Remote intelligent operation and maintenance system for coal mine industry
By designing a remote intelligent operation and maintenance system for the coal mine industry, using RocketMQ, XGBoost and natural language processing technology, the operation and maintenance problems in the complex network environment of the coal mine industry are solved, efficient and intelligent operation and maintenance management are achieved, and operation and maintenance efficiency is improved. and system stability.
Patent Information
- Application Number
- CN202411064332.6
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-05
- Publication Date
- 2024-11-08
AI Technical Summary
Due to the complex network environment and high requirements for the network environment, the operation and maintenance systems in the coal mine industry are difficult to effectively deploy and maintain traditional operation and maintenance systems, and cannot meet the needs of operation and maintenance personnel to simplify work processes, improve response speed and realize intelligent management.
A remote intelligent operation and maintenance system for the coal mining industry was designed, including cloud server, mining client, instruction management unit, data analysis unit, network adaptation unit, natural language processing module, etc., using RocketMQ, XGBoost, spaCy, AES encryption, machine learning models and other technologies realize instruction release, data analysis, network adaptation, natural language interaction and efficient operation and maintenance management.
The reliability and high availability of operation and maintenance operations in complex network environments are realized, the operation and maintenance efficiency and intelligence level are improved, the operation and maintenance costs are reduced, the data-driven fault warning and natural language interaction are supported, and efficient service deployment is achieved and upgrade.
Smart Images

Figure CN118921387A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the field of remote operation and maintenance technology, and more specifically, to a remote intelligent operation and maintenance system for the coal mining industry. Background Art
[0002] The current operation and maintenance system of the coal mining industry faces complex network environment problems, such as the lack of fixed public network IP and the refusal to use specific intranet penetration tools, which make it difficult to effectively deploy and maintain the traditional operation and maintenance system. In addition, the servers in the coal mining environment may be distributed under different network conditions, and some servers can only communicate on the intranet, which poses a major obstacle to the operation and maintenance efficiency and system flexibility. Therefore, the traditional operation and maintenance solutions based on the HTTP protocol and relying on Kubernetes (k8s) are limited in application in specific scenarios of the coal mining industry due to their high requirements and complexity for the network environment, and cannot fully meet the needs of operation and maintenance personnel for simplifying workflows, improving response speed, and realizing intelligent management.
[0003] In this context, traditional operation and maintenance methods, such as remotely logging into each server to deploy and upgrade operations individually, or traveling to the coal mine site for maintenance, are not only time-consuming and labor-intensive, but also greatly increase operation and maintenance costs and reduce operation and maintenance efficiency. Therefore, there is an urgent need for a new system that can adapt to the special network environment of the coal mining industry and realize remote intelligent operation and maintenance to overcome the limitations of existing technologies and meet the needs of efficient and intelligent operation and maintenance. Summary of the invention
[0004] The purpose of the present invention is to provide a remote intelligent operation and maintenance system and system for the coal mining industry to solve the problems in the prior art.
[0005] In order to solve the above problems, the remote intelligent operation and maintenance system for the coal mining industry involved in the present invention adopts the following technical solutions: A remote intelligent operation and maintenance system for coal mining industry, comprising: Cloud server: responsible for issuing instructions, data processing and analysis; Instruction management unit: used to publish upgrade, deployment, and management instructions through RocketMQ, using topics and tags to distinguish target mining areas and operation types; Data analysis unit: used to collect and analyze data from the mine host, use the XGBoost model to perform abnormal analysis on log data, predict faults and issue early warnings; Mine client: used to be deployed on the mine host to receive and execute instructions, and collect and upload data; Instruction receiving and execution unit: used to subscribe to RocketMQ messages, parse and execute cloud instructions, such as container management and configuration modification; Data collection unit: used to monitor system resources, network traffic, Docker container status and hardware sensor data, and upload them to the cloud server at regular intervals.
[0006] Preferably, it also includes a network adaptation unit: used to have the function of message forwarding between intranets, and support the relay of instructions and data through designated proxy nodes in a complex network environment; Natural language processing module: integrated into the cloud server to process the voice or text instructions of operation and maintenance personnel.
[0007] Preferably, it also includes an instruction parsing unit: based on spaCy, it performs text preprocessing, word segmentation, and entity recognition to convert the operation and maintenance instructions into an operation sequence that can be executed by the system; Interaction confirmation unit: interacts with operation and maintenance personnel to confirm the intention of command execution and ensure that the operation is accurate before execution.
[0008] Preferably, it also includes a data encryption unit: AES-CBC mode is used to encrypt uplink and downlink data to ensure the secure transmission of instructions and data; RocketMQ cluster: provides highly available message queue services; Message queue unit: ensures the reliability and orderliness of command and data transmission, and supports command forwarding and data reporting in complex networks.
[0009] Preferably, it also includes an intelligent resource optimization and scheduling subsystem, including: Resource monitoring unit: real-time monitoring of system resource usage to ensure that data is updated to the database in real time; Model optimizer: uses machine learning models, such as reinforcement learning, to predict future demand based on historical resource usage patterns and business loads, dynamically adjust resource allocation strategies, and optimize resource utilization; Dynamic scheduler: Automatically adjusts Docker container resource quotas based on model output to keep services running stably and efficiently; Policy execution unit: used to realize the automated execution of intelligent scheduling decisions, ensure load balancing, and avoid service interruptions caused by single point failures.
[0010] Preferably, it also includes an intelligent safety protection subsystem, including: Threat intelligence collection unit: real-time tracking of network security trends, collection of the latest threat intelligence, including virus databases, vulnerability information, and attack patterns, to ensure that the system security database is updated; Behavior Analyzer: Uses AI algorithms to learn system behavior patterns, identify abnormal operations, and issue timely warnings; Response unit: Integrates automated defense strategies to immediately perform isolation, blocking, repair, or rollback operations once a security incident is discovered; Strategy Optimizer: Continuously learn and optimize defense strategies, adjust defense thresholds based on historical event feedback, improve recognition accuracy, reduce false positives and false negatives, and enhance protection capabilities.
[0011] Preferably, it also includes an intelligent knowledge management subsystem, including: Knowledge collection unit: used to automatically aggregate system logs, operation logs, troubleshooting cases, and operation and maintenance manuals, build a knowledge base, and support multi-dimensional indexing; Adaptive learner: Based on NLP technology, it learns the queries, instructions, and operating habits of operation and maintenance personnel, dynamically optimizes search sorting, and makes personalized recommendations to improve the efficiency of knowledge acquisition. Knowledge Updater: Real-time updates, combining external resources with internal operations, continuously enriching the knowledge base and ensuring cutting-edge information; Interactive interface: Provides a natural language interactive interface to facilitate operation and maintenance queries, quickly obtain operation guides and troubleshooting solutions, support voice input, and improve operation and maintenance response speed and convenience.
[0012] The beneficial effects of the present invention are as follows: This application proposes a remote intelligent operation and maintenance system for the coal mining industry, which aims to solve the operation and maintenance problems in the complex network environment of coal mines and improve the operation and maintenance efficiency and intelligence level. Its core technical effects are summarized as follows: Adaptability to complex network environments: The system is designed with a unique network management mechanism that can operate effectively under the restricted network conditions unique to coal mines (such as only some devices have external network access rights, unstable networks, etc.). Through the intranet proxy and message forwarding strategy, reliable transmission of instructions and data in an external network restricted environment is achieved, ensuring the accessibility of operation and maintenance operations and the high availability of the system.
[0013] Data-driven fault warning and analysis: The system integrates data collection and analysis modules, which can not only collect multi-dimensional information including logs, system resource usage, network traffic, container status and hardware sensor data, but also use machine learning models for fault prediction and anomaly analysis. Through deep learning processing of logs and XGBoost modeling, the system can automatically classify and locate problems, warn of potential faults in advance, and provide a scientific basis for preventive maintenance.
[0014] Natural language processing command system: The natural language processing technology based on spaCy is introduced, allowing operation and maintenance personnel to issue complex commands through voice or text. The system can accurately understand and execute operations such as service upgrades and configuration modifications. Through model training and dictionary optimization in specific fields, the recognition accuracy of operation and maintenance commands is significantly improved, and the interactive friendliness and operation and maintenance efficiency of the system are enhanced.
[0015] Efficient service deployment and upgrade capabilities: Using RocketMQ as a message queue, remote deployment, upgrades, and batch operations of services are realized. Combining Docker container technology and cloud native concepts ensures rapid deployment, flexible management, and version control of applications. AES encryption ensures data transmission security, and RocketMQ's sequential message characteristics ensure the orderliness of instruction execution, improving the standardization and automation of operation and maintenance operations.
[0016] In summary, the present invention provides a comprehensive, intelligent and adaptable remote operation and maintenance solution for the coal mining industry by integrating complex network adaptation strategies, data intelligent analysis, natural language interaction and efficient operation and maintenance management technology, which significantly reduces operation and maintenance costs and improves operation and maintenance efficiency and system stability. BRIEF DESCRIPTION OF THE DRAWINGS
[0017] In order to more clearly illustrate the technical solution of the embodiment of the present invention, the following briefly introduces the drawings required for use in the embodiment: Figure 1 is a diagram of the overall architecture of the system of this application; FIG2 is a flow chart of deployment and upgrade of the application system service; FIG3 is a flowchart of remote management of the system of the present application; Figure 4 is a flowchart of data warning and analysis of the application system; FIG5 is a networking diagram of complex network management of the present application; FIG6 is a networking diagram of complex network management of the present application; FIG7 is a flowchart of complex network management of the present application; FIG8 is a flowchart of the corresponding operation of the natural language implementation of this application. DETAILED DESCRIPTION
[0018] In order to make the technical purpose, technical solution and beneficial effects of the present invention clearer, the technical solution of the present invention is further described below in conjunction with the drawings and specific embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention, that is, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings herein can be arranged and designed in various different configurations.
[0019] The current operation and maintenance systems generally need to ensure network connection, and most of them are based on the http protocol to obtain or push edge measurement data. Secondly, the operation and maintenance systems are mostly based on k8s management. Due to the complex maintenance of k8s, and the high professional requirements for the network environment and personnel, it is relatively heavyweight architecture. The network environment of the coal mining industry is relatively complex, and the systems are deployed on the intranet. Even if there is external network capability, there is no fixed public network IP, and the mine refuses to use tools such as nps as intranet penetration (it will be scanned as a vulnerability), and some mines have multiple servers. In the case of multiple servers, a server will have external network capabilities, while other servers can only communicate on the intranet. In this network environment, the operation and maintenance systems on the market are not sufficient to support it. Before this system was developed, our operation and maintenance personnel could only remotely deploy and upgrade services one by one or deploy and upgrade services on business trips to coal mines. This application supports the deployment architecture arm / x86, Os: windows10 and above / linux / kylin, which already includes the scope of current mining servers.
[0020] The overall architecture of this application system is as follows Figure 1 , The operation management system consists of two parts: the cloud server and the client deployed on the mine host. The functions of this system will be introduced below.
[0021] Deployment and upgrade of services Figure 2 ; The cloud server sends upgrade or deployment instructions to MQ through the corresponding topic and other distinguishing fields; The client of the mine obtains this instruction by distinguishing fields and parses it; After obtaining the image name and version number that needs to be operated this time, it will check whether the image of this version exists locally. If not, it will go to the cloud image repository to pull the image of the corresponding version; Determine whether this command requires modification / addition / deletion of the corresponding configuration file, and if so, perform the corresponding operation; Determine whether this instruction requires modifying the database file before starting the container. If so, execute the SQL script first. Create a container, mount files, map ports, set network mode, host name, etc. Start the container; Report the status of this task, whether it is successful; RocketMq is used as the data transmission middleware, and the sequential messages provided by RcoketMq are used to ensure the orderliness of the command data issuance. The topoic and tag fields of the message are used to distinguish each mining area. For example, if the topic name is "topic_iiot_upgrade" and the tag is "001001-1", it means that an upgrade command is sent to the No. 1 node in the mine numbered 001001.
[0022] Rocket has a consumption retry and flow control mechanism. After a message is received and consumed by a subscriber, an ack command will be sent to the server to ensure that the message has been consumed. If consumption fails due to network or other reasons, a consumption retry will be triggered to ensure data reliability. High availability is ensured by deploying a RocketMq cluster; The issuance of instructions and reporting of data use AES CBC mode for data encryption to ensure data security.
[0023] Remote management such as Figure 3 ; The server sends corresponding instructions to MQ (such as restarting the container, deleting the container, stopping the container, deleting the image, modifying the configuration file, database operation, upload frequency setting, etc.); The client obtains the corresponding instruction from MQ; Perform the appropriate action.
[0024] Data Collection: Use Docker to mount the logs on the host machine to obtain error and warning level logs for data upload; Obtain the usage of computer system resources such as CPU, memory, hard disk usage and other data for data upload; Obtain information such as computer network traffic for data upload; Get Docker image information and container information to upload data; Obtain basic computer hardware information for data upload (such as computer room temperature, smoke, water immersion, door magnet and other hardware sensor data); The above data are uploaded regularly / collected every five minutes by default. The upload frequency can be set by the host computer or configuration file; Due to the inconsistent environment of the mine computer room, it is impossible to determine the collection method of the hardware sensor. This aspect is relatively customized. Some cabinets have many sensors that need to be connected to the cabinet manufacturer, and some are installed by themselves in the computer room. The connection method is different. This is not a key item, but a icing on the cake.
[0025] Data warning and analysis are shown in Figure 4; 1. Upload the running status of Docker data and restart the container if any abnormality occurs in the container; if the restart fails, issue an early warning through email, SMS or other means; 2. Retrieve uploaded data log data and quickly locate abnormal problems; 3. Use machine learning modeling on log data and train the model to achieve anomaly analysis and location; take corresponding actions based on this analysis to achieve the function of automatic classification and processing of anomalies.
[0026] First, preprocess the log data to obtain the number of problem logs and corresponding labels, so as to obtain the model input data; then slice the data according to a certain length, combine the data with the same feature name, count them in groups, and add labels; filter the data determined in the previous step, use the topic model for feature extraction, and represent the log data as a vector of topic distribution to extract semantic information; finally, perform XGBoost modeling training, and after model evaluation and retraining, obtain a model for locating abnormal logs to achieve abnormality analysis and location; input the preprocessed log data into the model to obtain which type of abnormality the log data belongs to, so as to deal with the abnormal data in a targeted manner. (If the log reports a redis service abnormality, check the redis service to determine whether the redis server is started, and start it if it is not started); Figure 4 .
[0027] Analyze the uploaded resource data to provide recommendations on whether the computer hardware / network needs to be upgraded and the optimal configuration.
[0028] When the system detects a fault, it will do the following: Fault detection: The system monitors various performance indicators and log data in real time and uses machine learning models to detect anomalies. Once an anomaly is found, a preliminary analysis is performed immediately.
[0029] Preliminary analysis: The system analyzes the nature and possible causes of the fault based on preset rules and historical data. This step may include automatic classification and tagging of log data.
[0030] Decision making: Based on the analysis results, the system decides the next course of action. If the fault can be automatically repaired, the system will execute the corresponding automated script; if manual intervention is required, the system will transfer it to the corresponding operation and maintenance team.
[0031] Execution and feedback: The automated script executes the repair operation and feeds back the execution result to the system. If the repair is successful, the system records the fault handling log and ends the process; if the repair fails, the system will escalate the fault level and notify the higher-level operation and maintenance team.
[0032] Script automation: The system has a series of built-in automation scripts for common faults. Once a matching fault mode is detected, the corresponding script will be automatically executed to repair it.
[0033] Machine Learning Models: In some cases, the system can use machine learning models to predict the occurrence of failures and take preventive measures before failure occurs.
[0034] Application examples of automated script machine learning models. Taking the Redis service failure as an example, the system can apply machine learning models in the following ways: Historical data analysis: The system collects historical failure data of Redis services, including performance indicators, log patterns, and repair actions before the failure.
[0035] Feature engineering: Extract features from historical data that help predict failures, such as memory usage, request latency, error log frequency, etc.
[0036] Model training: Use the extracted features and fault labels to train a classification model that can predict whether the Redis service is about to fail.
[0037] Real-time monitoring and prediction: The system monitors various indicators of the Redis service in real time and inputs the data into the trained model to predict whether a failure will occur.
[0038] Preventive measures: If the model predicts a risk of failure, the system can automatically perform some preventive measures, such as increasing memory quotas, restarting services, etc., to avoid the failure.
[0039] In complex network management, such as Figure 5 , only host 1 has external network capability, and hosts 2 and 3 cannot directly connect to the MQ of the cloud server. In this case, only host 1 can perform operations such as command forwarding and image downloading. Or all three computers are connected to the external network, but the network environment is unstable and often disconnected. Figure 6 ; Suppose now you need to upgrade a service of the No. 3 host of this mine, such as Figure 7 , host number 3 has no network for some reason. Here are the steps: First, when installing the client, configure the network group data of nodes one, two, and three through the configuration file (that is, tell them that the three are one network).
[0040] The server determines whether it has network connection by accessing the cloud service; if it has external network capability, it broadcasts to other servers in the group; If no one in this group has the ability to connect to the Internet, it will retry at regular intervals. If a certain machine has no ability to connect to the Internet, but other networks in the group have the ability to connect to the Internet, it will select a machine. Here we select host number 1. Host No. 3 tells Host No. 1 to bind, and Host No. 1 will add the message subscription of Host No. 3 until Host No. 3 tells Host No. 1 to stop binding.
[0041] After the host No. 1 receives the command sent to the host No. 3, it sends the data to the host No. 3 through the intranet MQ / or other technologies, and the host No. 3 performs specific operations on the data. If an image needs to be downloaded, host No. 1 will first download the corresponding image and then send it to host No. 3 via the intranet.
[0042] Basic algorithm: The intranet IP of this group (such as 192.168.1.1, 192.168.1.2, 192.168.1.3); Determine whether you can access the cloud server (actually heartbeat detection); If possible, broadcast such as 192.168.1.1 - can access the external network. If not, see if there are other hosts broadcasting. The broadcast is refreshed every 10 minutes.
[0043] Assume that the broadcast list at this time is: (192.68.1.1-can access the external network, 192.68.1.2-can access the external network, 192.68.1.3-cannot); At this time, 192.68.1.3 will find the first node that can access the external network in order, and find 192.168.1.1 here. 192.68.1.1 will subscribe to the MQ message of 192.68.1.3, and 192.68.1.3 will send the data to be reported to 192.68.1.1 for reporting.
[0044] At this time, if there is a problem with the 192.168.1.1 network, it will continue to look for nodes that can access the external network, and 192.68.1.2 will subscribe to the MQ messages of 192.68.1.1 192.68.1.3.
[0045] The network will be retried every 10 minutes. If 192.168.1.1 is back online after the retry, 192.68.1.2 will be notified that it no longer needs to subscribe to 192.68.1.1's data and the binding will be canceled.
[0046] Notification means and instruction forwarding are implemented using the intranet redis message queue; Data transfer: Use the intranet SFTP server; The corresponding operations are implemented in natural language, as shown in Figure 8: The server supports voice or text input to implement corresponding operations, such as upgrading the xxx image of the xxx host in the xxx mine to the xxx version.
[0047] Use the spaCy open source natural language processing library; By downloading Wikipedia Chinese, preprocessing the data, cleaning the text, using Jieba word segmentation to segment the text, and converting traditional Chinese characters to simplified Chinese; Divide the text data and use spaCy for model training; Make adjustments based on this model and add dictionary data necessary for operation and maintenance.
[0048] After performing operations such as word segmentation, stop word filtering, part-of-speech tagging, and named entity tagging on the text sent by the user; Split a large instruction into small instructions (such as first finding the xxx mine topic code, then checking whether the image exists, pulling the corresponding image, creating container parameters, etc.); Prompt the user to confirm the command; Issue instructions; spaCy library version. In this system, we chose the spaCy natural language processing library. spaCy is a very efficient and feature-rich Python natural language processing library that supports pre-trained models in multiple languages. The specific spaCy version used is spaCy v2.x, because it provides a stable API and relatively complete Chinese processing capabilities.
[0049] Model training data sources,In order to train a model that can accurately understand operation and maintenance instructions,,we used the following data sources: Wikipedia Chinese data: Wikipedia provides a large amount of high-quality Chinese corpus, and we downloaded some of the data for model training.
[0050] Internal operation and maintenance manual: The company’s internal operation and maintenance manual contains a large number of operation and maintenance terms and common instructions. This data is crucial to improving the accuracy of the model.
[0051] Actual operation and maintenance logs: By collecting and analyzing log data from actual operation and maintenance, we can discover the expressions and terms commonly used by operation and maintenance personnel.
[0052] Dictionary optimization details,To further improve the accuracy of the model, we optimized spaCy’s Chinese dictionary as follows: Add professional terms: Add professional terms related to the coal mining industry and operation and maintenance to the dictionary to ensure that these terms can be correctly recognized by the model.
[0053] Custom stop words: According to the operation and maintenance scenarios, a batch of stop words are customized, such as common but insignificant words like "de" (的) and "le" (了), to reduce noise interference.
[0054] Part-of-speech tagging optimization: Adjusted the part-of-speech tagging strategy of spaCy to make it more suitable for the characteristics of Chinese operation and maintenance instructions.
[0055] Enhanced entity recognition: Conducted special training on entity recognition for the model to more accurately identify key information in instructions, such as mine names, host names, image names, etc.
[0056] Through the above optimization measures, we expect to improve the natural language processing module's ability to understand operation and maintenance instructions, thereby enhancing the intelligence level of the entire remote intelligent operation and maintenance system.
[0057] The embodiments to be protected in this application include: A remote intelligent operation and maintenance system for the coal mining industry, including: Cloud-based server: Responsible for instruction issuance, data processing, and analysis; Instruction management unit: Publishes upgrade, deployment, and management instructions through RocketMQ, and uses topics and tags to distinguish target mining areas and operation types; Data analysis unit: Collects and analyzes data from on-site mine hosts, uses the XGBoost model to perform anomaly analysis on log data, predicts faults, and issues early warnings; On-site mine client: Deployed on on-site mine hosts to achieve instruction reception and execution, and data collection and upload; Instruction reception and execution unit: Subscribes to RocketMQ messages, parses and executes cloud-based instructions, such as container management and configuration modification; Data collection unit: Monitors system resources, network traffic, Docker container status, and hardware sensor data, and regularly uploads them to the cloud server; Network adaptation unit: Has the function of message forwarding between intranets, and supports instruction and data relay through specified proxy nodes in complex network environments; Natural language processing module: Integrated into the cloud-based server to process voice or text instructions from operation and maintenance personnel; Instruction parsing unit: Based on spaCy, performs text preprocessing, word segmentation, and entity recognition, and converts operation and maintenance instructions into an operation sequence executable by the system; Interactive confirmation unit: Interacts with operation and maintenance personnel to confirm the instruction execution intention, and executes after ensuring the operation is accurate; Encryption and security mechanism: Ensures data transmission security; Data encryption unit: Encrypts uplink and downlink data using the AES-CBC mode to ensure the secure transmission of instructions and data; RocketMQ cluster: provides highly available message queue services; Message queue unit: ensures the reliability and orderliness of command and data transmission, and supports command forwarding and data reporting in complex networks; Brief description of the workflow: Service deployment and upgrade: The cloud server sends instructions through RocketMQ. After receiving the instructions, the mine client performs mirror operations, configuration modifications, database adjustments, and reports the status. Data collection and analysis: The client periodically collects various types of operation and maintenance data, uploads them to the cloud after encryption, and processes them by the data analysis unit to predict faults and issue warnings; Natural language operation: Operation and maintenance personnel issue instructions through voice or text, which are analyzed by the natural language processing module and converted into specific operation instructions, which are executed by the cloud server. Network adaptation and data relay: In a complex network environment, a network adaptation mechanism is formed between clients, and data and command relay is performed through nodes with external network capabilities; In the complex network environment of the coal mining industry, this application improves the operation and maintenance efficiency and intelligence level through a remote intelligent operation and maintenance system, realizes data-driven fault warning, supports natural language interaction, and has efficient service deployment and upgrade capabilities.
[0058] Other Embodiment 1: This application also discloses an intelligent resource optimization and scheduling subsystem, including: Resource monitoring unit: real-time monitoring of system resource usage, including but not limited to CPU, memory, storage, network bandwidth, and I / O resource utilization, to ensure that data is updated to the database in real time; Model optimizer: uses machine learning models, such as reinforcement learning, to predict future demand based on historical resource usage patterns and business loads, dynamically adjust resource allocation strategies, and optimize resource utilization; Dynamic scheduler: Automatically adjusts Docker container resource quotas, such as CPU shares, memory limits, and disk space, based on model output, dynamically scales capacity up or down to keep services running stably and efficiently. Policy execution unit: realizes the automated execution of intelligent scheduling decisions, such as container migration and service balancing between different hosts, ensuring load balancing and avoiding service interruptions caused by single point failures; Extension effect: Based on the basic solution, the intelligent resource optimization and dynamic scheduling system realizes efficient allocation and utilization of resources, automatically adapts to business changes, reduces resource waste, enhances system elasticity and stability, and further reduces operation and maintenance costs; Other Embodiment 2: This application also discloses an intelligent safety protection subsystem, including: Threat intelligence collection unit: real-time tracking of network security trends, collection of the latest threat intelligence, including virus databases, vulnerability information, and attack patterns, to ensure that the system security database is updated; Behavior Analyzer: Uses AI algorithms to learn system behavior patterns, identify abnormal operations, such as abnormal login time, abnormal traffic surge, cache retrieval, and abnormal request patterns, and issue timely warnings; Response unit: Integrates automated defense strategies to immediately perform isolation, blocking, repair, or rollback operations upon detection of security incidents, such as DDoS attacks and malware activities; Strategy Optimizer: Continuously learn and optimize defense strategies, adjust defense thresholds based on historical event feedback, improve recognition accuracy, reduce false positives and false negatives, and enhance protection capabilities; Extension effect: Based on the basic solution, the active defense system strengthens the security protection capability of the system. Through intelligent identification and response mechanisms, it effectively resists network attacks, ensures the security of the operation and maintenance system, reduces potential losses, and improves the operation and maintenance efficiency and the overall stability of the system. Other Embodiment 3: This application also discloses an intelligent knowledge management subsystem, including: Knowledge collection unit: Automatically aggregate system logs, operation logs, troubleshooting cases, and operation and maintenance manuals to build a knowledge base, supporting multi-dimensional indexing; Adaptive learner: Based on NLP technology, it learns the queries, instructions, and operating habits of operation and maintenance personnel, dynamically optimizes search sorting, and makes personalized recommendations to improve the efficiency of knowledge acquisition. Knowledge Updater: Real-time updates, combining external resources (such as the latest technical documents, forums, blogs) with internal operations to continuously enrich the knowledge base and ensure cutting-edge information; Interactive interface: Provides a natural language interactive interface to facilitate operation and maintenance queries, quickly obtain operation guides and troubleshooting solutions, support voice input, and improve operation and maintenance response speed and convenience; Extended effect: This solution not only provides an efficient way to acquire knowledge, but also continuously optimizes the knowledge acquisition process through a personalized learning mechanism, improves the work efficiency of operation and maintenance personnel, shortens the problem-solving cycle, promotes the skill growth of the operation and maintenance team, and adds an important link to the intelligent management of operation and maintenance.
[0059] Finally, it should be noted that the above embodiments are only used to illustrate but not limit the technical solutions of the present invention, and any equivalent substitutions of the present invention and any modifications or partial substitutions that do not depart from the spirit and scope of the present invention should be included in the scope of protection of the claims of the present invention.
Claims
1. A remote intelligent operation and maintenance system for the coal mining industry, characterized in that: include: Cloud server: responsible for issuing instructions, data processing and analysis; Instruction management unit: used to publish upgrade, deployment, and management instructions through RocketMQ, using topics and tags to distinguish target mining areas and operation types; Data analysis unit: used to collect and analyze data from the mine host, use the XGBoost model to perform abnormal analysis on log data, predict faults and issue early warnings; Mine client: used to be deployed on the mine host to receive and execute instructions, and collect and upload data; Instruction receiving and execution unit: used to subscribe to RocketMQ messages, parse and execute cloud instructions, such as container management and configuration modification; Data collection unit: used to monitor system resources, network traffic, Docker container status and hardware sensor data, and upload them to the cloud server at regular intervals.
2. The remote intelligent operation and maintenance system for the coal mining industry according to claim 1 is characterized in that: It also includes a network adaptation unit: it is used to have the function of message forwarding between intranets, and supports command and data relay through designated proxy nodes in complex network environments.
3. The remote intelligent operation and maintenance system for the coal mining industry according to claim 1 is characterized in that: Natural language processing module: integrated into the cloud server to process the voice or text instructions of operation and maintenance personnel.
4. The remote intelligent operation and maintenance system for the coal mining industry according to claim 1 is characterized in that: It also includes an instruction parsing unit: based on spaCy, it performs text preprocessing, word segmentation, and entity recognition to convert operation and maintenance instructions into an operation sequence that can be executed by the system; Interaction confirmation unit: interacts with operation and maintenance personnel to confirm the intention of command execution and ensure that the operation is accurate before execution.
5. The remote intelligent operation and maintenance system for the coal mining industry according to claim 1 is characterized in that: It also includes a data encryption unit: it uses AES-CBC mode to encrypt uplink and downlink data to ensure the secure transmission of instructions and data; RocketMQ cluster: provides highly available message queue services; Message queue unit: ensures the reliability and orderliness of command and data transmission, and supports command forwarding and data reporting in complex networks.
6. The remote intelligent operation and maintenance system for the coal mining industry according to claim 1 is characterized in that: It also includes intelligent resource optimization and scheduling subsystems, including: Resource monitoring unit: real-time monitoring of system resource usage to ensure that data is updated to the database in real time; Model optimizer: uses machine learning models, such as reinforcement learning, to predict future demand based on historical resource usage patterns and business loads, dynamically adjust resource allocation strategies, and optimize resource utilization; Dynamic scheduler: Automatically adjusts Docker container resource quotas based on model output to keep services running stably and efficiently; Policy execution unit: used to realize the automated execution of intelligent scheduling decisions, ensure load balancing, and avoid service interruptions caused by single point failures.
7. The remote intelligent operation and maintenance system for the coal mining industry according to claim 1 is characterized in that: It also includes an intelligent safety protection subsystem, including: Threat intelligence collection unit: real-time tracking of network security trends, collection of the latest threat intelligence, including virus databases, vulnerability information, and attack patterns, to ensure that the system security database is updated; Behavior Analyzer: Uses AI algorithms to learn system behavior patterns, identify abnormal operations, and issue timely warnings; Response unit: Integrates automated defense strategies to immediately perform isolation, blocking, repair, or rollback operations once a security incident is discovered; Strategy Optimizer: Continuously learn and optimize defense strategies, adjust defense thresholds based on historical event feedback, improve recognition accuracy, reduce false positives and false negatives, and enhance protection capabilities.
8. The remote intelligent operation and maintenance system for the coal mining industry according to claim 1 is characterized in that: It also includes an intelligent knowledge management subsystem, including: Knowledge collection unit: used to automatically aggregate system logs, operation logs, troubleshooting cases, and operation and maintenance manuals, build a knowledge base, and support multi-dimensional indexing; Adaptive learner: Based on NLP technology, it learns the queries, instructions, and operating habits of operation and maintenance personnel, dynamically optimizes search sorting, and makes personalized recommendations to improve the efficiency of knowledge acquisition. Knowledge Updater: Real-time updates, combining external resources with internal operations, continuously enriching the knowledge base and ensuring cutting-edge information; Interactive interface: Provides a natural language interactive interface to facilitate operation and maintenance queries, quickly obtain operation guides and troubleshooting solutions, support voice input, and improve operation and maintenance response speed and convenience.
Citation Information
Cited By
Network line intelligent operation and maintenance monitoring management system and method
CN119835143A
Multi-system remote collaborative management system and method for fully mechanized coal mining face of coal mine
CN120031300A
Intelligent acquisition method and system for real-time mine data
CN122087271A
