IT equipment monitoring system and method and storage medium thereof
By combining real-time monitoring and fault prediction model training with adaptive adjustment, the problem of the lack of predictive maintenance in IT equipment monitoring systems has been solved, thereby improving the reliability and stability of equipment operation and reducing operating costs.
Patent Information
- Application Number
- CN202411245277.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-06
- Publication Date
- 2024-11-12
- Estimated Expiration
- Not applicable · inactive patent
AI Technical Summary
Existing IT equipment monitoring systems lack predictive maintenance and adaptive optimization capabilities, making it difficult to prevent equipment failures, affecting equipment lifespan and operating costs, and causing discontinuity in equipment operation.
It employs a real-time monitoring module, a model training module, a predictive maintenance module, and a user interface and intelligent alarm module. Through real-time data acquisition, fault prediction model training, and adaptive adjustment of equipment operating parameters, it achieves predictive maintenance and adaptive optimization.
It improves the reliability and stability of equipment operation, extends equipment life, reduces operating costs, and ensures continuous operation of the equipment.
Smart Images

Figure CN118938780A_ABST
Abstract
Description
Technical Field
[0001] The present invention relates to the technical field of equipment monitoring, and more specifically, to an IT equipment monitoring system, method and storage medium thereof. Background Art
[0002] With the development of science and technology, the number and types of modern equipment are increasing, and the stable operation of modern equipment is crucial to the daily operation of the equipment. For example, in the medical industry, the rapid development of science and technology in the medical industry has gradually increased the demand for oxygen in hospitals, and automated oxygen production equipment has emerged. Among them, the stable operation of automated oxygen production equipment is crucial to the daily operation of hospitals.
[0003] In the prior art, IT equipment is usually used to monitor the operation of equipment.
[0004] However, traditional IT equipment monitoring systems usually lack predictive maintenance and adaptive optimization functions for equipment operation. Equipment failures are often difficult to prevent. When the equipment alarms for abnormal operation, it is too late. Equipment abnormal operation shortens the equipment life, wastes energy, and increases operating costs. At the same time, it also affects the continuity of equipment operation. Therefore, a new monitoring system is urgently needed to improve the reliability and stability of equipment operation. Summary of the invention
[0005] In view of the technical problem raised in the background technology that the IT equipment monitoring system in the prior art usually lacks predictive maintenance and adaptive optimization functions for equipment operation, the present invention provides an IT equipment monitoring system based on algorithm prediction.
[0006] The present invention provides the following technical solution 1:
[0007] An IT equipment monitoring system, comprising:
[0008] Real-time monitoring module: The real-time monitoring module is responsible for collecting various operating data of the equipment, providing the evaluation results of the current equipment operating status through analysis, and then uploading the equipment operating data to the cloud;
[0009] Model training module: The model training module is responsible for designing the fault prediction model and training the fault prediction model using historical data in the cloud;
[0010] Predictive maintenance module: The predictive maintenance module is responsible for real-time fault prediction based on the fault prediction model and feeding the real-time data of equipment operation into the fault prediction model to calculate and generate the final fault probability. At the same time, it is also responsible for real-time control of equipment to adjust operating parameters based on the final fault probability to prevent possible future faults.
[0011] The user interface and intelligent alarm module are responsible for displaying the equipment operating parameters on the user interface, identifying whether the probability of failure exceeds the set threshold, giving early alarm, and providing feedback on the operating status of the adjusted operating parameters to the equipment management personnel.
[0012] The present invention provides the following technical solution 2:
[0013] An IT equipment monitoring method comprises the following steps:
[0014] S1. Collecting equipment operation data: Using the monitoring module to collect various operation data of the equipment, and then uploading the equipment operation data to the cloud;
[0015] S2. Design a fault prediction model;
[0016] S3, model training: Use the two historical data of normal operation and failure occurrence in the cloud to train the designed fault prediction model;
[0017] S4. Calculate the failure probability: Based on the failure prediction model, feed the real-time data of equipment operation into the failure prediction model for real-time failure prediction, and calculate and generate the final failure probability;
[0018] S5, intelligent alarm: identify the final failure probability and send an alarm;
[0019] S6, Adaptive adjustment: Based on the final failure probability, the equipment is controlled in real time to adjust the operating parameters to prevent possible future failures;
[0020] S7, adjustment feedback: Identify the adaptive adjustment of equipment operating parameters and feedback the adjustment results.
[0021] The present invention provides the following technical solution three:
[0022] A storage medium stores a computer program, which, when executed by a processor, implements the IT equipment monitoring method in technical solution 2.
[0023] In summary, the present invention has the following beneficial effects:
[0024] (1) Improve the operational reliability of equipment;
[0025] (2) Improve the predictive maintenance function of equipment. BRIEF DESCRIPTION OF THE DRAWINGS
[0026] Figure 1 It is the structural principle diagram of IT equipment monitoring system;
[0027] Figure 2 It is the structural principle diagram of the real-time monitoring module;
[0028] Figure 3 This is the structural schematic diagram of the model training module;
[0029] Figure 4 This is the schematic diagram of the predictive maintenance module structure;
[0030] Figure 5 This is the structural schematic diagram of the user interface and intelligent alarm module.
[0031] Figure markings: 1. Real-time monitoring module; 1-1. Sensor data acquisition unit; 1-2. Data upload unit; 1-3. Real-time data processing and complex analysis unit; 2. Model training module; 2-1. Historical data acquisition and preprocessing unit; 2-2. Fault prediction model design and training unit; 3. Predictive maintenance module; 3-1. Real-time fault prediction and maintenance suggestion generation unit; 3-2. Adaptive adjustment unit of equipment operating parameters; 4. User interface and intelligent alarm module; 4-1. Data download unit; 4-2. User interface design and data display unit; 4-3. Intelligent alarm unit; 4-4. Feedback unit. DETAILED DESCRIPTION
[0032] The present invention is further described in detail below in conjunction with the embodiments and drawings, but the embodiments of the present invention are not limited thereto. In the following embodiments, except for IT equipment, other equipment refers to the objects monitored by the IT equipment - the monitored equipment, "IT" refers to computer technology, the full name of which is Internet Technology in English, and IT equipment refers to computer technology equipment used to monitor the monitored equipment.
[0033] An IT equipment monitoring system is mainly used to analyze the future failure probability of the equipment based on the equipment's operating parameters, obtain a predicted value of the future failure probability, and provide a predictive maintenance function for the equipment, so as to more intelligently know the possible future failure types of the equipment in advance and perform some adaptive fault recovery operations in advance.
[0034] Combination Figure 1-Figure 5 As shown, the IT equipment monitoring system includes a real-time monitoring module 1, a model training module 2, a predictive maintenance module 3, and a user interface and intelligent alarm module 4.
[0035] The real-time monitoring module 1 is responsible for real-time monitoring of the operating status of the equipment, collecting various operating parameters of the equipment during operation, accurately evaluating the operating status of the equipment by analyzing various operating parameters, and providing the current equipment operating status evaluation results so as to intuitively understand the current equipment operating status, and then uploading the equipment operating parameters to the cloud storage through network communication technologies such as the Internet of Things or the Internet. The functional units of the real-time monitoring module 1 include a sensor data acquisition unit 1-1, a data upload unit 1-2, and a real-time data processing and complex analysis unit 1-3. The functions of each subunit are as follows.
[0036] The sensor data acquisition unit 1-1 is used to achieve high-precision data acquisition and ensure real-time monitoring of key parameters of the equipment. The core steps of data acquisition include: ① Sensor layout optimization, that is, deploying sensors at key locations of the equipment to ensure full coverage. ② Collection frequency optimization, that is, optimizing the collection frequency through historical data analysis. Optimizing the collection frequency means setting different collection times for different periods of equipment operation, so that it can not only reflect the equipment status in real time, but also reduce data redundancy.
[0037] At the same time, when arranging sensors, multiple sensors are used to collect multiple data for one operating parameter, and the multi-sensor data is integrated through a weighted fusion algorithm to reduce the impact of single sensor errors. Usually, multiple sensors need to be arranged at different measurement locations of the equipment to generate high-precision real-time monitoring data, providing a basis for subsequent analysis and decision-making.
[0038] The data fusion formula of multiple sensors is described as follows:
[0039] Where D(t) is the fused data at time t,
[0040] S i (t) is the reading of the ith sensor, w i is the weight of the ith sensor.
[0041] The data upload unit 1-2 is used to upload data to the cloud, and to achieve efficient and secure data transmission, ensuring the real-time and consistency of data. Data upload usually requires data encryption and compression, that is, before data upload, the data is encrypted and compressed to ensure the security and efficiency of transmission. The network protocol for data upload can adopt the Internet of Things protocol: lightweight and efficient Internet of Things protocols such as MQTT to optimize data transmission speed and reduce network delay.
[0042] The real-time data processing and complex analysis unit 1-3 is used to provide the evaluation results of the current equipment operation status. Among them, the data processed by the real-time data processing and complex analysis unit 1-3 is the data uploaded to the cloud in real time by the data upload unit 1-2. When working, the real-time data processing and complex analysis unit 1-3 first needs to eliminate the noise and outliers in the data to ensure the data quality. In one embodiment, a Kalman filter can be used for data filtering to smooth the random noise in the data; a statistically based anomaly detection algorithm (DBSCAN clustering algorithm) is used to identify and remove outliers to ensure data quality.
[0043] Then, the real-time analysis algorithm is used to analyze and evaluate the current operating status of the equipment. The real-time analysis algorithm formula is described as follows:
[0044]
[0045] Among them, A(t) is the analysis result at time t, D(t) is the data at the current time, θ is the smoothing factor,
[0046] η is the dynamic adjustment coefficient, is the rate of change of the data.
[0047] The real-time analysis algorithm calculates the real-time data change rate first. Combined with historical data A(t-1), the current equipment status A(t) is evaluated, thereby ultimately outputting real-time equipment status analysis results and providing reliable equipment operating status evaluation.
[0048] The model training module 2 is used to design a fault prediction model and train the model using historical data. Designing a fault prediction model refers to selecting and mixing multiple prediction models to form a fault prediction model to improve the accuracy of subsequent predictions. Model training refers to training a fault prediction model using historical data.
[0049] In this embodiment, the fault prediction model is configured as three prediction models of the hybrid model, namely: random forest prediction model (RandomForest), support vector machine prediction model (SVM) and deep neural network prediction model (DNN) ensemble learning. The description of the three prediction models is as follows:
[0050] 1. The prediction formula of random forest is:
[0051] Among them, P f is the predicted value of failure probability,
[0052] N is the number of decision trees, h i (X) is the prediction result of the i-th tree for the input data X.
[0053] 2. Prediction formula of support vector machine:
[0054] in,
[0055] f(x) is the predicted value of the failure probability, x is the input data, x i is the support vector, α i and i
[0056] are the coefficients and labels of the support vectors, K(x i ,x) is the kernel function and b is the bias.
[0057] 3. The structure of the deep neural network is:
[0058] -Input layer: receives feature data obtained from feature extraction;
[0059] -Multiple hidden layers: Each layer uses nonlinear activation functions such as ReLU to capture complex patterns in the data;
[0060] -Output layer: output the failure probability P of the device.
[0061] The functional units of the model training module 2 include a historical data collection and preprocessing unit 2-1 and a fault prediction model design and training unit 2-2.
[0062] The historical data collection and preprocessing unit 2-1 is used to download the historical data of the equipment from the cloud and ensure that it covers both normal operation and failure occurrence historical data, and then cleans the data by using statistical methods and machine learning algorithms (isolation forest algorithm) to remove abnormal data and fill in missing values, thereby ensuring the quality of input data in the design and training unit 2-2 of the fault prediction model.
[0063] The fault prediction model design and training unit 2-2 is used to design a fault prediction model and use historical data to train the model. The fault prediction model designed in this embodiment is as described above, including a hybrid model of a random forest prediction model (RandomForest), a support vector machine prediction model (SVM) and a deep neural network prediction model (DNN). After the fault prediction model is completed, historical data is used to train the model. The historical data comes from the output of the historical data collection and preprocessing unit 2-1.
[0064] The model training process in Unit 2-2 of Design and Training of Fault Prediction Model is to input historical data into the fault prediction model. Since the historical data covers both normal operation and fault occurrence historical data, the fault prediction model can be repeatedly trained to adjust the fault prediction model formula / structure to ensure that the fault prediction results of the determined fault prediction model have the highest accuracy.
[0065] The predictive maintenance module 3 is used to input real-time data (the latest operating data of the current equipment) and perform real-time fault prediction.
[0066] The functional units of the predictive maintenance module 3 include a real-time fault prediction and maintenance suggestion generation unit 3-1 and an equipment operation parameter adaptive adjustment unit 3-2.
[0067] The real-time fault prediction and maintenance suggestion generation unit 3-1 inputs real-time data (the latest operating data of the current device) and calculates the current failure probability of the device. According to the above content, the failure probability model of the present invention is a hybrid model of a random forest prediction model (RandomForest), a support vector machine prediction model (SVM) and a deep neural network prediction model (DNN). Therefore, by inputting the latest operating data of the current device into the failure probability model of the present invention, three failure probabilities can be obtained.
[0068] Then use the following failure probability calculation formula in the real-time failure prediction and maintenance suggestion generation unit 3-1 to calculate the final failure probability:
[0069] Formula description:
[0070] Among them, P f (t) is the probability of failure, f i (t)
[0071] is the prediction result of different models, w i is the model weight.
[0072] The algorithm formula is to calculate the failure probability of the current device at a given time t by weighted integration of the prediction results of different models. In one embodiment, Bayesian optimization can be used to adjust the model weight w i , dynamically update the weight distribution according to the performance of historical forecasts to improve the accuracy of forecasts.
[0073] At the same time, the real-time fault prediction and maintenance suggestion generation unit 3-1 generates specific maintenance suggestions based on the final failure probability calculated, such as the need to replace parts, adjust operating parameters, etc.
[0074] The adaptive adjustment unit 3-2 of the equipment operating parameters is used to implement a feedback control mechanism, that is, the control device system adjusts the equipment operating parameters in real time according to the fault prediction results to prevent possible faults from occurring in the future.
[0075] The method for adjusting the equipment operating parameters is performed using the following adaptive adjustment algorithm formula, which is described as:
[0076] in,
[0077] O(t) is the optimized operating parameter, P f (t) is the failure probability, R(t) is the current equipment operating parameters, is the failure probability P f (t) is the cumulative integral at time t,
[0078] is the failure probability P f (t) is the rate of change, α, β, γ, ζ are the adjustment coefficients
[0079] The cumulative integral of the failure probability at time t in the formula represents the cumulative effect of the equipment failure probability over the past period of time. The cumulative integral can reflect the long-term accumulation of failure risks, thereby affecting the current adjustment of operating parameters. Regarding the rate of change of the failure probability, this term is used to capture the changing trend of the failure probability, such as whether the failure probability is rising or falling. By considering the rate of change of the failure probability, the system can respond to potential problems more quickly.
[0080] The algorithm combines the real-time fault probability, current equipment status, fault probability change rate and the cumulative effect of fault probability to dynamically adjust equipment operating parameters and balance equipment performance and safety. Among them, the use of reinforcement learning algorithms can be considered to gradually optimize the adjustment coefficient through simulation and actual feedback to improve the accuracy of equipment adaptive adjustment.
[0081] The adaptive adjustment unit 3-2 of the equipment operating parameters ultimately generates optimized equipment operating parameters, reduces the probability of failure, extends the service life of the equipment, and improves energy efficiency.
[0082] The user interface and intelligent alarm module 4 is used to display device parameters on the user interface and to identify potential abnormalities of the device and issue an alarm in advance.
[0083] The functional units of the user interface and intelligent alarm module 4 include a data download unit 4-1, a user interface design and data display unit 4-2, an intelligent alarm unit 4-3 and a feedback unit 4-4.
[0084] The data download unit 4-1 downloads the data (including real-time data and historical data) uploaded to the cloud by the data upload unit 1-2 from the cloud server, and the data download unit 4-1 transmits the downloaded device operation data to the user interface design and data display unit 4-2.
[0085] User interface design and data display unit 4-2 can support historical data query. To enhance the readability of data visualization, Bezier curve is used to smoothly display data trend. The formula description is:
[0086] B(t) is the curve point, P iis the control point and t is the time point, so that an intuitive and interactive interface can be output to reflect the device status and historical trends in real time.
[0087] The intelligent alarm unit 4-3 identifies real-time fault predictions and generates maintenance suggestions. The unit 3-1 calculates and generates the final fault probability and determines whether it exceeds the set threshold. Based on the severity of the anomaly (exceeding the set threshold), the system sends an alarm through a priority management algorithm to ensure that key issues are handled first.
[0088] The feedback unit 4-4 identifies the adaptive adjustment unit 3-2 of the equipment operating parameters and adjusts the equipment operating parameters in real time according to the fault prediction results, so that after learning the alarm of the fault prediction value through the intelligent alarm unit 4-3, the system can learn the key problem processing results of the fault prediction value through the feedback unit 4-4.
[0089] The following is a detailed description of an oxygen production device in a hospital as an example.
[0090] The sensor data acquisition unit 1-1 collects parameters such as oxygen concentration %, oxygen delivery flow ml, flow rate m / s, pressure pa, temperature ℃, etc. of the oxygen production equipment. The data upload unit 1-2 uploads the operating parameters of the oxygen production equipment to the cloud server. The real-time data processing and complex analysis unit 1-3 first eliminates noise and outliers in the uploaded data to ensure data quality, and then uses the real-time analysis algorithm to analyze and evaluate the current operating status of the equipment. The real-time analysis algorithm formula is described as follows:
[0091]
[0092] Among them, A(t) is the analysis result at time t, D(t) is the data at the current time, θ is the smoothing factor,
[0093] η is the dynamic adjustment coefficient, is the rate of change of the data.
[0094] Substitute each oxygen production parameter into D(t) in the real-time analysis algorithm formula and calculate the analysis result A(t). Then, the operating status of the current oxygen production equipment can be evaluated based on the numerical value of the analysis result A(t) and the numerical range.
[0095] In addition to evaluating the current operating status of the equipment, the IT equipment monitoring system of the present invention also downloads the historical data of the oxygen production equipment from the cloud server through the historical data collection and preprocessing unit 2-1, and feeds the historical data to the fault prediction model in the fault prediction model design and training unit 2-2. The fault prediction model includes a hybrid model of a random forest prediction model (RandomForest), a support vector machine prediction model (SVM) and a deep neural network prediction model (DNN). The hybrid model is trained using historical data, and the prediction model formula or model structure is adjusted to ensure that the fault prediction result of the prediction model is accurate. After the fault prediction model is designed and trained, the predictive maintenance module 3 gives the final fault probability.
[0096] Specifically, the real-time fault prediction and maintenance suggestion generation unit 3-1 uses the following fault probability calculation formula to calculate and generate the final fault probability:
[0097] Formula description:
[0098] Among them, P f (t) is the probability of failure, f i (t)
[0099] is the prediction result of different models, w i is the model weight.
[0100] The algorithm formula is to integrate the prediction results of different models by weighting P f , f(x) and P, calculate the failure probability P of the current device at a given time t f (t). At the same time, the real-time fault prediction and maintenance suggestion generation unit 3-1 generates specific maintenance suggestions based on the final fault probability generated by calculation, such as the need to replace oxygen production components, adjust oxygen production operation parameters, etc.
[0101] In determining the final failure probability P f (t), the adaptive adjustment unit 3-2 of the equipment operating parameters is used to implement the feedback control mechanism, that is, to control the equipment system according to the fault probability P f (t) Adjust equipment operating parameters in real time to prevent possible future failures.
[0102] The method for adjusting the equipment operating parameters is performed using the following adaptive adjustment algorithm formula, which is described as:
[0103] in,
[0104] O(t) is the optimized operating parameter, P f (t) is the failure probability, R(t) is the current equipment operating parameters, is the failure probability Pf (t) is the cumulative integral at time t,
[0105] is the failure probability P f (t) is the rate of change, α, β, γ, ζ are the adjustment coefficients
[0106] The cumulative integral of the failure probability at time t in the formula represents the cumulative effect of the failure probability of the oxygen production equipment over the past period of time. The cumulative integral can reflect the long-term accumulation of failure risks, thereby affecting the current adjustment of operating parameters. Regarding the rate of change of the failure probability, this term is used to capture the changing trend of the failure probability, such as whether the failure probability is rising or falling. By considering the rate of change of the failure probability, the system can respond to potential problems more quickly.
[0107] Therefore, the predictive maintenance module 3 can calculate the final failure probability of the hospital oxygen production equipment and implement adaptive parameter adjustment.
[0108] Finally, the user interface design and data display unit 4-2 displays the real-time operation data of the oxygen generator on the user interface and provides a historical data query function. f (t) When the fault probability threshold is met, the intelligent alarm unit 4-3 identifies the final fault probability P f (t), and based on the severity of the anomaly, the system sends an alarm through a priority management algorithm to ensure that key issues are handled first.
[0109] At the same time, since the adaptive adjustment unit 3-2 of the equipment operating parameters has the function of controlling the oxygen production equipment system according to the failure probability P f (t) The function of adjusting the equipment operating parameters in real time, the feedback unit 4-4 identifies the adjustment result of the equipment operating parameters adjusted by the adaptive adjustment unit 3-2 in real time according to the fault prediction result, and feeds back the adjustment result to the oxygen production equipment manager.
[0110] Finally, based on the above system structure, the present invention also provides:
[0111] A method for monitoring IT equipment. The method details refer to the above system structure. The main operation steps are:
[0112] S1. Collecting equipment operation data: Using the monitoring module to collect various operation data of the equipment, and then uploading the equipment operation data to the cloud;
[0113] S2. Design a fault prediction model;
[0114] S3, model training: Use the two historical data of normal operation and failure occurrence in the cloud to train the designed fault prediction model;
[0115] S4. Calculate the failure probability: Based on the failure prediction model, feed the real-time data of equipment operation into the failure prediction model for real-time failure prediction, and calculate and generate the final failure probability;
[0116] S5, intelligent alarm: identify the final failure probability and send an alarm;
[0117] S6, Adaptive adjustment: Based on the final failure probability, the equipment is controlled in real time to adjust the operating parameters to prevent possible future failures;
[0118] S7, adjustment feedback: Identify the adaptive adjustment of equipment operating parameters and feedback the adjustment results.
[0119] A storage medium stores a computer program, which, when executed by a processor, implements the IT equipment monitoring method in technical solution 2.
[0120] The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above embodiments. All technical solutions under the concept of the present invention belong to the protection scope of the present invention. It should be pointed out that for ordinary technicians in this technical field, some improvements and modifications without departing from the principle of the present invention should also be regarded as the protection scope of the present invention.
Claims
1. An IT equipment monitoring system, characterized in that: include: Real-time monitoring module: The real-time monitoring module is responsible for collecting various operating data of the equipment, providing the evaluation results of the current equipment operating status through analysis, and then uploading the equipment operating data to the cloud; Model training module: The model training module is responsible for designing the fault prediction model and training the fault prediction model using historical data in the cloud; Predictive maintenance module: The predictive maintenance module is responsible for real-time fault prediction based on the fault prediction model and feeding the real-time data of equipment operation into the fault prediction model to calculate and generate the final fault probability. At the same time, it is also responsible for real-time control of equipment to adjust operating parameters based on the final fault probability to prevent possible future faults. The user interface and intelligent alarm module are responsible for displaying the equipment operating parameters on the user interface, identifying whether the probability of failure exceeds the set threshold, giving early alarm, and providing feedback on the operating status of the adjusted operating parameters to the equipment management personnel.
2. An IT equipment monitoring system according to claim 1, characterized in that: The real-time monitoring module comprises: Sensor data acquisition unit, used to realize data collection and ensure real-time monitoring of equipment parameters; A data uploading unit, used to upload the collected data to the cloud; The real-time data processing and complex analysis unit is used to analyze and evaluate the current operating status of the equipment using the real-time analysis algorithm. The real-time analysis algorithm formula is described as follows: Among them, A(t) is the analysis result at time t, D(t) is the data at the current time, θ is the smoothing factor, η is the dynamic adjustment coefficient, is the rate of change of the data.
3. An IT equipment monitoring system according to claim 2, characterized in that: The model training module includes: Historical data collection and preprocessing unit, used to download historical data from the cloud and ensure that both normal operation and fault occurrence historical data are covered; A fault prediction model design and training unit is used to design a fault prediction model, which includes a hybrid model of a random forest prediction model (RandomForest), a support vector machine prediction model (SVM) and a deep neural network prediction model (DNN); The prediction formula of random forest is: Among them, P f is the predicted value of failure probability, N is the number of decision trees, h i (X) is the prediction result of the i-th tree for the input data X. The prediction formula of support vector machine is: in, f(x) is the predicted value of the failure probability, x is the input data, x i is the support vector, α i and i are the coefficients and labels of the support vectors, K(x i ,x) is the kernel function and b is the bias. The structure of a deep neural network is: -Input layer: receives feature data obtained from feature extraction; -Multiple hidden layers: Each layer uses nonlinear activation functions such as ReLU to capture complex patterns in the data; -Output layer: output device failure probability P; The unit is also used to perform model training on the fault prediction model using two types of historical data.
4. The IT equipment monitoring system according to claim 3, characterized in that: The predictive maintenance module includes: The real-time fault prediction and maintenance suggestion generation unit is used to calculate the final fault probability using the following fault probability calculation formula: Formula description: Among them, P f (t) is the probability of failure, f i (t) is the prediction result of different models, w i is the model weight; The adaptive adjustment unit of the equipment operating parameters is used to control the equipment system to adjust the equipment operating parameters in real time according to the fault prediction results to prevent possible faults from occurring in the future.
5. The IT equipment monitoring system according to claim 4, characterized in that: The user interface and intelligent alarm module include: Data download unit, used to download equipment operation data from the cloud, including real-time data and historical data; A user interface design and data display unit, used to receive the data from the data download unit and display it to the user for reading; An intelligent alarm unit is used to identify the final failure probability calculated by the real-time failure prediction and maintenance suggestion generation unit, determine whether it exceeds a set threshold, and send an alarm through a priority management algorithm according to the severity of the anomaly to ensure that key issues are handled first; The feedback unit is used to feed back the operating status of the operating parameters after the adaptive adjustment unit of the operating parameters of the equipment adjusts them.
6. The IT equipment monitoring system according to claim 2, characterized in that: The sensor data acquisition unit uses multiple sensors to collect multiple data for one operating parameter of the equipment, and the multiple sensors are arranged at different measurement positions of the equipment, thereby generating high-precision real-time monitoring data; The data fusion formula of multiple sensors is described as follows: S i (t), where D(t) is the fused data at time t, S i (t) is the reading of the ith sensor, w i is the weight of the ith sensor.
7. An IT equipment monitoring method, characterized in that: The following steps are involved: S1. Collecting equipment operation data: Using the monitoring module to collect various operation data of the equipment, and then uploading the equipment operation data to the cloud; S2. Design a fault prediction model; S3, model training: Use the two historical data of normal operation and failure occurrence in the cloud to train the designed fault prediction model; S4. Calculate the failure probability: Based on the failure prediction model, feed the real-time data of equipment operation into the failure prediction model for real-time failure prediction, and calculate and generate the final failure probability; S5, intelligent alarm: identify the final failure probability and send an alarm; S6, Adaptive adjustment: Based on the final failure probability, the equipment is controlled in real time to adjust the operating parameters to prevent possible future failures; S7, adjustment feedback: Identify the adaptive adjustment of equipment operating parameters and feedback the adjustment results.
8. The IT equipment monitoring method according to claim 7, characterized in that: The fault prediction model includes a hybrid model of a random forest prediction model (RandomForest), a support vector machine prediction model (SVM) and a deep neural network prediction model (DNN); The prediction formula of random forest is: Among them, P f is the predicted value of failure probability, N is the number of decision trees, h i (X) is the prediction result of the i-th tree for the input data X. The prediction formula of support vector machine is: in, f(x) is the predicted value of the failure probability, x is the input data, x i is the support vector, α i and i are the coefficients and labels of the support vectors, K(x i ,x) is the kernel function and b is the bias. The structure of a deep neural network is: -Input layer: receives feature data obtained from feature extraction; -Multiple hidden layers: Each layer uses nonlinear activation functions such as ReLU to capture complex patterns in the data; -Output layer: output the failure probability P of the device.
9. The IT equipment monitoring method according to claim 7, characterized in that: The calculation of the failure probability uses the following failure probability calculation formula to calculate the final failure probability: Formula description: Among them, P f (t) is the probability of failure, f i (t) is the prediction result of different models, w i is the model weight; The adaptive adjustment unit 3-2 of the equipment operating parameters is used to control the equipment system to adjust the equipment operating parameters in real time according to the fault prediction results to prevent possible faults from occurring in the future.
10. A storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the IT equipment monitoring method according to any one of claims 7 to 9 is implemented.
Citation Information
Cited By
WebGL rendering-based device state display system and method
CN119417457A
Multi-sensor integrated system for remote monitoring of industrial equipment
CN119511910A
Multi-dimensional monitoring alarm and intelligent water affair management system and method based on MQTT protocol
CN119854107A
