating deviations, and judgment of whether subsystems are abnormal. The system simulation results are usually expressed in the form of a multi-dimensional time series matrix, where each column corresponds to a characteristic variable of a subsystem, and each row corresponds to a sampling time point. They can also be interfaced with standard model output interfaces (such as FMU and Modelica) for compatibility with engineering software platforms.

[0107] Before inputting equipment operation data into the coarse-grained simulation model, the data can be standardized in units, cleaned, and have missing data filled in. Then, time-series vectors are assembled using timestamp alignment and input into the model. During co-simulation calculations, different subsystem models are assembled according to their structural connections, and flow boundaries and coupling conditions are set between them. This allows the simulation process to realistically reflect the actual operating logic of the system, achieving multi-variable thermal-fluid linkage prediction from the wellhead to the grid outlet.

[0108] A subsystem to be analyzed refers to a target subsystem identified after comparing the system simulation results with real-time equipment operation data, showing significant state deviations and potentially containing fault sources. This identification process can be based on residual calculation and threshold determination mechanisms. For example, if the deviations between the simulated and measured values ​​of multiple key parameters in a subsystem exceed a set threshold range, or if the deviations exhibit a continuous increasing trend or non-physical jump characteristics, then that subsystem can be marked as a subsystem to be analyzed. The subsystem to be analyzed serves as the input range target for fine-grained simulation models, used to narrow the diagnostic search space, improve simulation efficiency, and enhance the accuracy of fault location. The identification of subsystems to be analyzed can support single-point marking or parallel marking of multiple subsystems, and the identification results can at least include the subsystem number, abnormal parameter items, deviation statistics, and trigger time windows for parameter context passing in subsequent equipment-level simulation processes.

[0109] Equipment simulation results refer to the set of predicted operating states of each independent device under theoretically normal conditions, obtained after inputting the operating data of the equipment corresponding to the subsystem under analysis into a fine-grained simulation model. These simulation results are based on a single device as the smallest unit of granularity. The simulation output of each device can include its main performance indicators. For example, the simulation output of a fine-grained simulation model can be time-series predicted values ​​for heat exchanger outlet temperature, pump head, fan airflow, and heat pump heating efficiency. The simulation results can be expressed as floating-point multidimensional vectors or include diagnostic auxiliary quantities, such as the first derivatives of various performance indicators and nonlinear response characteristics, to assist in subsequent comparisons and identification of abnormal behaviors.

[0110] In equipment-level simulation, to ensure the consistency and physical rationality of the predicted results, the subsystem boundary states (such as pressure boundaries and inlet temperatures) output by the coarse-grained simulation can be used as boundary condition inputs for the fine-grained model to guarantee the consistency of upstream and downstream coupling. The comparison error between the equipment simulation results and the actual equipment operating data can be used to construct a diagnostic residual vector, which can then be mapped to the fault type identification results, forming a closed-loop support for the diagnostic chain.

[0111] By comparing equipment simulation results and equipment operation data, the current fault point and type of a multi-level geothermal well cluster system can be determined. During the comparison process, a simulation-measurement comparison table is established for each device, and multi-factor comprehensive analysis is performed based on fault type templates, feature offset trends, and a diagnostic rule base. For example, if a water pump's simulation predicts a pressure increase but the actual measured pressure decreases, accompanied by behavioral characteristics such as decreased current and increased temperature, it can be identified as a "pump blockage" type mechanical fault according to the rule base. The judgment result can include the current fault point (e.g., equipment ID), the current fault type (e.g., structural fault, operational fault, control fault), and the corresponding confidence level, serving as an important basis for subsequent maintenance and strategy generation.

[0112] By introducing a collaborative discrimination mechanism between coarse-grained and fine-grained simulation models, a multi-scale fault analysis path from the system level to the device level is constructed. While ensuring the consistency of the overall operating trend, it achieves accurate backtracking and type determination of local abnormal behaviors, effectively reducing false alarm and false diagnosis rates. At the same time, it has good scalability and model generalization ability, and is especially suitable for multi-level energy systems with complex subsystem structures and easily changing boundary conditions.

[0113] In an optional embodiment of this disclosure, equipment operation data can be input into a pre-trained fault identification and classification model to determine the estimated fault points and estimated fault types corresponding to the multi-level geothermal well group system; based on the estimated fault points and estimated fault types, data items to be compared are filtered from the system simulation operation results, equipment simulation operation results, and equipment operation data.

[0114] The fault identification and classification model refers to a machine learning or deep learning model structure used to classify and judge the operating status of a multi-level geothermal well group system based on equipment operation data, thereby identifying potential fault points and corresponding fault types. This model is trained using supervised learning based on historical operating data samples, including normal operation samples and known fault-labeled samples. The model input is a multi-dimensional sensor data tensor after time-series processing and standardization, and the output is a multi-class classification result, representing the fault type category most likely corresponding to the current operating status.

[0115] The fault identification and classification model can employ Convolutional Neural Networks (CNN), Recurrent Nearal Networks (RNN), Convolutional Recurrent Neural Networks (CRNN), or Transformer-like structures. It can also incorporate attention mechanisms or graph neural networks. This embodiment does not impose any specific limitations on the model structure. The core of the fault identification and classification model lies in its ability to extract and classify weak anomalies in time-series data using high-dimensional features, thereby identifying potential system fault trends in advance even without obvious sensor anomaly alarms, thus improving the foresight of the system's intelligent diagnosis.

[0116] An estimated fault point refers to the target device or set of devices with the highest abnormal response characteristics, judged by the fault identification and classification model to be most likely to fail. Essentially, it is a high-risk node selected by the model after scoring and ranking the status of each device in the system's device space based on the input current operating data. It is usually represented by device number, location information, or the name of its subsystem. An estimated fault point can include a single device or multiple devices. When outputting the fault point location result, the fault identification and classification model can combine attention weights, confidence thresholds, or multi-layered voting mechanisms to rank the devices.

[0117] The estimated fault type refers to the category of the most likely fault mode determined by the fault identification and classification model based on the input operational data samples. It describes the type of failure risk the system faces in its current state. The classification dimensions of the estimated fault type can be predefined according to the system settings. For example, it can include fault categories such as decreased heat exchange efficiency, abnormal flow of pump equipment, heat exchanger blockage, valve jamming, frequent start-stop of heat pumps, and motor overload operation. This embodiment does not impose any special limitations on this. The estimated fault type is usually the highest probability term in the multi-class Softmax probability distribution output by the model as the final result, and a confidence score and error bar chart can be attached to evaluate the robustness of the judgment.

[0118] The data items to be compared refer to the set of data selected from system simulation results, equipment simulation results, and equipment operation data for comparative analysis, based on the equipment or subsystems involved in the estimated fault point and type. The data items to be compared may include simulation output values ​​(e.g., predicted flow rate, outlet temperature) of the equipment suspected of being faulty, actual operating data of the equipment (e.g., real-time sensor values), and coupling parameters (e.g., boundary flow rate, inlet pressure difference) of the equipment or its upstream subsystems in the system simulation. These are used to determine whether the actual operating deviation matches the characteristic template of the estimated fault type through error calculation, residual construction, or trend offset analysis. The selection process for the data items to be compared supports time alignment, parameter matching, and sampling frequency consistency processing. They are typically organized in the form of a triplet of equipment ID, timestamp, and parameter name to support the comparison and verification between model predictions and mechanism simulations.

[0119] During the model usage phase, equipment operation data can be organized into time series samples according to a preset time window sliding pattern. Each sample is in units of fixed-length time intervals (e.g., the first 5 minutes, 10 minutes, 30 minutes), forming a two-dimensional tensor input fault identification and classification model. The fault identification and classification model can perform forward inference on this sample, outputting a set of classification probability vectors, where each dimension represents the discrimination probability of the corresponding fault type. The term with the highest probability is the estimated fault type. Simultaneously, by combining attention weight visualization or location information decoding, the equipment number or location most significantly affected by the fault can be further determined, achieving the localization of the estimated fault point.

[0120] Since the model's output estimation results can be used to identify the location of the most likely problematic device or subsystem in the system, the estimated fault point and fault type can serve as a basis for selecting key comparison data items in subsequent simulation diagnostic processes. In practice, the set of subsystem predicted values ​​corresponding to the estimated fault point location can be extracted from the system simulation results, filtering out data from areas irrelevant to the fault to reduce the computational load of simulation comparison. Then, the predicted values ​​of devices matching the estimated fault point device ID are extracted from the device simulation results, and their key operating parameters are extracted to construct residual vectors. Furthermore, actual operating observation data of the same device is extracted from the device operating data and compared with the simulation output. This comparison process can employ residual absolute value comparison, Dynamic Time Warping (DTW) distance, or KL divergence to quantify the degree of difference. Alternatively, a pattern matching method based on fuzzy rules can be used to determine whether the deviation category matches the characteristic pattern of the estimated fault type.

[0121] By introducing a machine learning-based fault identification and classification model as a priori judgment mechanism, the fault point and fault type are estimated before simulation analysis, thereby providing target guidance and data screening basis for the system simulation process, avoiding redundant calculations caused by full data comparison, improving the execution efficiency of the diagnostic path and the discrimination accuracy of key data, and realizing the integrated diagnostic effect of data-driven and mechanism model.

[0122] Furthermore, it can be done through Figure 3 The steps described in the text involve inputting equipment operation data into a pre-trained fault identification and classification model to determine the estimated fault points and types corresponding to the multi-level geothermal well group system, referencing... Figure 3 As shown, it can specifically include:

[0123] Step S310: Construct a sensor time series based on the device operation data, and perform time normalization processing on the sensor time series to obtain a time normalization result;

[0124] Step S320: Map the time normalization result to an angle in a polar coordinate system, calculate the cosine value of the angle difference, and generate the Gram angle difference field matrix;

[0125] Step S330: The Gram angle difference field matrix is ​​used to extract features through the convolutional neural network in the fault identification and classification model, and the extracted feature map is input into the classification layer to determine the estimated fault point and estimated fault type corresponding to the multi-level geothermal well group system.

[0126] Specifically, a sensor time series is constructed based on the equipment operation data, and the sensor time series is then time-normalized to obtain the time-normalized result. The equipment operation data consists of continuous time series data composed of various key parameters such as temperature, pressure, flow rate, current, voltage, and start / stop status. To facilitate processing by the neural network model, each type of sensor data is first reorganized into equally spaced time series segments along the time axis, and multi-dimensional feature splicing is performed. On this basis, the min-max normalization method is applied to the original numerical sequence to uniformly map all feature data to the [0, 1] interval, thereby eliminating the influence of dimensional differences on model convergence. For example, for each historical fault event (T... fail Type fail ), where T fail It can represent historical failure points, Type fail It can represent the historical fault type corresponding to the historical fault point, and can obtain the sensor time series from a period of time before the fault. The sensor time series can be normalized, specifically through the following relationship:

[0127] ;

[0128] in, This can represent the normalized observation value of the i-th sensor. It can represent the sensor time series The observations from the i-th sensor are then normalized and mapped to angles in polar coordinates.

[0129] ;

[0130] in, This can represent the normalized angle mapped to the i-th sensor observation in polar coordinates. Furthermore, the cosine of the angle difference can be calculated to generate the Gram angle difference field matrix. :

[0131] ;

[0132] in, The mapping angle can represent the observation value of the i-th sensor. Mapping angle with the j-th sensor observation The matrix formed by the cosine of the difference between the time series and the time series is called the Gramian Angular Field (GAF) matrix. It is used to encode a one-dimensional time series into a two-dimensional spatial structure, thereby capturing the global dependencies and sequence shape features between time points.

[0133] Next, features can be extracted from the Gram angle difference field matrix using a convolutional neural network in the fault identification classification model. The extracted feature map is then input into the classification layer to determine the estimated fault points and types corresponding to the system. The convolutional neural network structure typically consists of multiple convolutional layers, activation function layers, pooling layers, and fully connected layers, used to extract multi-scale image-level temporal features from the GAF matrix. Specifically, in each convolutional computation, a fixed-size convolutional kernel is used to perform a sliding convolution operation on the input feature map. The convolution result is then non-linearly mapped using the ReLU activation function (corrected linear unit) to form the output feature map. For example, the output feature map can be represented by the following relationship:

[0134] ;

[0135] in, It can represent the first In the output feature map of a convolutional network, the activation value at the m-th row and n-th column is; ReLU() can represent the rectified linear unit (ReLU) activation function, which is a commonly used nonlinear activation function used to introduce nonlinearity and enhance the expressive power of the network. It can represent the first The weight coefficients in the i-th row and j-th column of the convolutional kernel are one of the learnable parameters during model training, which determine the weighted sensitivity of the convolutional kernel to the local input region. It can represent the size of the convolution kernel (filter); It can represent the first The pixel value or activation value at the position of the m+i-1th row and n+j-1th column in the input feature map of layer -1; It can represent the first The bias term of the convolutional operation is a learnable scalar parameter used to perform a linear shift after weighted summation, enhancing the model's representational power.

[0136] After several layers of convolution and pooling operations, the final feature map is flattened and input into a fully connected layer to obtain the final class logits vector. This vector is then normalized by a Softmax layer, outputting the probability distribution of each fault type category. The category number corresponds to the fault type, and the term with the highest probability is the estimated fault type determined by the model. Furthermore, a localization module (such as CAM or Grad-CAM) uses the estimated fault type to deduce the key region that led to this determination, thereby identifying the device number or subsystem location of the estimated fault point.

[0137] To further improve the accuracy and generalization ability of the classification model, cross-entropy is used as the loss function throughout the training process. For example, in this embodiment, cross-entropy as the loss function can be expressed by the following relationship:

[0138] ;

[0139] in, It can represent the overall loss function value, that is, the error metric produced by the training samples under the current model prediction results; It can represent the total number of training samples. This can represent the total number of categories in a classification task, corresponding to the number of predefined fault types in a geothermal system. It can represent the category weighting coefficient. This can represent the label of the i-th sample in the k-th class. It can be represented as given the i-th input sample Under the given conditions, the model predicts the probability value of it belonging to the k-th class. This can represent the categorical variable predicted by the model. Furthermore, training optimization can utilize the Adaptive Moment Estimation (Adam) optimizer to automatically adjust the learning rate; the specific parameter updates can be expressed as the following relationship:

[0140] ;

[0141] in, This can represent the parameters of the model to be optimized at the t-th iteration. It can represent the new parameter values ​​of the model obtained after the t-th gradient update, that is, the weights that will be used in the next iteration; It can represent the learning rate, which controls the step size of gradient updates. It is a preset hyperparameter that affects the convergence speed and stability of the model. This can represent the bias correction value for the first-order moment estimate. This can represent the bias correction value for the second-order moment estimation. It can represent a small positive constant introduced to prevent numerical instability caused by a denominator of zero; for example, it can take the value 10. -8 Through the aforementioned time-series image modeling, GAF transformation, and convolutional neural network classification process, it is possible to automatically identify hidden abnormal patterns in the input operational data, and simultaneously provide equipment location results and fault mode classification results, thus constructing an intelligent auxiliary diagnostic path from data-driven to structural semantic analysis.

[0142] By mapping sensor time-series data into Gram difference field images through polar angle transformation, and using convolutional neural networks for high-dimensional feature extraction and classification output, the original one-dimensional operational data is given the ability to express spatial topological structure. This enhances the model's ability to learn weak anomaly trends, nonlinear features, and interaction patterns between devices, thereby improving the model's diagnostic stability and interpretability under complex working conditions.

[0143] In one example embodiment of this disclosure, the future failure prediction results of a multi-stage geothermal well cluster system can be determined by using equipment wear and tear data through the following steps, specifically including:

[0144] Equipment wear and tear data can be input into a pre-trained equipment life prediction model based on a long short-term memory network to generate the remaining lifespan of each piece of equipment in a multi-level geothermal well cluster system under different output load percentages. Equipment wear and tear data can also be input into a pre-trained future failure prediction model to estimate the future failure points and types of the multi-level geothermal well cluster system. The remaining lifespan of the equipment under different output load percentages, as well as the future failure points and types of the multi-level geothermal well cluster system, can be used as the future failure prediction results.

[0145] The equipment life prediction model can employ a Long Short-Term Memory (LSTM) network as its primary modeling structure. This structure possesses the ability to model long-term dependencies and can be used to extract latent, slowly changing characteristics and nonlinear dynamic patterns during equipment degradation. The model structure can consist of an input layer, several LSTM hidden unit layers, and a fully connected regression output layer. Within each LSTM unit, historical information is selectively remembered and updated through memory units and forgetting gates. The model training process can employ supervised learning. Training samples consist of historical equipment operating cycle sequences and known failure times. The loss function can be either mean squared error (MSE) or a weighted regression loss function to penalize high-error prediction points. During model deployment, the system inputs current equipment wear data into the model and outputs an estimated remaining lifespan of the equipment at the current or given load level. Lifespan is calculated at various load percentages (e.g., 50%, 70%, 90%) to generate a "load-lifespan" curve. This curve can be used to assist in developing operational strategies under various load scenarios and support the development of load-aware preventative maintenance plans.

[0146] The future fault prediction model is a multi-label, multi-time-window classification prediction network, designed to predict the device nodes most likely to fail and their fault types within a future period before a fault occurs. The model structure of the future fault prediction model can be the same as that of the fault identification and classification model in this embodiment; please refer to the model structure and training process of the fault identification and classification model, which will not be elaborated here.

[0147] The lifetime curve output by the LSTM model and the fault prediction events output by the classification model can be merged and encoded to form a predictive status record for each device. This record may include, for example, device ID, current load level, RUL curve slope, most recent possible failure time window, expected fault type, and prediction confidence level. This result can be directly used as input to the system operation and maintenance strategy generation module to support the development of optimized load scheduling, start-up and shutdown strategies, and maintenance window arrangements based on remaining lifetime thresholds, failure probability thresholds, or risk ranking mechanisms. This allows for the construction of a predictive closed-loop logic from device degradation modeling and fault event prediction to maintenance task execution.

[0148] By constructing a lifetime prediction model based on long short-term memory networks and supplementing it with a future fault type classification prediction model, we can realize trend modeling of the future health status of equipment and early warning of potential fault events. This not only generates a lifetime assessment curve reflecting the relationship between operating load and remaining lifetime, but also provides high-confidence prediction results of future fault points and types, enabling operation and maintenance strategies to have predictive capabilities and effectively transforming from "fault response" to "predictive assurance".

[0149] Optional, can be done through Figure 4 The steps in the document enable the generation of system operation and maintenance strategies based on equipment wear and tear data, current fault locations, current fault types, and future fault prediction results. Figure 4 As shown, it can specifically include:

[0150] Step S410: Determine the fault severity score for each device based on the device wear and tear data, the current fault point, and the current fault type;

[0151] Step S420: Determine the failure rate score of each device based on the equipment wear and tear data, the current failure point, and the future failure prediction results;

[0152] Step S430: Based on the coverage of the monitoring equipment and the diagnostic system corresponding to the multi-level geothermal well group system, determine the fault detection score of each device;

[0153] Step S440: Determine the risk order value through the fault severity score, the fault occurrence rate score, and the fault detectability score;

[0154] Step S450: Sort the maintenance tasks corresponding to each device according to the risk order value, and generate a system operation and maintenance strategy.

[0155] The fault severity score is a quantitative indicator used to measure the impact of an identified fault in a device on system operation. Its calculation is based on multiple factors, including the fluctuation range of key operating parameters of the device before and after the fault, the weight of the subsystem in the energy transfer path, and the system's sensitivity to power supply stability after the fault. In practical implementation, the severity function can be designed as a comprehensive function. Input variables include the simulation-measured deviation value of the current device (such as temperature residual, flow drop magnitude), load level, and redundancy replacement capability; the output is a normalized value of 0 to 1, with a larger value indicating a more significant impact of the fault on system performance. For example, when a main water pump experiences a flow drop and the system lacks a backup pump, its severity score will be significantly higher than that of auxiliary equipment in standby mode.

[0156] The failure incidence rate score quantifies the probability of a device failing again within its future operating cycle. It is a failure probability assessment value generated based on a statistical learning model or a time series prediction model. In its implementation, it can utilize the output of the future failure prediction model in this embodiment. Specifically, it analyzes the failure probability value of the device within the future prediction window and weights the distance between the current time point and the expected future failure time to construct a risk curve integral as the score. If the device's LSTM prediction result indicates that its remaining lifespan is close to zero, or if the failure type probability output by the future classification model exceeds the confidence threshold, the corresponding incidence rate score will approach 1; if the device's current state is stable and its historical degradation trend is gradual, the score will approach 0.

[0157] The fault detection score is used to evaluate the system's ability to detect a fault in a timely and accurate manner when the equipment malfunctions. In practice, this involves statistically analyzing the physical location of the equipment and the coverage density of its subsystems within a sensor distribution map to confirm whether key indicators are being monitored in real-time by the main control system or auxiliary sensors. Secondly, it assesses the observability of the equipment's characteristic variables under fault conditions, such as drastic pressure fluctuations, temperature jumps, or abnormal flow rates, to determine if these reflect potential anomalies. Thirdly, it combines the diagnostic model's accuracy in identifying historical fault samples to assess the equipment's sensitivity within the existing intelligent diagnostic system. The fault detection score is generated by considering all these factors. A lower score indicates a greater difficulty in early detection of the equipment's fault tendency, suggesting that manual inspections or enhanced monitoring methods should be prioritized.

[0158] Risk order values ​​can be determined using fault severity scores, fault occurrence scores, and fault detectability scores. For example, risk order values ​​can be calculated using the following formula:

[0159] ;

[0160] in, This can represent the risk order value of the i-th device. This can represent the severity score of the fault of the i-th device. This can represent the failure rate score of the i-th device. This can represent the fault detectability score of the i-th device. To control the consistency of the scoring scale, each score value is normalized, and the value range is usually set to 0 to 10, or 0 to 1. If a device fault simultaneously has high impact, high probability of occurrence, and low detectability, its RPN value will increase significantly, and it will be prioritized in subsequent maintenance tasks.

[0161] Maintenance tasks for each device can be sorted based on risk priority values ​​to generate system operation and maintenance strategies. This sorting process can follow descending order of RPN values, generating maintenance task queues sequentially from highest to lowest. System-level maintenance execution plans are then generated by considering constraints such as resource requirements, manpower, and availability within time windows for each task. A maintenance strategy optimization engine can be introduced during strategy generation to prioritize preventative maintenance or replacement strategies for high RPN devices and implement inspection or delayed intervention strategies for medium- and low RPN devices, thereby maximizing overall system availability. The final system operation and maintenance strategy can be output as a structured instruction list, containing metadata such as the target device to be maintained, estimated response time, suggested maintenance methods, scheduling priority, and RPN score, supporting automatic task dispatch and execution by downstream operation and maintenance platforms or scheduling centers.

[0162] By quantifying and scoring three dimensions—fault severity, fault occurrence rate, and fault detection rate—and integrating them into a single RPN value, a normalized comparison of risk levels for different devices under different health conditions can be achieved. This can be used to guide the prioritization of maintenance tasks, avoid situations where improper resource allocation leads to delays in the repair of critical equipment, and significantly improve the operability and accuracy of maintenance plans.

[0163] In one example embodiment of this disclosure, the optimization of system operation and maintenance strategies can be achieved through the following steps, which may specifically include:

[0164] With the goals of meeting minimum energy supply requirements and maximizing equipment lifespan, a multi-objective optimization model is established, taking equipment capacity and health status as constraints. A non-dominated sorting genetic algorithm is used to solve the multi-objective optimization model to determine the target equipment operation strategy. Based on the target equipment operation strategy, the system operation and maintenance strategy is optimized to obtain the optimized system operation and maintenance strategy. The optimized system operation and maintenance strategy is then used to control each piece of equipment in the multi-level geothermal well group system in real time.

[0165] The multi-objective optimization model aims to balance the operational lifespan and health status of key equipment while ensuring continuous power supply to the system, thereby achieving system-level operational stability and long-term benefits. The minimum energy supply demand is the basic objective, constrained by the requirement that the total output of the system in each time period should not be less than the predicted user load. Maximizing equipment lifespan is achieved by minimizing the cumulative degradation of each device within the scheduling cycle, typically calculated based on equipment lifespan models (e.g., load-life curves).

[0166] In the construction of a multi-objective optimization model, multiple operational constraints can be incorporated into the optimization solution framework. For example, operational constraints can be equipment capacity limits, i.e., at any given time, the output of a single piece of equipment must not exceed its maximum allowable output under its current health state, and should be greater than or equal to its minimum stable output; operational constraints can also be start-stop limit, i.e., the number of start-stop cycles for each piece of equipment within a scheduling cycle must not exceed the equipment durability threshold, which is achieved by recording the output status changes at continuous moments; operational constraints can also be equipment health state limits, i.e., at any time during the scheduling process, the cumulative operational degradation of the equipment must not exceed its current remaining service life, which is usually measured by an integral model or a recursive degradation function.

[0167] For example, a multi-objective optimization model can be represented by the following set of relations:

[0168]

[0169] in, The objective function representing the energy supply-demand imbalance is... This represents the total runtime of the system, and it seeks the longest stable operating cycle while satisfying power supply constraints. This represents the system heat supply at time t. This represents the heat demand at time t. The objective function representing system damage is... This represents the instantaneous damage rate at time t. This represents the geothermal water flow rate at time t. , These represent the minimum and maximum safe flow limits for the system's pump equipment, respectively. Indicates the rate of change of flow rate. This represents the upper limit of the rate of change of flow rate, and is the safe adjustment rate of the equipment. , Let represent the inlet water temperature and the outlet water temperature at time t, respectively. , These represent the minimum allowable inlet water temperature and the maximum allowable outlet water temperature, respectively.

[0170] After establishing the model, the Non-dominated Sorting Genetic Algorithm II (NSGA-II) can be used to solve the multi-objective optimization model and determine the target equipment operation strategy. NSGA-II is a population optimization algorithm based on the principle of multi-objective evolution. Its core feature is that it can simultaneously handle multiple conflicting objective functions and output a set of non-dominated optimal solutions. The non-dominated sorting genetic algorithm is a common technique in this field, and its solution process will not be described in detail here.

[0171] The target equipment operation strategy refers to the set of executable scheduling schemes and control commands formulated for each operating device in the system, based on the solution results of a multi-objective optimization model, under the premise of meeting system energy supply requirements and equipment lifespan constraints. The target equipment operation strategy takes into account both system-level operational objectives and device-level health status, achieving a dynamic balance between system safety, economy, and continuity. The system operation and maintenance strategy can be optimized based on the target equipment operation strategy to obtain an optimized system operation and maintenance strategy. The optimized system operation and maintenance strategy can include the start / stop status, output level, load ratio, and parallel switching logic of each device in each scheduling period. The optimized system operation and maintenance strategy is mapped into the system's real-time control logic and dynamically corrected in conjunction with the fault-risk devices and available resources identified in the operation and maintenance plan. Specifically, the optimized system operation and maintenance strategy not only covers the scheduling logic of critical equipment but can also include backup plans for some non-critical equipment, thereby improving the system's redundancy and fault tolerance capabilities.

[0172] By modeling the system's energy supply satisfaction and equipment lifespan maximization as a multi-objective optimization problem and solving it using the Non-Dominated Sorting Genetic Algorithm (NSGA-II), an operating strategy that can simultaneously satisfy supply and demand balance and equipment durability requirements is obtained. The optimization model outputs the optimal scheduling solution while taking into account equipment operating constraints (capacity boundaries, start-stop frequency, and health status), effectively improving the overall coordination, flexibility, and economy of the system's operation, and meeting the needs of complex energy networks for intelligent scheduling and multi-objective collaborative management.

[0173] Figure 5 The illustration schematically shows a process framework diagram for fault identification and maintenance of a multi-stage geothermal well group system according to some embodiments of the present disclosure.

[0174] refer to Figure 5As shown, step S510, data acquisition and preprocessing, can be used to collect historical and real-time operating data of various equipment in a multi-level geothermal well group system, including but not limited to parameters such as temperature, pressure, flow rate, current, voltage, and start / stop status; and the collected data is cleaned, standardized, and time-series normalized to form the data input format required for subsequent model analysis; step S520, constructing a fault identification model, by extracting features from the processed equipment operating time-series data, a fault identification classification model is established, and the estimated fault type and estimated fault point in the current system are identified and output, which is used to narrow the diagnostic scope of subsequent simulation analysis and improve identification efficiency and accuracy; step S530, multi-level collaborative mechanism modeling, after identifying the fault point, coarse-grained subsystem models and fine-grained equipment models are constructed respectively, and hierarchical simulation is carried out on the subsystem and equipment to which the fault point belongs; By comparing the system simulation results with the equipment simulation results, the current fault point and type are finally confirmed and located. Step S540: Construct a fault prediction model. Based on equipment wear and tear data, a prediction model is constructed, including equipment health modeling, life prediction curve modeling, and future fault event window identification. The model outputs future fault prediction results, including remaining useful life (RUL), fault trends, and estimated fault types, providing a forward-looking basis for subsequent operation and maintenance strategies. Step S550: Develop a guaranteed operation strategy. Based on fault diagnosis results and prediction information, a corresponding preventative maintenance plan is generated, a multi-objective optimization model that meets energy supply demand and lifespan optimization is constructed, and scheduling solutions are performed in conjunction with operational constraints such as equipment capacity, start / stop frequency, and health status. The target equipment operation strategy and system optimization operation and maintenance plan are output. Step S560: Execute the strategy and achieve real-time monitoring. The generated system operation and maintenance strategy is sent to the equipment controller for execution, controlling the start / stop status and output allocation of each device. Simultaneously, the monitoring system provides dynamic feedback on the system response status during execution, triggering rolling optimization or strategy correction when necessary to achieve closed-loop adjustment of operation control.

[0175] Figure 6 The diagram illustrates a framework of a mechanistic model corresponding to a multi-stage geothermal well cluster system according to some embodiments of the present disclosure.

[0176] refer to Figure 6 As shown, the mechanistic model 600 of a multi-stage geothermal well group system may include at least a well group subsystem 610, a power subsystem 620, a heat extraction subsystem 630, a pipeline subsystem 640, and a hydraulic distribution subsystem 650. Wherein:

[0177] The well group subsystem 610 may include shallow geothermal well groups, medium-deep geothermal well groups, and deep geothermal well groups. Each well group consists of several numbered geothermal wells, corresponding to different burial depths of geothermal resources. The well group subsystem is the heat source input end of the entire system, used to extract geothermal energy from different strata to supply subsequent heat extraction and power systems.

[0178] The power subsystem 620 may include a wellside pump system, a primary circulation system, and a water supply system. The wellside pumps at each depth level (shallow, medium-deep, and deep) are connected to their corresponding heat pump equipment and primary circulation pumps to drive the flow of geothermal fluids and form a closed-loop energy transfer path. In addition, a water supply pump and water supply loop are provided to maintain the hydraulic balance of the system and avoid the risk of low-pressure operation or cavitation. The power subsystem is the key object of analysis of coarse-grained system operating parameters (such as pressure difference, temperature rise, and flow coupling) in simulation modeling.

[0179] The heat extraction subsystem 630 may include various levels of heat pump devices, thermal storage equipment, air source heat pumps, cooling towers, and plate heat exchangers, etc., for the extraction, conversion, storage, and distribution of heat from geothermal fluids. The operating status of this subsystem has a significant impact on the system's energy supply efficiency and equipment failure characteristics, making it an important object for fine-grained simulation modeling and equipment health status assessment.

[0180] The hydraulic distribution subsystem 650 may include hydraulic control components such as a water collector, a water distributor, and a softened water tank, which are used to collect, distribute, and regulate the water quality of the heat transfer medium within the system. In the simulation, it is used to set boundary conditions and node connection relationships to ensure the consistency of material transport paths between subsystems.

[0181] The pipeline subsystem 640 may include equipment such as valves, heat meters, water meters, and electricity meters in the pipeline network. It is used for data acquisition, flow measurement, and hydraulic control. This system is the main source of operational data and the primary input source for equipment operation and wear data. It can be used to construct data-driven life prediction and fault analysis models. The subsystems are coupled through boundary nodes. In coarse-grained modeling, a centralized module interconnection structure can be used, while in fine-grained modeling, the subsystems can be split according to the accuracy of the equipment-level model. Each subsystem interfaces with the simulation platform through a unified interface, enabling data input, state output, and boundary transfer for the mechanistic model.

[0182] It should be noted that although the steps of the method in this disclosure are described in a specific order in the accompanying drawings, this does not require or imply that the steps must be performed in that specific order, or that all the steps shown must be performed to achieve the desired result. Additional or alternative steps may be omitted, multiple steps may be combined into one step, and / or a step may be broken down into multiple steps.

[0183] Furthermore, this example embodiment also provides a fault identification and maintenance device for a multi-level geothermal well group system. (Refer to...) Figure 7 As shown, the fault identification and maintenance device 700 for the multi-level geothermal well group system includes: an equipment data acquisition module 710, an equipment fault identification module 720, an equipment fault prediction module 730, an equipment operation and maintenance strategy generation module 740, and an equipment maintenance module 750. Wherein:

[0184] The equipment data acquisition module 710 is used to collect equipment operation data of each level of equipment in the multi-level geothermal well group system in real time, and to simultaneously acquire equipment wear and tear data of each of the equipment.

[0185] The equipment fault identification module 720 is used to determine the current fault point and the current fault type of the multi-level geothermal well group system by using the equipment operation data and the pre-established coarse-grained simulation model and fine-grained simulation model.

[0186] The equipment failure prediction module 730 is used to determine the future failure prediction results of the multi-stage geothermal well group system based on the equipment wear and tear data.

[0187] The equipment operation and maintenance strategy generation module 740 is used to generate a system operation and maintenance strategy based on the equipment wear and tear data, the current fault point, the current fault type, and the future fault prediction results.

[0188] The equipment maintenance module 750 is used to control each device in the multi-level geothermal well group system in real time through the system operation and maintenance strategy, and to push the current fault point, the current fault type and the system operation and maintenance strategy to the operation and maintenance object.

[0189] In some exemplary embodiments of this disclosure, based on the foregoing scheme, the coarse-grained simulation model uses each level of subsystem in the multi-level geothermal well group system as the basic unit to describe the energy and mass transfer relationships between the subsystems; the fine-grained simulation model uses each independent device in the subsystem as the basic unit to describe the detailed operating characteristics of the device; the device fault identification module 720 is configured as follows:

[0190] The equipment operation data is input into the coarse-grained simulation model to perform collaborative simulation calculations on each level of subsystems in the multi-level geothermal well group system, and to determine the system simulation operation results.

[0191] By comparing the system simulation results with the equipment operation data, we can identify the subsystems to be analyzed that have abnormal equipment operation data.

[0192] The equipment operation data corresponding to the subsystem to be analyzed is input into the fine-grained simulation model so as to perform collaborative simulation calculations on each device in the subsystem to be analyzed through the fine-grained simulation model and determine the equipment simulation operation results.

[0193] By comparing the simulation results and the operating data of the equipment, the current fault point and the current fault type of the multi-level geothermal well group system are determined.

[0194] In some example embodiments of this disclosure, based on the foregoing scheme, the equipment fault identification module 720 is further configured to:

[0195] The equipment operation data is input into a pre-trained fault identification and classification model to determine the estimated fault point and estimated fault type of the multi-level geothermal well group system.

[0196] Based on the estimated fault point and the estimated fault type, data items to be compared are filtered from the system simulation results, the equipment simulation results, and the equipment operation data.

[0197] In some example embodiments of this disclosure, based on the foregoing scheme, the equipment fault identification module 720 is further configured to:

[0198] A sensor time series is constructed based on the device operation data, and the sensor time series is time normalized to obtain the time normalization result.

[0199] The time normalization result is mapped to an angle in a polar coordinate system, and the cosine value of the angle difference is calculated to generate the Gram angle difference field matrix.

[0200] The convolutional neural network in the fault identification and classification model extracts features from the Gram angle difference field matrix, and inputs the extracted feature map into the classification layer to determine the estimated fault point and estimated fault type corresponding to the multi-level geothermal well group system.

[0201] In some example embodiments of this disclosure, based on the foregoing scheme, the equipment fault prediction module 730 is configured as follows:

[0202] The equipment wear data is input into a pre-trained equipment life prediction model based on a long short-term memory network to generate the remaining life of each piece of equipment in the multi-level geothermal well group system under different output load percentages.

[0203] The equipment wear data is input into a pre-trained future failure prediction model to estimate the future failure points and future failure types of the multi-level geothermal well group system.

[0204] The remaining service life of the equipment under different output load percentages, as well as the future failure points and future failure types of the multi-stage geothermal well group system, are used as the future failure prediction results.

[0205] In some example embodiments of this disclosure, based on the foregoing scheme, the device operation and maintenance strategy generation module 740 is configured as follows:

[0206] Based on the equipment wear and tear data, the current fault point, and the current fault type, determine the fault severity score for each device;

[0207] Based on the equipment wear and tear data, the current fault points, and the future fault prediction results, a fault occurrence rate score is determined for each piece of equipment.

[0208] Based on the coverage of the monitoring equipment and the diagnostic system corresponding to the multi-level geothermal well group system, the fault detection score of each device is determined.

[0209] The risk order values ​​are determined by the fault severity score, the fault occurrence rate score, and the fault detectability score.

[0210] The maintenance tasks corresponding to each device are sorted based on the risk priority values ​​to generate a system operation and maintenance strategy.

[0211] In some example embodiments of this disclosure, based on the foregoing scheme, the equipment operation and maintenance strategy generation module 740 is further configured to:

[0212] To meet the minimum energy supply demand and maximize equipment lifespan, a multi-objective optimization model is established with equipment capacity, start-up and shutdown frequency, and equipment health status as constraints.

[0213] The multi-objective optimization model is solved using a non-dominated sorting genetic algorithm to determine the target equipment operation strategy;

[0214] The system operation and maintenance strategy is optimized based on the target device operation strategy to obtain the optimized system operation and maintenance strategy;

[0215] The optimized system operation and maintenance strategy is used to control each device in the multi-level geothermal well group system in real time.

[0216] In some example embodiments of this disclosure, based on the foregoing scheme, the multi-objective optimization model is as follows:

[0217]

[0218] in, This represents the energy supply and demand deviation function. This represents the total runtime of the system, and it seeks the longest stable operating cycle while satisfying power supply constraints. This represents the total system energy demand in the t-th time period. This represents the total number of schedulable devices in the system. This represents the power output of the i-th device during time period t. This represents the output load of the equipment during time period t. This indicates the minimum safe output load of the equipment. Indicates the maximum allowable output load of the equipment. This indicates the start / stop status of the device during the time period t+1, with a value of 1 or 0. This indicates the start / stop status of the device during time period t, with a value of 1 or 0. This indicates the maximum number of times the equipment is allowed to start and stop within the entire scheduling cycle. This represents the cumulative lifespan loss of the equipment due to load operation over the entire operating cycle. This indicates the upper limit of the remaining service life of the equipment.

[0219] The specific details of each module of the fault identification and maintenance device for the multi-level geothermal well group system mentioned above have been described in detail in the corresponding fault identification and maintenance method for the multi-level geothermal well group system, so they will not be repeated here.

[0220] It should be noted that although several modules or units of the fault identification and maintenance device for multi-stage geothermal well cluster systems have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this disclosure, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided into multiple modules or units.

[0221] Furthermore, in an exemplary embodiment of this disclosure, an electronic device capable of implementing the above-described method for fault identification and maintenance of a multi-level geothermal well group system is also provided.

[0222] Those skilled in the art will understand that various aspects of this disclosure can be implemented as a system, method, or program product. Therefore, various aspects of this disclosure can be embodied in the following forms: a completely hardware embodiment, a completely software embodiment (including firmware, microcode, etc.), or an embodiment combining hardware and software aspects, collectively referred to herein as a "circuit," "module," or "system."

[0223] The following reference Figure 8 To describe an electronic device 800 according to such an embodiment of the present disclosure. Figure 8The electronic device 800 shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments disclosed herein.

[0224] like Figure 8 As shown, the electronic device 800 is presented in the form of a general-purpose computing device. The components of the electronic device 800 may include, but are not limited to: at least one processing unit 810, at least one storage unit 820, a bus 830 connecting different system components (including storage unit 820 and processing unit 810), and a display unit 840.

[0225] The storage unit stores program code that can be executed by the processing unit 810, causing the processing unit 810 to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure. For example, the processing unit 810 can perform actions such as... Figure 1 Step S110 involves real-time acquisition of equipment operation data from each level of equipment in the multi-level geothermal well group system, and simultaneous acquisition of equipment wear and tear data for each device. Step S120 involves determining the current fault point and current fault type of the multi-level geothermal well group system using the equipment operation data and pre-established coarse-grained and fine-grained simulation models. Step S130 involves determining the future fault prediction result of the multi-level geothermal well group system using the equipment wear and tear data. Step S140 involves generating a system operation and maintenance strategy based on the equipment wear and tear data, the current fault point, the current fault type, and the future fault prediction result. Step S150 involves real-time control of each device in the multi-level geothermal well group system using the system operation and maintenance strategy, and pushing the current fault point, the current fault type, and the system operation and maintenance strategy to the operation and maintenance target.

[0226] Storage unit 820 may include readable media in the form of volatile storage units, such as random access memory (RAM) 821 and / or cache memory 822, and may further include read-only memory (ROM) 823.

[0227] The storage unit 820 may also include a program / utility 824 having a set (at least one) of program modules 825, including but not limited to: an operating system, one or more application programs, other program modules, and program data, each or some combination of these examples may include an implementation of a network environment.

[0228] Bus 830 can represent one or more of several types of bus structures, including a memory cell bus or memory cell controller, a peripheral bus, a graphics acceleration port, a processing unit, or a local bus using any of the various bus structures.

[0229] Electronic device 800 can also communicate with one or more external devices 870 (e.g., keyboard, pointing device, Bluetooth device, etc.), and with one or more devices that enable a user to interact with electronic device 800, and / or with any device that enables electronic device 800 to communicate with one or more other computing devices (e.g., router, modem, etc.). This communication can be performed via input / output (I / O) interface 850. Furthermore, electronic device 800 can also communicate with one or more networks (e.g., local area network (LAN), wide area network (WAN), and / or public networks, such as the Internet) via network adapter 860. As shown, network adapter 860 communicates with other modules of electronic device 800 via bus 830. It should be understood that, although not shown in the figures, other hardware and / or software modules can be used in conjunction with electronic device 800, including but not limited to: microcode, device drivers, redundant processing units, external disk drive arrays, RAID systems, tape drives, and data backup storage systems.

[0230] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, terminal device, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0231] In exemplary embodiments of this disclosure, a computer-readable storage medium is also provided, on which a program product capable of implementing the methods described above is stored. In some possible embodiments, various aspects of this disclosure may also be implemented as a program product including program code that, when the program product is run on a terminal device, causes the terminal device to perform the steps described in the "Exemplary Methods" section of this specification according to various exemplary embodiments of this disclosure.

[0232] refer to Figure 9As shown, a program product 900 for implementing the above-described method for fault identification and maintenance of a multi-stage geothermal well group system according to an embodiment of the present disclosure is described. It may employ a portable compact disk read-only memory (CD-ROM) and include program code, and may run on a terminal device, such as a personal computer. However, the program product of the present disclosure is not limited thereto. In this document, the readable storage medium may be any tangible medium containing or storing a program that may be used by or in conjunction with an instruction execution system, apparatus, or device.

[0233] The program product may employ any combination of one or more readable media. A readable medium may be a readable signal medium or a readable storage medium. A readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of readable storage media (a non-exhaustive list) include: an electrical connection having one or more wires, a portable disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof.

[0234] Computer-readable signal media may include data signals propagated in baseband or as part of a carrier wave, carrying readable program code. Such propagated data signals may take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. A readable signal medium may also be any readable medium other than a readable storage medium, capable of sending, propagating, or transmitting programs for use by or in conjunction with an instruction execution system, apparatus, or device.

[0235] The program code contained on the readable medium may be transmitted using any suitable medium, including but not limited to wireless, wired, optical fiber, RF, etc., or any suitable combination thereof.

[0236] Program code for performing the operations of this disclosure can be written in any combination of one or more programming languages, including object-oriented programming languages ​​such as Java and C++, and conventional procedural programming languages ​​such as C or similar languages. The program code can execute entirely on the user's computing device, partially on the user's computing device, as a standalone software package, partially on the user's computing device and partially on a remote computing device, or entirely on a remote computing device or server. In cases involving remote computing devices, the remote computing device can be connected to the user's computing device via any type of network, including a local area network (LAN) or a wide area network (WAN), or it can be connected to an external computing device (e.g., via the Internet using an Internet service provider).

[0237] Furthermore, the above figures are merely illustrative of the processes included in the method according to exemplary embodiments of this disclosure and are not intended to be limiting. It is readily understood that the processes shown in the above figures do not indicate or limit the temporal order of these processes. Additionally, it is readily understood that these processes may be executed synchronously or asynchronously, for example, in multiple modules.

[0238] From the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this disclosure can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the methods according to the embodiments of this disclosure.

[0239] Other embodiments of this disclosure will readily occur to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not disclosed herein. The specification and embodiments are to be considered exemplary only, and the true scope and spirit of this disclosure are indicated by the claims.

[0240] It should be understood that this disclosure is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this disclosure is limited only by the appended claims.

Claims

1. A method for fault identification and maintenance of a multi-stage geothermal well cluster system, characterized in that, The method comprises: real-time acquisition of equipment operation data of each level of equipment in the multi-level geothermal well group system, and synchronous acquisition of equipment loss and use data of each equipment; determination of the current fault point and the current fault type of the multi-level geothermal well group system through the equipment operation data, and a pre-established coarse-grained simulation model and a fine-grained simulation model; determination of the future fault prediction result of the multi-level geothermal well group system through the equipment loss and use data; generation of a system operation and maintenance strategy according to the equipment loss and use data, the current fault point, the current fault type, and the future fault prediction result; real-time control of each equipment in the multi-level geothermal well group system through the system operation and maintenance strategy, and pushing of the current fault point, the current fault type, and the system operation and maintenance strategy to an operation and maintenance object; wherein the coarse-grained simulation model takes each level of subsystem in the multi-level geothermal well group system as a basic unit, and describes the energy and material transfer relationship between each subsystem; and the fine-grained simulation model takes each independent equipment in a subsystem as a basic unit, and describes the detailed operation characteristics of the equipment; the determination of the current fault point and the current fault type of the multi-level geothermal well group system through the equipment operation data, and the pre-established coarse-grained simulation model and fine-grained simulation model comprises: inputting the equipment operation data into the coarse-grained simulation model to perform cooperative simulation calculation on each level of subsystem in the multi-level geothermal well group system through the coarse-grained simulation model, and determining a system simulation operation result; comparing the system simulation operation result and the equipment operation data to determine a to-be-analyzed subsystem with abnormal equipment operation data; inputting the equipment operation data corresponding to the to-be-analyzed subsystem into the fine-grained simulation model to perform cooperative simulation calculation on each equipment in the to-be-analyzed subsystem through the fine-grained simulation model, and determining an equipment simulation operation result; comparing the equipment simulation operation result and the equipment operation data to determine the current fault point and the current fault type of the multi-level geothermal well group system.

2. The method of claim 1, wherein, The method further comprises: inputting the equipment operation data into a pre-trained fault identification classification model to determine an estimated fault point and an estimated fault type corresponding to the multi-level geothermal well group system; screening to-be-compared data items from the system simulation operation result, the equipment simulation operation result, and the equipment operation data through the estimated fault point and the estimated fault type.

3. The method of claim 2, wherein, inputting the equipment operation data into a pre-trained fault identification classification model to determine an estimated fault point and an estimated fault type corresponding to the multi-level geothermal well group system, comprises: constructing a sensor time series based on the equipment operation data, and performing time normalization processing on the sensor time series to obtain a time normalization result; mapping the time normalization result into an angle in a polar coordinate system, and calculating a cosine value of the angle difference to generate a Gram angle difference field matrix; The convolutional neural network in the fault identification classification model is used for feature extraction on the Gram angle difference field matrix, and the extracted feature map is input into a classification layer to determine an estimated fault point and an estimated fault type corresponding to the multi-stage geothermal well group system.

4. The method of claim 1, wherein, The future fault prediction result of the multi-stage geothermal well group system is determined by the equipment wear and tear usage data, including: The equipment wear and tear usage data is input into a pre-trained equipment life prediction model based on a long short-term memory network to generate the remaining service life of each equipment in the multi-stage geothermal well group system under different output load percentages; The equipment wear and tear usage data is input into a pre-trained future fault prediction model to estimate the future fault point and the future fault type of the multi-stage geothermal well group system; The remaining service life of the equipment under different output load percentages, and the future fault point and the future fault type of the multi-stage geothermal well group system are used as the future fault prediction result.

5. The method of claim 1, wherein, The system operation and maintenance strategy is generated according to the equipment wear and tear usage data, the current fault point, the current fault type, and the future fault prediction result, including: The fault severity score of each equipment is determined according to the equipment wear and tear usage data, the current fault point, and the current fault type; The fault occurrence rate score of each equipment is determined according to the equipment wear and tear usage data, the current fault point, and the future fault prediction result; The fault detection degree score of each equipment is determined based on the monitoring equipment coverage range and the diagnostic system coverage range corresponding to the multi-stage geothermal well group system; The risk order value is determined by the fault severity score, the fault occurrence rate score, and the fault detection degree score; The maintenance tasks corresponding to each equipment are sorted in combination with the risk order value to generate the system operation and maintenance strategy.

6. The method of claim 5, wherein, The method further includes: A multi-objective optimization model is established by taking the equipment capacity and the equipment health state as constraint conditions, with the goal of meeting the minimum energy supply demand and maximizing the equipment service life; The multi-objective optimization model is solved by using a non-dominated sorting genetic algorithm to determine a target equipment operation strategy; The system operation and maintenance strategy is optimized based on the target equipment operation strategy to obtain an optimized system operation and maintenance strategy; Each equipment in the multi-stage geothermal well group system is controlled in real time by the optimized system operation and maintenance strategy.

7. The method of claim 6, wherein, The multi-objective optimization model is: wherein, represents the energy supply-demand deviation objective function, represents the total running time of the system, and the longest stable running period is sought under the premise of meeting the energy supply constraint, represents the system heat supply at the tth moment, represents the demand heat at the tth moment, represents the system damage objective function, represents the instantaneous damage rate at the tth moment, represents the geothermal water flow at the tth moment, , respectively represent the minimum and maximum safe flow limits of the system pump equipment, represents the flow rate of change, represents the upper limit value of the flow rate of change, which is the safe adjustment rate of the equipment, , respectively represent the inlet water temperature and outlet water temperature at the tth moment, , respectively represent the minimum value allowed for the inlet water temperature, and the maximum value allowed for the outlet water temperature.

8. A fault identification and maintenance system for a multi-stage geothermal well cluster system, the system comprising: including: An equipment data acquisition module is configured to acquire equipment operation data of each equipment in the multi-stage geothermal well group system in real time, and to synchronously acquire equipment wear and tear usage data of each equipment; An equipment fault identification module is configured to determine a current fault point and a current fault type of the multi-stage geothermal well group system by using the equipment operation data, and a pre-established coarse-grained simulation model and a fine-grained simulation model; An equipment fault prediction module is configured to determine a future fault prediction result of the multi-stage geothermal well group system by using the equipment wear and tear usage data; The device operation and maintenance policy generation module is configured to generate a system operation and maintenance policy according to the device wear and tear usage data, the current fault point, the current fault type, and the future fault prediction result. The device maintenance module is configured to perform real-time control on each device in the multi-stage geothermal well group system through the system operation and maintenance policy, and push the current fault point, the current fault type, and the system operation and maintenance policy to an operation and maintenance object. The coarse-grained simulation model takes each subsystem in the multi-stage geothermal well group system as a basic unit, and describes the energy and material transfer relationship between the subsystems. The device fault identification module is configured to: input the device operation data into the coarse-grained simulation model to perform cooperative simulation calculation on each subsystem in the multi-stage geothermal well group system through the coarse-grained simulation model, and determine a system simulation operation result; compare the system simulation operation result with the device operation data to determine a to-be-analyzed subsystem in which the device operation data is abnormal; input the device operation data corresponding to the to-be-analyzed subsystem into the fine-grained simulation model to perform cooperative simulation calculation on each device in the to-be-analyzed subsystem through the fine-grained simulation model, and determine a device simulation operation result; compare the device simulation operation result with the device operation data to determine a current fault point and a current fault type of the multi-stage geothermal well group system.

9. An electronic device, comprising: The device operation and maintenance policy generation module is configured to generate a system operation and maintenance policy according to the device wear and tear usage data, the current fault point, the current fault type, and the future fault prediction result. The device maintenance module is configured to perform real-time control on each device in the multi-stage geothermal well group system through the system operation and maintenance policy, and push the current fault point, the current fault type, and the system operation and maintenance policy to an operation and maintenance object. The coarse-grained simulation model takes each subsystem in the multi-stage geothermal well group system as a basic unit, and describes the energy and material transfer relationship between the subsystems. The device fault identification module is configured to: input the device operation data into the coarse-grained simulation model to perform cooperative simulation calculation on each subsystem in the multi-stage geothermal well group system through the coarse-grained simulation model, and determine a system simulation operation result; compare the system simulation operation result with the device operation data to determine a to-be-analyzed subsystem in which the device operation data is abnormal; input the device operation data corresponding to the to-be-analyzed subsystem into the fine-grained simulation model to perform cooperative simulation calculation on each device in the to-be-analyzed subsystem through the fine-grained simulation model, and determine a device simulation operation result; compare the device simulation operation result with the device operation data to determine a current fault point and a current fault type of the multi-stage geothermal well group system. The device operation and maintenance policy generation module is configured to generate a system operation and maintenance policy according to the device wear and tear usage data, the current fault point, the current fault type, and the future fault prediction result. The device maintenance module is configured to perform real-time control on each device in the multi-stage geothermal well group system through the system operation and maintenance policy, and push the current fault point, the current fault type, and the system operation and maintenance policy to an operation and maintenance object. The coarse-grained simulation model takes each subsystem in the multi-stage geothermal well group system as a basic unit, and describes the energy and material transfer relationship between the subsystems. The device fault identification module is configured to: input the device operation data into the coarse-grained simulation model to perform cooperative simulation calculation on each subsystem in the multi-stage geothermal well group system through the coarse-grained simulation model, and determine a system simulation operation result; compare the system simulation operation result with the device operation data to determine a to-be-analyzed subsystem in which the device operation data is abnormal; input the device operation data corresponding to the to-be-analyzed subsystem into the fine-grained simulation model to perform cooperative simulation calculation on each device in the to-be-analyzed subsystem through the fine-grained simulation model, and determine a device simulation operation result; compare the device simulation operation result with the device operation data to determine a current fault point and a current fault type of the multi-stage geothermal well group system. The device operation and maintenance policy generation module is configured to generate a system operation

Images