Medium and high speed generator set intelligent operation and maintenance system based on cloud edge cooperation

Abstract

The embodiment of the invention provides a medium and high speed generator set intelligent operation and maintenance system based on cloud edge collaboration, and the system comprises an intelligent operation and maintenance center end which is provided with a Tdensine time sequence database and an operation and maintenance data warehouse, is used for storing high-frequency operation data from edge equipment, and operates fault prediction, health assessment and control strategy algorithms. Intelligent decision support, visual interface, strategy management and remote control functions are provided, and man-machine interaction and multi-site collaborative operation and maintenance are realized; and a plurality of edge acquisition and control terminals which are distributed and deployed on a high-speed generator set in each vehicle and are responsible for collecting various kinds of field operation data and uploading the data to the cloud end, and receiving a cloud end control instruction and executing local adjustment at the same time. According to the embodiment of the invention, the real-time data processing of the edge side is combined with the big data analysis of the cloud end, and the comprehensive monitoring, real-time analysis, state prediction and intelligent control of the operation state of the generator set are realized while the operation and maintenance efficiency and the decision accuracy are improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent operation and maintenance technology for equipment, and in particular to an intelligent operation and maintenance system for medium and high speed generator sets based on cloud-edge collaboration. Background Technology

[0002] With the continuous advancement of intelligentization in mining power equipment, heavy-duty truck generator sets are developing towards high automation, intelligence, and unmanned operation. As a critical power source, diesel generator sets on heavy-duty trucks face enormous operational pressure in extreme environments such as high altitudes, frigid zones, and high dust levels, and their reliability directly affects the vehicle's transportation capacity and task execution efficiency. However, most generator set operation and maintenance management systems are still at a traditional level, mainly relying on parameter monitoring and alarm functions. That is, when a key parameter exceeds a preset threshold, the system will issue an alarm signal and record the corresponding information. However, this passive response-based fault management method has many problems:

[0003] 1) Delayed response: The system usually generates an alarm only after a fault occurs or worsens, lacking a real-time prediction mechanism, which leads to a delayed maintenance response.

[0004] 2) Lack of intelligent decision-making mechanism: Existing systems rely heavily on manual judgment and processing, and cannot automatically adjust operating strategies or conduct control interventions based on operating conditions and historical trends.

[0005] 3) Poor environmental adaptability: Under harsh conditions such as high altitude and strong load fluctuations, the existing system is unable to correct key thermal parameters according to environmental changes, affecting the accuracy of operation and maintenance judgments.

[0006] Therefore, heavy-duty truck generator sets urgently need an intelligent operation and maintenance system with real-time predictive analysis, remote intelligent control, and flexible adaptation to operating status. Summary of the Invention

[0007] In view of the above-mentioned problems in the prior art, the present invention provides an intelligent operation and maintenance system for medium and high speed generator sets based on cloud-edge collaboration to solve the technical problems of slow response, lack of intelligent decision-making mechanism and poor environmental adaptability in the prior art.

[0008] This invention provides an intelligent operation and maintenance system for medium- and high-speed generator sets based on cloud-edge collaboration, comprising:

[0009] The intelligent operation and maintenance center is equipped with a cloud data and algorithm server and a cloud operation and maintenance server. The cloud data and algorithm server is deployed with a Tdengine time series database and an operation and maintenance data warehouse to store high-frequency operation data from edge devices and run fault prediction, health assessment and control strategy algorithms to provide intelligent decision support. The cloud operation and maintenance server provides a visual interface, policy management and remote control functions to realize human-computer interaction and multi-site collaborative operation and maintenance.

[0010] Several edge data acquisition and control terminals are distributed and deployed on each vehicle-mounted medium and high speed generator set. The edge data acquisition and control terminal serves as the local control core, integrating HTTPS and Modbus multi-protocol communication capabilities. It is responsible for collecting various on-site operating data and uploading it to the cloud, while receiving cloud control commands and executing local adjustments.

[0011] The intelligent operation and maintenance center and several edge acquisition and control terminals achieve two-way data interaction and control command issuance through wireless communication.

[0012] In one embodiment, the cloud data and algorithm server is equipped with an algorithm service module and a data service module, and the cloud operation and maintenance server is equipped with an operation and maintenance service module. The algorithm service module is used to implement functions such as anomaly monitoring, parameter prediction, parameter correction, health assessment, and heat load calculation. The data service module is used to implement functions such as data cleaning, data storage management, real-time data query, historical data query, trend data query, and anomaly data management. The operation and maintenance service module is used to implement functions such as operation status diagram drawing, health score diagram drawing, anomaly prompts, fault records, alarm lists, and coolant filling prompts.

[0013] In one embodiment, the edge acquisition and control terminal is deployed with,

[0014] The oil monitoring module is used to collect lubrication status data, including lubricating oil viscosity, water activity, and metal particles.

[0015] Vibration monitoring module, used to monitor the vibration characteristics of key parts of generator set;

[0016] The generator set monitoring module is used to monitor various operating parameters of the engine, including temperature, pressure, and load, and to perform protective control.

[0017] In one embodiment, the algorithm service module has multivariate fusion state modeling and environmental correction functions. The system combines operating data with environmental factors including atmospheric pressure, temperature and altitude to construct a correction model, dynamically correct thermal parameters, and establish equipment health scoring models and operating condition prediction models to achieve multivariate fusion modeling, which is used to improve the diagnostic accuracy of key operating parameters, the consistency of horizontal comparison, and the robustness of the model to environmental disturbances.

[0018] In one embodiment, the vibration monitoring module has an embedded feature calculation unit for real-time processing of the acquired signals, extracting key feature parameters including vibration RMS value, power spectral density, kurtosis, and skewness, and constructing a device health index matrix.

[0019] In one embodiment, the edge acquisition and control terminal has functions including network interruption resume transmission, local policy caching, and remote configuration update.

[0020] In one embodiment, the system adopts a combined architecture of edge caching and cloud time-series database, which enables fast buffering and short-term response at the edge acquisition and control end, and enables large-capacity data storage and long-term trend analysis at the intelligent operation and maintenance center end.

[0021] When the system is running normally, the edge acquisition and control terminal is in cloud-edge collaborative mode, collecting various types of on-site operating data and uploading them to the cloud, while receiving cloud control commands and executing local adjustments;

[0022] When the system malfunctions, the edge acquisition and control terminal switches to edge autonomous mode to ensure that basic functions operate independently, and then resumes linkage after communication is restored.

[0023] In one embodiment, when the system identifies a significant risk, the system immediately enters an emergency response process, as follows:

[0024] The edge acquisition and control terminal automatically executes "load-limited operation" or "active shutdown" commands;

[0025] Generate event logs and status snapshots;

[0026] Send a high-priority alarm to the remote platform.

[0027] In one embodiment, the system supports local manual reset and remote expert intervention, allowing for manual intervention or strategy adjustments when necessary.

[0028] In one embodiment, the system supports the construction of short-term, medium-term, and long-term data warehouses in four time domains: 5 minutes, hours, days, and weeks, and supports differentiated sampling rate storage strategies.

[0029] Compared with existing technologies, the beneficial effects of the cloud-edge collaborative intelligent operation and maintenance system for medium and high-speed generator sets provided by this invention are as follows:

[0030] 1. By integrating multiple sensors and intelligent algorithms, and through a cloud-edge collaborative architecture, this invention combines real-time data processing at the edge with big data analysis in the cloud, thereby improving operational efficiency and decision-making accuracy while achieving comprehensive monitoring, real-time analysis, status prediction, and intelligent control of generator set operating status.

[0031] 2. By monitoring key indicators such as fuel inlet pressure and temperature, lubricating oil inlet pressure and temperature, cylinder exhaust temperature, generator stator and bearing temperature, and vibration parameters, and combining data analysis and machine learning models, this invention enables real-time assessment and early warning of equipment status. Compared with existing technologies, this invention can significantly improve the operational stability and safety of medium- and high-speed diesel generator sets in complex environments, and reduce failure rates and maintenance costs. Attached Figure Description

[0032] Figure 1 A schematic diagram of the composition architecture of a cloud-edge collaborative intelligent operation and maintenance system for medium and high-speed generator sets provided in an embodiment of the present invention;

[0033] Figure 2 This is a schematic diagram illustrating the boundary regression fitting model for the exhaust temperature of a certain cylinder of an engine in an intelligent operation and maintenance system for medium and high-speed generator sets based on cloud-edge collaboration, provided as an embodiment of the present invention. Detailed Implementation

[0034] To enable those skilled in the art to better understand the technical solution of the present invention, the present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0035] Various embodiments and features of this application are described herein with reference to the accompanying drawings.

[0036] These and other features of this application will become apparent from the following description of preferred forms of embodiments given as non-limiting examples, with reference to the accompanying drawings.

[0037] It should also be understood that although this application has been described with reference to some specific examples, those skilled in the art can certainly implement many other equivalent forms of this application, which have the features described in the claims and are therefore all within the scope of protection defined herein.

[0038] The above and other aspects, features and advantages of this application will become more apparent when taken in conjunction with the accompanying drawings and in view of the following detailed description.

[0039] Specific embodiments of this application are described below with reference to the accompanying drawings; however, it should be understood that the claimed embodiments are merely examples of this application, which can be implemented in various ways. Well-known and / or repeated functions and structures are not described in detail to ascertain the true intent based on the user's historical operations, and to avoid unnecessary or redundant details that would obscure this application. Therefore, the specific structural and functional details claimed herein are not intended to be limiting, but merely serve as the basis and representative basis for the claims to teach those skilled in the art to use this application in various ways with substantially any suitable detailed structure.

[0040] This specification may use the phrases “in one embodiment,” “in another embodiment,” “in yet another embodiment,” or “in other embodiments,” all of which may refer to one or more of the same or different embodiments according to this application.

[0041] The principles and features of the present invention are described below with reference to the accompanying drawings. The embodiments described are for illustrative purposes only and are not intended to limit the scope of the invention. The following description, in conjunction with... Figure 1-2 The preferred embodiments of the present invention will be described in further detail below:

[0042] like Figure 1 As shown, this embodiment of the invention provides an intelligent operation and maintenance system for medium- and high-speed generator sets based on cloud-edge collaboration, including:

[0043] The intelligent operation and maintenance center is equipped with a cloud data and algorithm server and a cloud operation and maintenance server. The cloud data and algorithm server is deployed with a Tdengine time series database and an operation and maintenance data warehouse to store high-frequency operation data from edge devices and run fault prediction, health assessment and control strategy algorithms to provide intelligent decision support. The cloud operation and maintenance server provides a visual interface, policy management and remote control functions to realize human-computer interaction and multi-site collaborative operation and maintenance.

[0044] The cloud-based data and algorithm server is equipped with an algorithm service module and a data service module. The cloud-based operations and maintenance server is equipped with an operations and maintenance service module. The algorithm service module includes sub-modules such as an anomaly monitoring module, parameter prediction module, parameter correction module, health assessment module, and heat load calculation module to realize anomaly monitoring, parameter prediction, parameter correction, health assessment, and heat load calculation functions. The data service module includes sub-modules such as a data cleaning module, data storage management module, real-time / historical data query module, trend query module, and anomaly data management module to realize data cleaning, data storage management, real-time data query, historical data query, trend data query, and anomaly data management functions. The operations and maintenance service module is used to realize functions such as operation status diagram drawing, health score diagram drawing, anomaly prompts, fault records, alarm lists, and coolant filling prompts.

[0045] Several edge data acquisition and control terminals, numbered 1 to n, are distributed and deployed on each vehicle-mounted medium- and high-speed generator set. Each terminal is responsible for local data acquisition and control response. As the local control core, each edge data acquisition and control terminal integrates HTTPS and Modbus multi-protocol communication capabilities. It is responsible for collecting various on-site operating data and uploading it to the cloud, while simultaneously receiving cloud control commands and executing local adjustments. Edge data acquisition and control modules are deployed on each edge data acquisition and control module. Each edge data acquisition and control module includes three sub-modules: an oil monitoring module, a vibration monitoring module, and a generator set monitoring module. The oil monitoring module collects lubrication status data, including lubricating oil viscosity, water activity, and metal particles. The vibration monitoring module monitors the vibration characteristics of key parts of the generator set. The generator set monitoring module monitors various engine operating parameters, including temperature, pressure, and load, and performs protective control.

[0046] The intelligent operation and maintenance center and several edge acquisition and control terminals achieve two-way data interaction and control command issuance through wireless communication.

[0047] like Figure 1 As shown in the diagram, the system embodies an integrated architecture of "cloud-edge collaboration + multi-source fusion perception + intelligent decision support", supporting multi-dimensional intelligent operation and maintenance functions such as operation status monitoring, health assessment, trend prediction, intelligent control and fault management of vehicle generator sets.

[0048] In one embodiment, the implementation scheme and steps of the present invention are illustrated using the system configuration of a certain type of diesel generator set as an example:

[0049] Table 1: Main Components and Internal Communication Methods of the Operation and Maintenance System

[0050]

[0051]

[0052] As shown in Table 1, the system mainly consists of a cloud-based data and algorithm server, a cloud-based operation and maintenance server, an edge data acquisition and control module, an online oil acquisition module, a vibration monitoring module, and a unit monitoring box. Each module is interconnected efficiently through industrial standard communication protocols such as HTTPS and Modbus TCP / RTU.

[0053] The cloud-based data and algorithm server deploys the Tdengine time-series database and operations and maintenance data warehouse to store high-frequency operational data from edge devices and run fault prediction, health assessment, and control strategy algorithms to provide intelligent decision support. The cloud-based operations and maintenance server provides a visual interface, policy management, and remote control functions to enable human-computer interaction and multi-site collaborative operations and maintenance.

[0054] The edge data acquisition and control module serves as the local control core, integrating HTTPS and Modbus multi-protocol communication capabilities. It is responsible for collecting various operational data from the field and uploading them to the cloud, while also receiving control commands from the cloud and executing local adjustments.

[0055] The online oil acquisition module monitors lubricating oil parameters in real time, including temperature, viscosity, water activity, dielectric constant, and metal / non-metal particle count, providing fundamental data for equipment wear analysis and fault prediction. The vibration monitoring module collects high-frequency vibration signals from key components of the generator set and extracts characteristic parameters (such as RMS, power spectrum, skewness, and kurtosis) for early identification of structural faults.

[0056] The generator set monitoring box is used to monitor the generator set's operating status and safety protection functions, and transmits key operating parameters to the edge acquisition and control module via Modbus RTU, serving as the basis for system status assessment and control strategy adjustment.

[0057] The embodiments of this invention cover the complete process of system construction, data acquisition and processing, intelligent analysis, predictive control, remote linkage, and emergency response, as detailed below.

[0058] 1) Multi-source operational data acquisition

[0059] The system deploys multiple types of sensors in key parts of the generator set, such as the diesel engine body, oil pan, turbocharger, bearings, and cooling system, to collect data in real time, including lubricating oil pressure and temperature, shaft temperature, stator temperature, common rail pressure, cylinder exhaust temperature, diesel quality parameters (metal particles, non-metal particles, water content, dielectric constant, etc.), and vibration signals, so as to achieve full coverage of the unit's operating status and full life cycle monitoring.

[0060] In one embodiment, sensor modules for temperature, pressure, particle, and vibration are deployed in key components of the vehicle-mounted truck generator set, including the diesel engine body, oil pan, turbocharger, lubrication system, bearings, and cooling unit. The sensors connect to an edge data acquisition and control module via Modbus RTU / TCP protocol to collect key parameters such as lubrication oil inlet pressure / temperature, water activity, particle content, shaft temperature, stator temperature, common rail pressure, exhaust temperature, and fan speed, and report them in real time according to a set sampling period (e.g., 1s to 5s). Typical parameters are shown in Table 2-3.

[0061] Table 2. Measurement Points for Thermal Parameters of Generator Sets

[0062]

[0063]

[0064]

[0065] Table 3 Engine Oil Condition Monitoring Table

[0066]

[0067]

[0068] 2) Operation feature extraction and status index construction

[0069] The edge computing unit processes and extracts features from the raw sensor data, including key indicators such as RMS value, power spectral density, kurtosis, skewness, particle growth rate, and temperature rise rate. A unified set of health status indicators is constructed to provide structured input for fault identification and modeling analysis.

[0070] The edge vibration monitoring module has an embedded feature calculation unit that processes the collected signals in real time and extracts key feature parameters such as vibration RMS value, power spectral density, kurtosis, and skewness to construct an equipment health index matrix, which serves as the input vector for subsequent modeling, trend prediction, and early warning.

[0071] Table 4. Feature Extraction of Vibration Monitoring Locations

[0072]

[0073]

[0074]

[0075] 3) Data cleaning, synchronization and preprocessing

[0076] The system is designed with a multi-channel data stream preprocessing mechanism to perform timestamp alignment, missing value imputation, outlier removal, noise filtering, and normalization for data from different sources, ensuring the quality of the modeling input data and the accuracy of the analysis results.

[0077] All sensor data is preprocessed at the edge acquisition and control terminal, and the process includes:

[0078] (1) Timestamp alignment (synchronizing asynchronous data collected from different modules in a unified time);

[0079] (2) Determine data quality and measurement point refresh status;

[0080] The data preprocessing process at the intelligent operation and maintenance center includes:

[0081] (1) Outlier removal (removing noise points such as abrupt changes and constant values);

[0082] (2) Missing completion (interpolation or rule-based completion).

[0083] 4) Construction of a multi-time-domain data warehouse system

[0084] The system supports building short-term, medium-term, and long-term data warehouses in four time domains: 5 minutes, hours, days, and weeks. It supports differentiated sampling rate storage strategies, which improves storage efficiency and provides support for multi-level data applications such as trend modeling, degradation analysis, and operation and maintenance strategy optimization.

[0085] The system constructs a multi-layered data warehouse:

[0086] Short-term data warehouse (5 minutes to 1 hour): used for real-time trend analysis;

[0087] Mid-term data warehouse (1 hour to 1 day): used for periodic performance evaluation and operational analysis;

[0088] Long-term data warehouse (1 day to 1 month): used for fault history tracing and model training.

[0089] All data is stored in layers with compression according to a set sampling rate and time window, supporting fast retrieval and playback. The specific time period depends on project requirements. The multi-time domain data warehouse system provides data warehouse creation and management tools, as well as layered extraction ETL templates.

[0090] 5) Data storage and historical analysis in edge-cloud collaboration

[0091] The system employs a combined architecture of edge caching and cloud-based time-series databases (such as Tdengine), enabling rapid buffering and short-term response at the edge, while providing large-capacity data storage and long-term trend analysis in the cloud. It supports functions such as historical data backtracking for key events, anomaly reconstruction, operational status monitoring, and model training data extraction.

[0092] A cached database is deployed at the edge data acquisition and control terminal to enable rapid reading and analysis of short-term data, and an in-memory database (such as Redis or local storage) is used to support rapid response. A high-performance time-series database (such as Tdengine) is deployed in the cloud (intelligent operation and maintenance center) to support the storage of large-capacity long-term data and multi-task analysis. After a critical event is triggered, it supports backtracking the trajectory of relevant parameters by time window to achieve trend reconstruction and behavior analysis.

[0093] 6) State modeling and environment correction mechanism

[0094] The system combines operational data with environmental factors (such as atmospheric pressure, temperature, and altitude) to construct a correction model, dynamically correct thermal parameters, and establish equipment health scoring models and operating condition prediction models to achieve multi-variable fusion modeling, providing high-accuracy support for fault early warning and operation scheduling.

[0095] The modules deployed on the cloud-based data and algorithm servers in this system possess multivariate fusion state modeling and environmental correction capabilities, which are used to improve the diagnostic accuracy of key operating parameters, the consistency of cross-sectional comparisons, and the robustness of the model to environmental disturbances. This mechanism is particularly suitable for typical mining area scenarios such as high-altitude, frigid, large diurnal temperature ranges, and drastic seasonal variations.

[0096] The specific functions consist of three parts:

[0097] (1) Environmental parameter prediction modeling

[0098] The system uses Facebook Prophet time series modeling tool to build predictive models for key environmental factors such as altitude, temperature, and atmospheric pressure. It has good fitting and extrapolation capabilities for long-term trends, seasonal fluctuations, and diurnal variation patterns, and can predict the environmental conditions in the next few hours to days.

[0099] Input parameters include:

[0100] Historical environmental monitoring data (temperature, air pressure, humidity, altitude, etc.);

[0101] Historical operating data of the equipment (load, power, temperature response);

[0102] After model training, the system can periodically output environmental prediction data streams for the area where the equipment is located based on meteorological data over a period of time, providing an input basis for the subsequent dynamic thermal correction module.

[0103] (2) Dynamic correction mechanism for thermal parameters

[0104] Based on the predicted environmental factors and the currently collected real-time operational data, the system inputs the environmental parameters into the thermal parameter correction module in the operational model to dynamically correct and normalize the following indicators, in order to eliminate the risk of baseline drift and false alarms caused by changes in the external environment:

[0105] Fuel inlet temperature, lubricating oil inlet temperature, high-temperature water inlet temperature, high-temperature water outlet temperature, compressor inlet temperature, turbine exhaust temperature 1-4, cylinder exhaust temperature A1-A8, cylinder exhaust temperature B1-B8, main bearing temperature 1-8, generator stator temperature, generator bearing temperature, etc.

[0106] The main correction methods include:

[0107] The current measured value is compared with the standard operating curve under the same environmental conditions, and the baseline drift is corrected by residual correction.

[0108] When modeling fault trends, the input data is first normalized to improve the robustness and accuracy of the health scoring model and anomaly detection algorithm.

[0109] (3) Modeling output and system interaction

[0110] The revised modeling results will serve as input to the core health assessment (and fault prediction) module and will be simultaneously fed back to the cloud-based visualization system and edge control devices. The environmentally compensated baseline operating index values ​​will be used for alarm threshold comparison and strategy logic reference. This function can significantly reduce the false alarm rate caused by environmental disturbances and improve the consistency of horizontal comparisons of equipment operating data across different mining areas.

[0111] 7) Fault detection and early warning mechanism

[0112] The system supports a combined fault identification mechanism based on rule-based thresholds and model prediction, and can detect anomalies including but not limited to: low lubricating oil level (indirectly predicted via water temperature), abnormal common rail pressure, air leakage in the intake system, abnormal single-cylinder power output, bearing failure, excessive shaft / stator temperature, and oil deterioration. The system can issue early warnings based on changes in characteristic values ​​and respond in stages according to severity (e.g., yellow, orange, and red alarms).

[0113] The system integrates two types of fault identification methods:

[0114] Rule-based detection: Set threshold rules, such as "lubricating oil temperature > 95℃" or "lubricating oil pressure < 1.5 bar" to determine abnormal lubrication;

[0115] A dynamic threshold monitoring mechanism based on operating condition statistics: This mechanism detects dynamic threshold anomalies based on current operating conditions. Due to the drastic fluctuations in the operating conditions of truck generator sets, especially in actual mining environments where equipment experiences frequent start-stop cycles, sudden load changes, hill climbing, and speed variations, traditional fixed threshold alarm mechanisms are prone to false alarms or missed alarms. Therefore, this system constructs a dynamic threshold status monitoring mechanism based on operating condition modeling on a cloud-based algorithm server to achieve intelligent and adaptive anomaly detection of operating parameters. This mechanism specifically includes the following steps and structure:

[0116] (1) Construction of working condition modeling data

[0117] Based on the generator set operating data, the following two types of core information are extracted:

[0118] Operating condition characteristics include parameters such as diesel engine speed, fuel rack displacement, load power, and throttle opening.

[0119] Status response dimensions: such as cylinder block exhaust temperature, common rail pressure, stator temperature, shaft temperature, lubricating oil pressure, etc.

[0120] The two dimensions are fused to construct a two-dimensional sample dataset, which is then used as input for dynamic modeling.

[0121] (2) Construction of State Detection Boundary Model

[0122] During the system deployment phase, the following statistical and modeling process is performed by accumulating long-term data on key operating parameters:

[0123] Mathematical statistics methods were used to obtain the mean and variance of state parameters within different operating condition ranges;

[0124] A nonlinear regression model is used to construct the upper and lower boundary mapping relationship between "state parameters and operating condition factors";

[0125] The model output is the upper / lower limit range of a certain state parameter at a certain operating point (such as 75% load).

[0126] For example Figure 2 As shown, the upper and lower boundary fitting curves of "the change of exhaust temperature of a certain cylinder of the engine with load" can be used to determine whether the real-time temperature is within the normal fluctuation range.

[0127] (3) Real-time data deviation calculation

[0128] During the actual operation of the system, deviation analysis is performed on the real-time data within a specified time window from the model boundary, including:

[0129] Deviation calculation: Calculate the deviation distance of the state parameters from the upper and lower boundary intervals within this time period;

[0130] Deviation ratio calculation: The percentage of data points that fall outside the boundary is the total number of data points.

[0131] (4) Health score generation and abnormality determination

[0132] Combining two types of indicators, deviation degree and deviation ratio, the system generates a health score H for each measurement point over a certain period of time. j :

[0133]

[0134] Among them, D j : The deviation distance at point j; Deviation point ratio; α, β: adjustment weights;

[0135] H j ∈[0-1]: 1 represents health, 0 represents severe abnormality;

[0136] The system is based on H j The alarm response is tiered, and alarm levels can be set as follows: yellow (early warning), orange (medium-level anomaly), and red (severe anomaly).

[0137] (5) Adaptive threshold alarm logic

[0138] Unlike traditional fixed threshold alarm mechanisms, this system can automatically call the matching boundary model according to different operating conditions to achieve linkage between "operating conditions and alarm thresholds", thereby avoiding false alarms caused by seasonal changes and load fluctuations.

[0139] 8) Intelligent feedforward control mechanism for cooling water temperature change

[0140] The system has the capability to analyze the dynamic trend of cooling water temperature, and can monitor the heating rate and absolute temperature value in real time. When the system detects that the cooling water temperature is showing a continuous upward trend but has not yet reached the alarm threshold, it determines that the cooling capacity is reduced due to potential load increase or insufficient lubricating oil. At this stage, the system can trigger an early warning to automatically replenish lubricating fluid or cooling water, preventing the temperature from exceeding the limit.

[0141] In automotive diesel generator sets, the adequacy of lubricating oil and coolant directly affects engine cooling and lubrication. However, due to the limitations of harsh operating conditions (such as high altitude, vibration, and dusty environments), the accuracy of oil pan level sensors is insufficient or they frequently fail, often leading to inaccurate level monitoring and delayed fluid replenishment, which in turn causes engine overheating, poor lubrication, and other malfunctions. Therefore, constructing a feedforward fluid replenishment reminder mechanism based on the trend of key parameters has become an important means to improve system reliability.

[0142] On the cloud data and algorithm server, the system runs the following supporting logic in parallel:

[0143] (1) Liquid level prediction model modeling: Combine historical data to construct a "cooling water temperature change - liquid level change" mapping model for long-term trend correction and system health score input;

[0144] (2) Replenishment suggestion generation: If the system has not reached the automatic replenishment threshold but the trend is judged to be that it will continue to heat up and exceed the warning value in the next 30 minutes, the smart operation and maintenance center will push a "replenishment suggestion" prompt message through the operation and maintenance platform to assist operation and maintenance personnel to intervene in advance;

[0145] (3) Edge-cloud data verification: The intelligent operation and maintenance center and the edge acquisition and control end verify the cooling system parameters, liquid replenishment execution records and response efficiency in a two-way manner, which is used for subsequent model correction and strategy optimization.

[0146] This intelligent fluid replenishment mechanism does not rely on the accuracy of oil pan level measurement. Instead, it achieves feedforward control through the thermal response trend of the cooling system, which not only improves the ability to identify faults before they occur but also effectively reduces the risk of overheating or wear caused by inaccurate fluid level monitoring. Meanwhile, the intelligent operation and maintenance center's alert mechanism enhances human-machine collaboration efficiency and adapts to the actual needs of remote maintenance and safety assurance under complex operating conditions.

[0147] 9) Intelligent wind turbine speed regulation mechanism based on generator temperature change trend

[0148] The system supports dynamic trend identification of key temperature parameters of the generator set. Combined with the current load conditions, ambient temperature, and cooling system response capabilities, it automatically adjusts the speed of the cooling fan to achieve precise temperature control and energy consumption optimization.

[0149] Generator sets face challenges in mining operations, including frequent start-stop cycles, high load impacts, and drastic temperature changes. The generator windings and bearings are the components most prone to heat accumulation and significant thermal degradation in the system; their temperature rise not only reflects changes in heat load but also directly affects equipment lifespan and failure risk. Traditional fan control methods mostly employ constant speed or simple temperature triggering, resulting in slow response, low energy efficiency, and an inability to adapt to complex dynamic operating conditions.

[0150] To improve the real-time performance and intelligence of cooling response, this system designs and deploys an intelligent fan speed regulation mechanism based on the collaborative analysis of generator winding temperature, shaft temperature change trend, load status and environmental factors. It achieves adaptive control of active heat dissipation and energy efficiency balance through edge computing.

[0151] (1) Fusion modeling

[0152] The system collects and calculates the following key parameters in real time:

[0153]

[0154]

[0155] (2) Constructing the heat load index HSI

[0156] The generator heat load index is calculated using a multivariate weighted model.

[0157]

[0158] Among them, w1, w2, w3, w4, and w5 are initially set to equal weights, and the total weight parameter can be trained and optimized based on historical sample data and dynamically distributed to edge devices for dynamic adjustment.

[0159] (3) Normalization of heat load index HSI

[0160] The raw HSI values ​​are linearly scaled according to their historical lower and lower limits, mapping them to the 0–100 range:

[0161]

[0162] HSI min (Can be set to historical minimum load or theoretical minimum value).

[0163] HSI max : Can be set to the historical highest value or the thermal limit threshold allowed by the system.

[0164] (4) Determination of heat load level and fan control strategy

[0165] Based on real-time calculated HSI norm The system automatically adjusts the PWM duty cycle of the fan.

[0166] 10) Cloud-edge collaborative architecture and operating mechanism

[0167] The system supports communication protocols such as HTTPS and Modbus TCP / RTU, forming a closed-loop communication architecture between the cloud server and edge devices. Edge devices can perform real-time data acquisition, rapid response, and control command execution locally, while the cloud handles centralized policy management, model training, remote fault analysis, and command distribution. It supports bidirectional data synchronization between the edge and cloud, as well as a fault-tolerant operation mode for data loss.

[0168] The cloud (intelligent operation and maintenance center) and the edge (edge ​​acquisition and control) maintain bidirectional data communication via an HTTPS channel, supporting real-time command issuance, data upload, and reconnection after interruption. The edge acquisition and control module has capabilities such as resume transmission after network outage, local policy caching, and remote configuration updates. In case of system anomalies, it can switch to "edge autonomous mode" to ensure the independent operation of basic functions, and resume linkage after communication is restored.

[0169] 11) Remote visual monitoring and control center interface

[0170] The system provides a remote monitoring platform that supports various visualization modules such as operation status graphs, health score graphs, fault records, alarm lists, and control command issuance interfaces, allowing users to view the unit's operating status in real time on a PC or mobile terminal.

[0171] The cloud-based (intelligent operations and maintenance center) operations and maintenance server deploys a web-based monitoring platform that supports a visual operating interface, such as real-time status dashboards, health score radar charts, historical data curves, alarm pop-ups, and automatic generation of operating reports. The remote control interface supports functions such as start / stop, mode switching, and parameter setting, and maintains synchronized display with mobile devices.

[0172] 12) Emergency response strategies and proactive safety protection mechanisms

[0173] When the system detects a serious fault risk (such as cooling system failure or severe deterioration of lubricating oil quality), it can automatically trigger emergency control logic, including load limiting operation, active shutdown, remote alarm, and entering protection mode. The system supports manual intervention access points, allowing remote maintenance personnel to manually intervene according to the policy. It also has a task maintenance / emergency return capability assessment mechanism to determine whether to continue executing the task or request a return to the field.

[0174] When the system identifies a major risk (such as severe deterioration of oil viscosity or cooling system failure), the system immediately enters the emergency response process:

[0175] • The edge device automatically executes "load-limited operation" or "active shutdown" commands;

[0176] • Generate event logs and status snapshots;

[0177] • Send high-priority alarms to the remote platform;

[0178] The system also supports local manual reset and remote expert intervention, allowing for manual intervention or strategy adjustments when necessary.

[0179] This invention integrates multi-source data acquisition modules for fluid, vibration, and temperature, combined with cloud-based health assessment and heat load calculation models, to achieve dynamic perception and precise early warning of the operating status of diesel generator sets. Employing environmental correction and trend modeling technologies, it eliminates the interference of altitude and season on parameter baselines, improving diagnostic accuracy. The edge device features intelligent fluid replenishment prompts and fan speed adjustment functions, enabling rapid local response. The cloud supports fault tracing and remote intervention, significantly enhancing the stability, intelligence, and safety of the equipment under extreme conditions such as high altitudes and frigid environments. Compared with existing technologies, this invention can significantly improve the operational stability and safety of medium- and high-speed diesel generator sets in complex environments, reducing failure rates and maintenance costs.

[0180] The above embodiments are merely exemplary embodiments of the present invention and are not intended to limit the present invention. The scope of protection of the present invention is defined by the claims. Those skilled in the art can make various modifications or equivalent substitutions to the present invention within its spirit and scope of protection, and such modifications or equivalent substitutions should also be considered to fall within the scope of protection of the present invention.

Claims

1. A cloud-edge collaboration-based intelligent operation and maintenance system for medium and high-speed generator sets, characterized in that, The intelligent operation and maintenance center end is provided with a cloud data and algorithm server and a cloud operation and maintenance server. The cloud data and algorithm server is deployed with a Tdengine time series database and an operation and maintenance data warehouse, is used for storing high-frequency operation data from the edge device, and runs fault prediction, health assessment and control strategy algorithms to provide intelligent decision support. The cloud operation and maintenance server provides a visual interface, strategy management and remote control functions to realize man-machine interaction and multi-site collaborative operation and maintenance.

2. The intelligent operation and maintenance system of the medium and high speed generator set based on cloud edge collaboration according to claim 1, wherein: The cloud data and algorithm server is loaded with an algorithm service module and a data service module, and the cloud operation and maintenance server is loaded with an operation and maintenance service module. The algorithm service module is used to realize abnormal monitoring, parameter prediction, parameter correction, health assessment and thermal load calculation functions.

3. The cloud-edge collaboration based intelligent operation and maintenance system for a medium-high speed generator set according to claim 1, characterized in that: The data service module is used to realize data cleaning, data storage management, real-time data query, historical data query, trend query data query and abnormal data management functions. The operation and maintenance service module is used to realize operation situation diagram drawing, health score diagram drawing, abnormal prompt, fault record, alarm list and cooling liquid filling prompt functions. The edge control end is deployed with an oil monitoring module, a vibration monitoring module and a unit monitoring module. The oil monitoring module is used to collect lubrication state data including oil viscosity, water activity and metal particles.

4. The cloud-edge collaboration based intelligent operation and maintenance system for a medium and high speed generator set according to claim 2, characterized in that: The vibration monitoring module is used to monitor the vibration characteristics of the key parts of the generator set.

5. The cloud-edge collaboration based intelligent operation and maintenance system for a medium and high speed generator set according to claim 3, characterized in that: The unit monitoring module is used to monitor various operating parameters of the engine including temperature, pressure and load and perform protection control. 6.The cloud-edge collaborative based intelligent operation and maintenance system for medium and high speed generator set of claim 1, characterized in that: The algorithm service module has multivariate fusion state modeling and environment correction functions. The system combines operation data and environmental factors including atmospheric pressure, temperature and altitude to build a correction model for dynamic correction of thermal parameters. The vibration monitoring module is embedded with a feature calculation unit for real-time processing of collected signals to extract key feature parameters including vibration RMS value, power spectral density, kurtosis and skewness to construct a device health index matrix. The edge control end has functions including network interruption and continuous transmission, local strategy caching and remote configuration updating.

7. The intelligent operation and maintenance system of the medium and high speed generator set based on cloud edge collaboration according to claim 6, wherein: The system adopts a combination architecture of edge cache and cloud time series database, realizes fast buffering and short-term response at the edge control end, and realizes large-capacity data storage and long-term trend analysis at the intelligent operation and maintenance center end. When the system is running normally, the edge control end is in cloud-edge collaborative mode, collects various types of operation data on site and uploads them to the cloud, and receives cloud control instructions and executes local adjustment. When the system is abnormal, the edge control end switches to the edge autonomous mode to ensure the independent operation of basic functions, and then restores the linkage after communication is restored.

8. The cloud-edge collaboration based intelligent operation and maintenance system for a medium and high speed generator set according to claim 1, characterized in that: When the system identifies a major risk, the system immediately enters the emergency handling process, as follows, The edge control end automatically executes the "load limiting operation" or "active shutdown" command. Generate event logs and state snapshots. Send high-priority alarms to the remote platform. 9.The cloud-edge collaborative based intelligent operation and maintenance system for a medium-high speed generator set according to claim 1, characterized in that: The system supports local manual reset and remote expert intervention, and manual intervention or strategy adjustment when necessary.

10. The cloud-edge collaboration based intelligent operation and maintenance system for a medium-high speed generator set according to claim 1, characterized in that: The system supports the construction of short-term, medium-term and long-term data warehouses in four time domains: 5 minutes, hours, days and weeks, and supports differentiated sampling rate storage strategies.

Images