Intelligent operation and maintenance system of intelligent calculation center

Abstract

The invention relates to the technical field of intelligent operation and maintenance, in particular to an intelligent operation and maintenance system of an intelligent calculation center, which comprises an intelligent calculation center infrastructure sensor system, an operation and maintenance robot acquisition and analysis system and a cloud intelligent calculation center processing system. Wherein the intelligent calculation center infrastructure sensor system comprises an intelligent calculation center infrastructure, the intelligent calculation center infrastructure internally covers communication, signal, data and power supply equipment, and a monitoring sensor, a moving ring sensor, a dangerous temperature sensor and a water immersion sensor are arranged in the intelligent calculation center infrastructure; according to the operation and maintenance robot collection and analysis system, all sensors are connected into a wireless receiver through a unified interface, and the wireless receiver transmits collected data to a robot end in real time. Through AI self-learning, real-time adjustment and optimization, power supply, temperature control, energy efficiency and other aspects are intelligently monitored.

Description

Technical Field

[0001] The present invention relates to the field of intelligent operation and maintenance technology, and in particular to an intelligent operation and maintenance system for an intelligent computing center. Background Art

[0002] The Intelligent Computing Center is a new type of public computing infrastructure based on the latest artificial intelligence theory and the leading artificial intelligence computing architecture. It provides artificial intelligence computing institutions and computing services, data services and algorithm services required for artificial intelligence applications. It plays a significant role in promoting the industrialization of AI, enabling the AIization of industries, helping intelligent governance, and promoting industrial clustering.

[0003] At present, the operation and maintenance of the intelligent computing center is still limited to the traditional manual inspection mode, with many and complex tasks and heavy workload. In some special cases, such as multiple failures, the working environment has become severe, and traditional manual work cannot be handled in a timely and safe manner, resulting in the loss of some important data, causing irreversible damage to the infrastructure and causing huge losses. In order to solve the many problems that exist in the existing operation and maintenance, an AI-based intelligent operation and maintenance system is provided. Summary of the invention

[0004] The purpose of the present invention is to provide an intelligent operation and maintenance system for an intelligent computing center to solve the problems raised in the above-mentioned background technology.

[0005] To achieve the above-mentioned purpose, the present invention provides the following technical solutions: an intelligent operation and maintenance system of an intelligent computing center, comprising three parts: an intelligent computing center infrastructure sensor system, an operation and maintenance robot collection and analysis system, and a cloud-based intelligent computing center processing system;

[0006] The intelligent computing center infrastructure sensor system includes the intelligent computing center infrastructure, which includes communication, signal, data, and power supply equipment. The intelligent computing center infrastructure is equipped with monitoring sensors, dynamic environment sensors, dangerous temperature sensors, and water immersion sensors.

[0007] The operation and maintenance robot collection and analysis system: each sensor is connected to the wireless receiver through a unified interface, and the wireless receiver transmits the collected data to the robot in real time; then the intelligent operation and maintenance robot with a built-in IoT gateway is located at the center of the system, and the system center has a built-in edge computing module to build a model of the corresponding infrastructure, walk to various places according to the command line, collect audio and video data through external sensors in real time, and perform calculations and analysis through the robot edge algorithm platform, so as to differentiate between normalization and accidents, and then immediately feedback the problem information to the remote cloud computing center;

[0008] The cloud-based intelligent computing center processing system includes data preprocessing and analysis, fault detection and location, solution recommendation and optimization, automated operation and maintenance and iteration, as well as visualization and user interaction.

[0009] Optimally, data preprocessing and analysis: eliminate redundant and erroneous data, ensure data consistency and accuracy, flexibly select analysis methods based on data type and timeliness requirements, use deep learning models to identify abnormal patterns, and perform anomaly detection through autoencoders: in Input data, g(x): encoder function, f(g(x)): decoder function, anomaly score: Where x is a given observation.

[0010] Preferably, fault detection and location: Use predictive maintenance models to predict potential faults, provide early warnings, and use time series prediction models (such as ARIMA):

[0011] y t =c+φ1y t-1 +φ2y t-2 +…+φ p y t-p +ε t

[0012] Among them, y t : current value, φ: model parameter, ε t : white noise;

[0013] Analyze log files through natural language processing, quickly identify fault causes, build fault trees, analyze fault propagation paths, quickly locate root causes, and use decision trees for fault diagnosis:

[0014]

[0015] Where, D: dataset, A: attribute, D v : The data subset whose attribute value is v.

[0016] Optimize, recommend and optimize solutions: collect and manage all troubleshooting cases to form a knowledge base; continuously optimize troubleshooting steps through machine learning, and then establish a feedback mechanism.

[0017] Optimize automation and iteration of operation and maintenance: Execute standardized operation and maintenance scripts through AI operation and maintenance robots. After each fault handling, record the operation steps and results, and evaluate the effectiveness of different processing strategies through A / B testing.

[0018] Optimize visualization and user interaction: Provide a real-time updated visualization interface, display key performance indicators and alarms, generate customized reports based on user needs, support decision analysis, and further optimize the visualization interface and data display through user feedback.

[0019] Preferably, the adaptive optimization algorithm includes the following steps:

[0020] Step 1: Data collection and preprocessing;

[0021] Step 2: Model selection and training;

[0022] Step 3: Optimization strategy generation;

[0023] Step 4: Adaptive adjustment, monitor system status and fault handling process, obtain real-time data, dynamically adjust optimization strategy according to real-time data and feedback information, and continuously update model and optimization strategy.

[0024] Preferably, step 1: data collection and preprocessing, using sensors and log collectors to monitor the system status, storing the collected data in the cloud for subsequent analysis, identifying fault signals through anomaly detection algorithms, and triggering feedback mechanisms.

[0025] Preferably, step 2: model selection and training, using a deep neural network machine learning model, training the model with historical data, predicting the possibility of failure and optimizing the processing steps, predicting the model:

[0026]

[0027] Where X: input feature vector, θ k : Model parameters, φ k (X): Feature transformation function, using cross-validation method to evaluate model performance.

[0028] Preferably, step 3: generating an optimization strategy, generating an optimization strategy according to the model prediction results, such as adjusting the fault handling steps and resource allocation, considering multiple objectives, such as processing time, resource consumption, and fault impact, and optimizing the objective function:

[0029]

[0030] Where T i (x): time of the i-th fault handling step, R j (x): consumption of the jth resource, w i , c j : Weight coefficient, indicating the importance of processing time and resource consumption, using heuristic algorithm for multi-objective optimization, heuristic algorithm update:

[0031] x t+1=x t +α·(p t -x t )+β·(g t -x t )

[0032] Among them, x t : Current solution, p t The individual optimal solution, g t : global optimal solution, α, β: learning factors.

[0033] Compared with the prior art, the present invention has the following beneficial effects:

[0034] 1. Through AI self-learning, real-time tuning, intelligent monitoring of power supply, temperature control, energy efficiency, etc.;

[0035] 2. Through AI predictive algorithms, "passive alarms" are transformed into "active AI predictive maintenance" (such as fan and capacitor life prediction, equipment temperature detection);

[0036] 3. With the rapid response capability of AI, once a failure occurs, the traditional manual response is changed to an active response, so as to proactively discover the problem, analyze the problem, and restore the failure;

[0037] 4. Make full use of AI in daily operation and maintenance, and combine it with corresponding intelligent sensing / LOT / 5G / sound recognition / image recognition / optical character recognition and other technologies to transform the traditional manual inspection mode into the AI ​​remote inspection mode;

[0038] 5. Use the operation and maintenance robot to build a three-dimensional model of the corresponding infrastructure. Through inspection, timely identification and response can be made to different situations of the infrastructure (such as safety and danger) in real time, and the model can be uploaded to the cloud computing center in a timely manner;

[0039] The cloud-based intelligent computing center service system created by us integrates functions such as work, inspection, reception, maintenance, and low-power standby (when there is nothing to do) through on-site intelligent operation and maintenance robots. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] In order to more clearly illustrate the specific implementation methods of the present invention or the technical solutions in the prior art, the drawings required for use in the specific implementation methods or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are some implementation methods of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0041] Figure 1 This is a diagram of the intelligent operation and maintenance system of the intelligent computing center of the present invention. DETAILED DESCRIPTION

[0042] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0043] See also Figure 1 , the present invention provides a technical solution:

[0044] An intelligent operation and maintenance system for an intelligent computing center, comprising an intelligent computing center infrastructure sensor system, an operation and maintenance robot collection and analysis system, and a cloud-based intelligent computing center processing system;

[0045] The intelligent computing center infrastructure sensor system includes the intelligent computing center infrastructure, which includes communication, signal, data, and power supply equipment. The intelligent computing center infrastructure is equipped with monitoring sensors, dynamic environment sensors, dangerous temperature sensors, and water immersion sensors.

[0046] Among them, the operation and maintenance robot collection and analysis system: each sensor is connected to the wireless receiver with a unified interface, and the wireless receiver transmits the collected data to the robot end in real time; then the intelligent operation and maintenance robot with a built-in IoT gateway is located at the center of the system, and the system center has a built-in edge computing module to establish a model of the corresponding infrastructure, go to various places according to command lines, collect audio and video data through external sensors in real time, and perform calculations and analysis through the robot edge algorithm platform, so as to differentiate between normal and accident responses, and then immediately feedback the problem information to the remote cloud-based intelligent computing center.

[0047] Cloud computing center processing system:

[0048] System Architecture

[0049] Data collection layer:

[0050] Operation and maintenance robots: deployed on various devices and system nodes, collect real-time data through sensors and log collectors, including temperature, load, bandwidth, error logs and other information.

[0051] Edge devices: Perform preliminary data processing and filtering at the edge to reduce the amount of data transmitted and improve response speed.

[0052] Data transmission layer:

[0053] Secure transmission protocol: Use encryption protocols (such as TLS / SSL) to ensure the security of data during transmission.

[0054] High-bandwidth network: supports large-scale data traffic transmission and ensures real-time performance.

[0055] Data storage layer:

[0056] Cloud storage platform: Use distributed storage systems (such as HDFS, S3) to store massive data and support high-concurrency reading and writing.

[0057] Data warehouse: Use structured data warehouses (such as Amazon Redshift, Google BigQuery) for analysis and storage of historical data.

[0058] Data analysis layer:

[0059] Big data processing framework: Use Spark, Hadoop and other big data processing frameworks to perform batch and stream processing on data.

[0060] Machine learning models: Use tools such as TensorFlow and PyTorch to train models for predictive analysis.

[0061] Visualization layer:

[0062] Visualization platform: Use tools such as Tableau, PowerBI, and Grafana to visualize data and display system status, performance indicators, fault alerts, etc.

[0063] Customized dashboard: Provides a personalized visual interface for different user roles (such as administrators, engineers, and management).

[0064] Intelligent decision-making layer:

[0065] Fault detection and location: Automatically identify fault modes through machine learning algorithms and locate the specific device and location where the fault occurred.

[0066] Solution recommendation: Use knowledge graphs and case libraries to recommend the best solutions.

[0067] Automated repair: Automated repair operations are performed through AI operation and maintenance robots.

[0068] The cloud-based intelligent computing center processing system includes data preprocessing and analysis, fault detection and location, solution recommendation and optimization, automated operation and maintenance and iteration, as well as visualization and user interaction.

[0069] Data preprocessing and analysis: Eliminate redundant and erroneous data, ensure data consistency and accuracy, flexibly select analysis methods based on data type and timeliness requirements, use deep learning models to identify abnormal patterns, and perform anomaly detection through autoencoders: in Input data, g(x): encoder function, f(g(x)): decoder function, anomaly score: Where x is a given observation.

[0070] Fault detection and location: Use predictive maintenance models to predict potential failures and provide early warnings, using time series forecasting models (such as ARIMA):

[0071] y t =c+φ1y t-1 +φ2y t-2 +…+φ p y t-p +ε t

[0072] Among them, y t : current value, φ: model parameter, ε t : white noise;

[0073] Analyze log files through natural language processing, quickly identify fault causes, build fault trees, analyze fault propagation paths, quickly locate root causes, and use decision trees for fault diagnosis:

[0074]

[0075] Where, D: dataset, A: attribute, D v : The data subset whose attribute value is v.

[0076] Solution recommendation and optimization: Collect and manage all troubleshooting cases to form a knowledge base; continuously optimize troubleshooting steps through machine learning, and then establish a feedback mechanism.

[0077] Automated operation and maintenance and iteration: Standardized operation and maintenance scripts are executed through AI operation and maintenance robots. After each fault handling, the operation steps and effects are recorded, and the effects of different processing strategies are evaluated through A / B testing.

[0078] Visualization and user interaction: Provide a real-time updated visualization interface to display key performance indicators and alarms, generate customized reports based on user needs, support decision analysis, and further optimize the visualization interface and data display through user feedback.

[0079] Implementation steps:

[0080] 1. Data collection and monitoring

[0081] Configure sensors and log collectors to collect key parameter data in real time.

[0082] Use edge devices to pre-process data and reduce the data transmission load.

[0083] 2. Data transmission and storage

[0084] Transfer data to cloud storage systems using secure protocols.

[0085] The data warehouse is used to store and analyze structured data.

[0086] 3. Data analysis and fault diagnosis

[0087] Use big data frameworks for batch and stream processing.

[0088] Identify potential faults and locate the root cause of the problem through anomaly detection models.

[0089] 4. Solution recommendation and implementation

[0090] Search the case library and recommend the best solution.

[0091] Use AI operation and maintenance robots to perform automated repairs and record operation steps.

[0092] 5. Feedback and Optimization

[0093] Collect feedback information, analyze processing effects, and update the knowledge base.

[0094] Optimize process flow through iterative learning and A / B testing.

[0095] 6. Visualization and report generation

[0096] Visualize key indicators and process results.

[0097] Generate customized reports based on demand to support decision-making analysis.

[0098] Improve troubleshooting efficiency: Automated repair and intelligent decision-making reduce manual intervention and increase processing speed.

[0099] Reduce operation and maintenance costs: reduce the use of flow meters, optimize the amount of flocculants, and reduce operating costs.

[0100] Enhance system stability: Ensure continuous and stable operation of the system through dynamic monitoring and feedback mechanisms.

[0101] Accumulate knowledge and experience: Continuously accumulate experience through case libraries and knowledge graphs to improve processing accuracy.

[0102] The adaptive optimization algorithm includes the following steps:

[0103] Step 1: Data collection and preprocessing, using sensors and log collectors to monitor system status, storing the collected data in the cloud for subsequent analysis, identifying fault signals through anomaly detection algorithms, and triggering feedback mechanisms.

[0104] Step 2: Model selection and training, using a deep neural network machine learning model, training the model with historical data, predicting the possibility of failure and optimizing the processing steps, predicting the model:

[0105]

[0106] Where X: input feature vector, θ k : Model parameters, φ k (X): Feature transformation function, using cross-validation method to evaluate model performance.

[0107] Step 3: Generate optimization strategies. Generate optimization strategies based on the model prediction results, such as adjusting fault handling steps and resource allocation, considering multiple goals, such as processing time, resource consumption, and fault impact, and optimizing the objective function:

[0108]

[0109] Where T i (x): time of the i-th fault handling step, R j (x): consumption of the jth resource, w i , c j : Weight coefficient, indicating the importance of processing time and resource consumption, using heuristic algorithm for multi-objective optimization, heuristic algorithm update:

[0110] x t+1 =x t +α·(p t -x t )+β·(g t -x t )

[0111] Among them, x t : Current solution, p t The individual optimal solution, g t : global optimal solution, α, β: learning factors

[0112] Through the above optimizations and improvements, the cloud-based intelligent computing center processing system will achieve more efficient operation and maintenance management, more accurate fault detection and more optimized solutions, and improve the reliability and response speed of the system.

[0113] Step 4: Adaptive adjustment, monitor system status and fault handling process, obtain real-time data, dynamically adjust optimization strategy according to real-time data and feedback information, and continuously update model and optimization strategy.

[0114] The feedback loop mechanism includes:

[0115] Step 1: Data Collection and Monitoring

[0116] 1. Real-time monitoring: Use sensors and log collectors to monitor system status.

[0117] 2. Data storage: Store the collected data in the cloud for subsequent analysis.

[0118] 3. Anomaly detection: Identify fault signals through anomaly detection algorithms and trigger feedback mechanisms.

[0119] Step 2: Feedback information analysis

[0120] 1. Fault analysis: Analyze the cause and impact of the fault and identify key issues.

[0121] 2. Optimization suggestions: Generate optimization suggestions based on the analysis results to guide troubleshooting.

[0122] 3. Learning and improvement: Use feedback information to update models and optimize strategies to improve processing results.

[0123] Feedback Function

[0124] F(x)=α·Δx+β·(x t -x t-1 )

[0125] : The deviation between the current state and the expected state.

[0126] α, β: feedback control coefficients.

[0127] Step 3: Optimization and Adjustment

[0128] 1. Automated adjustment: Automatically adjust fault handling steps and resource allocation based on feedback information.

[0129] 2. Continuous optimization: Continuously optimize processing processes and strategies to improve system efficiency.

[0130] Optimization Update

[0131] x t+1 =x t +γ·F(x t ).

[0133] : learning rate.

[0134] 3. Iterative updates: Regularly evaluate and update the model to ensure its adaptability and accuracy.

[0135] Adaptive Adjustment

[0136]

[0137] : Model parameters.

[0138] : learning rate.

[0139] 4.: Gradient of the loss function.

[0140] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or replace some or all of the technical features therein with equivalents. However, these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. An intelligent operation and maintenance system for an intelligent computing center, characterized in that: It consists of three parts: the intelligent computing center infrastructure sensor system, the operation and maintenance robot collection and analysis system, and the cloud-based intelligent computing center processing system; The intelligent computing center infrastructure sensor system includes the intelligent computing center infrastructure, which includes communication, signal, data, and power supply equipment. The intelligent computing center infrastructure is equipped with monitoring sensors, dynamic environment sensors, dangerous temperature sensors, and water immersion sensors. The operation and maintenance robot collection and analysis system: each sensor is connected to the wireless receiver through a unified interface, and the wireless receiver transmits the collected data to the robot end in real time; Then, the intelligent operation and maintenance robot with built-in IoT gateway is located at the center of the system. The system center also has built-in edge computing modules to build a model of the corresponding infrastructure. It moves to various places according to command lines, collects audio and video data in real time through external sensors, and performs calculations and analysis through the robot edge algorithm platform. It responds differently to normalization and accidents, and then immediately feeds back problem information to the remote cloud computing center. The cloud-based intelligent computing center processing system includes data preprocessing and analysis, fault detection and location, solution recommendation and optimization, automated operation and maintenance and iteration, as well as visualization and user interaction.

2. The intelligent operation and maintenance system of the intelligent computing center according to claim 1, characterized in that: The data preprocessing and analysis described above: eliminate redundant and erroneous data, ensure data consistency and accuracy, flexibly select analysis methods based on data type and timeliness requirements, use deep learning models to identify abnormal patterns, and perform anomaly detection through autoencoders: in Input data, g(x): encoder function, f(g(x)): decoder function, anomaly score: Where x is a given observation.

3. The intelligent operation and maintenance system of the intelligent computing center according to claim 2 is characterized by: Fault detection and location: Use predictive maintenance models to predict potential faults, provide early warnings, and use time series prediction models (such as ARIMA): y t =c+φ1y t-1 +φ2y t-2 +…+φ p y t-p +e t Among them, y t : current value, φ: model parameter, ε t : white noise; Analyze log files through natural language processing, quickly identify fault causes, build fault trees, analyze fault propagation paths, quickly locate root causes, and use decision trees for fault diagnosis: Where, D: dataset, A: attribute, D v : The data subset whose attribute value is v.

4. The intelligent operation and maintenance system of the intelligent computing center according to claim 3 is characterized by: The solution recommendation and optimization described above: collect and manage all troubleshooting cases to form a knowledge base; continuously optimize the troubleshooting steps through machine learning, and then establish a feedback mechanism.

5. The intelligent operation and maintenance system of the intelligent computing center according to claim 4, characterized in that: The automated operation and maintenance and iteration described above: Standardized operation and maintenance scripts are executed through AI operation and maintenance robots. After each fault handling, the operation steps and effects are recorded, and the effects of different processing strategies are evaluated through A / B testing.

6. The intelligent operation and maintenance system of the intelligent computing center according to claim 5, characterized in that: The visualization interacts with the user: providing a real-time updated visualization interface, displaying key performance indicators and alarms, generating customized reports based on user needs, supporting decision analysis, and further optimizing the visualization interface and data display through user feedback.

7. The intelligent operation and maintenance system of the intelligent computing center according to claim 6, characterized in that: The adaptive optimization algorithm includes the following steps: Step 1: Data collection and preprocessing; Step 2: Model selection and training; Step 3: Optimization strategy generation; Step 4: Adaptive adjustment, monitor system status and fault handling process, obtain real-time data, dynamically adjust optimization strategy according to real-time data and feedback information, and continuously update model and optimization strategy.

8. The intelligent operation and maintenance system of the intelligent computing center according to claim 7, characterized in that: The step 1: data collection and preprocessing, using sensors and log collectors to monitor system status, storing the collected data in the cloud for subsequent analysis, identifying fault signals through anomaly detection algorithms, and triggering feedback mechanisms.

9. The intelligent operation and maintenance system of the intelligent computing center according to claim 8, characterized in that: Step 2: Model selection and training, using a deep neural network machine learning model, training the model with historical data, predicting the possibility of failure and optimizing the processing steps, predicting the model: Where X: input feature vector, θ k : Model parameters, φ k (X): Feature transformation function, using cross-validation method to evaluate model performance.

10. The intelligent operation and maintenance system of the intelligent computing center according to claim 9, characterized in that: Step 3: Optimization strategy generation, generating an optimization strategy based on the model prediction results, such as adjusting the fault handling steps and resource allocation, considering multiple goals, such as processing time, resource consumption, and fault impact, and optimizing the objective function: Where T i (x): time of the i-th fault handling step, R j (x): consumption of the jth resource, w i , c j : Weight coefficient, indicating the importance of processing time and resource consumption, using heuristic algorithm for multi-objective optimization, heuristic algorithm update: x t+1 =x t +α·(p t -x t )+β·(g t -x t ) Among them, x t : Current solution, p t The individual optimal solution, g t : global optimal solution, α, β: learning factors.

Images