Intelligent machine room management method and system based on BIM

Abstract

The invention discloses a BIM-based intelligent machine room management method and system, and relates to the technical field of building information models and intelligent management, and the method comprises the steps: S1, obtaining real-time data in a machine room, and uploading the real-time data to a cloud server, S2, building and training a machine learning model for fault prediction and diagnosis based on historical data and the real-time data, and carrying out the fault prediction and diagnosis of the machine learning model. S3, analyzing the real-time data through a machine learning model, and recognizing a potential fault, S4, feeding back a recognition result to a BIM system, generating a maintenance work order, and notifying maintenance personnel to process the maintenance work order; according to the BIM-based intelligent machine room management method and system, the judgment standard of abnormal data can be adaptively optimized according to the actual working condition, so that the phenomena of false report and missing report are effectively reduced, the credibility of equipment operation parameters is improved, and the reliability of the equipment operation parameters is improved. The problem that in the prior art, in the fault prediction and diagnosis process, due to the fact that abnormal data are not accurately recognized through an algorithm model, misinformation or missed information is caused is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of building information modeling and intelligent management, and in particular to a method and system for managing an intelligent computer room based on BIM. Background Art

[0002] In the management and maintenance of building facilities, the computer room is a core facility, and its operating status directly affects the stability and reliability of the overall system. In order to improve the efficiency of computer room management, intelligent computer room management methods based on building information models have gradually been introduced in existing technologies. This method uses the visualization and data integration capabilities of BIM technology to build a digital model of the computer room to achieve real-time monitoring, data collection and analysis of equipment status.

[0003] Existing BIM-based intelligent computer room management methods usually combine sensor data and algorithm models for fault prediction and diagnosis. By collecting equipment operating parameters such as temperature, humidity, voltage, and current, the system can identify and alarm abnormal conditions. However, in actual applications, due to the inaccurate recognition of abnormal data by the algorithm model, the following problems may occur: false alarm problem: the system misjudges normal operating data as abnormal, generates unnecessary alarms, and increases maintenance costs and workload; missed alarm problem: the system fails to identify real abnormal data, delays the discovery and handling of faults, and may lead to serious consequences. The main reason for the above problems is that the abnormal data judgment criteria of the current algorithm model lack dynamic adaptability and it is difficult to fully consider the complexity and diversity of actual working conditions. In addition, there may be noise interference or instability of data acquisition equipment during data input, which further reduces the accuracy of fault diagnosis. Summary of the invention

[0004] The purpose of the present invention is to provide a BIM-based intelligent computer room management method and system to solve the problem of false alarm or missed alarm caused by the inaccurate recognition of abnormal data by the algorithm model in the fault prediction and diagnosis process in the prior art.

[0005] To achieve the above object, the present invention provides the following technical solution: a BIM-based intelligent computer room management method, the method comprising: S1, obtain the real-time data in the computer room and upload it to the cloud server; S2. Based on historical data and real-time data, build and train a machine learning model for fault prediction and diagnosis, including combining historical data with real-time data to generate time series data and denoising the time series. The specific formula is: ; Among them, S t represents the state parameters of the equipment in the computer room at the current moment after exponential smoothing, α represents the importance weight of the current real-time data, t represents time, and xt represents the real-time data collected from the sensors in the computer room, S t－1 Indicates the state parameters of the previous moment after smoothing; The smoothed data is fed into a machine learning model to train a model for fault prediction, which is then used to analyze new data in real time to detect potential fault trends. S3, analyze real-time data through machine learning models to identify potential failures; S4. Feedback the identification results to the BIM system, generate a maintenance work order and notify the maintenance personnel to handle it.

[0006] Preferably, the S1 comprises: Collect data from sensors in real time, including temperature, humidity, current, and voltage, smooth the data, and reduce instantaneous abnormal values ​​caused by jitter of the collection equipment. The specific formula is: ; Among them, M t Indicates the smoothed value of the current operating status of the equipment in the computer room, w i Represents the real-time data points collected from the sensors in the computer room, N represents the number of data points considered when calculating the moving average, t represents time, and i represents the index of the sampling moment.

[0007] Preferably, S3 includes: Extract abnormal features from historical fault data to form a template vector, and calculate the difference between real-time data and historical fault templates. The specific formula is: ; Among them, d represents the difference between real-time data and historical fault templates, n represents the characteristic number of equipment room operation status, x j represents the specific value of the real-time data collected from the sensor on the jth feature, y j It represents the specific value of the template value extracted from the historical fault data on the jth feature, where j represents the index of the data feature; Set the threshold T h , if the difference between the real-time data and the historical fault template is less than the threshold T h , the current data is judged as a potential fault, otherwise it is judged as a normal state.

[0008] Preferably, S4 includes: Calculate the marginal utility of maintenance tasks, sort the tasks according to the size of marginal utility, give priority to the tasks with the highest utility, generate maintenance work orders from the sorted task list, and send them to maintenance personnel through the BIM system. The specific formula for calculating the marginal utility of maintenance tasks is: ; Among them, U represents the utility value of the maintenance task, ΔD represents the improvement of the health score of the equipment operation status after performing the maintenance task, and ΔR represents the total amount of resources required to complete the maintenance task.

[0009] Preferably, the real-time data in the computer room in S1 includes the temperature, current, humidity and voltage of the equipment in the computer room.

[0010] Preferably, the importance weight α of the current real-time data in S2 ranges from 0.1 to 0.9.

[0011] Preferably, the calculation formula for the number of data points N considered when calculating the sliding average in S1 is: ; Where N is the number of data points considered when calculating the sliding average, L is the time range covered by the sliding average, and Δt is the time interval between each sampling of the sensor. Represents the ceiling operator.

[0012] Preferably, after the maintenance task is performed in S4, the calculation formula for the improvement ΔD of the health score of the equipment operation status is: ; Among them, ΔD represents the improvement of the health score of the equipment operation status after the maintenance task is performed, D after represents the health score of the equipment after completing the maintenance task, D before Indicates the health score of the equipment before the maintenance task begins.

[0013] Preferably, the utility value U of the maintenance task in S4 is calculated and adjusted, and the specific formula is: U′=U×(1+E); Among them, U′ represents the utility value of the adjusted maintenance task, U represents the utility value of the maintenance task, and E represents the task urgency factor.

[0014] A BIM-based intelligent computer room management system, used to implement the steps of the BIM-based intelligent computer room management method, the system comprising: Data acquisition module, used to obtain real-time data in the computer room and upload it to the cloud service; A model building module connected to the data acquisition module, used to build and train a machine learning model for fault prediction and diagnosis based on historical data and real-time data; A fault analysis module connected to the data acquisition module and the model building module is used to analyze real-time data through a machine learning model to identify potential faults; The task management module connected to the fault analysis module is used to feed back the identification results to the BIM system, generate a maintenance work order and notify the maintenance personnel to handle it.

[0015] It can be seen from the above technical solution that the present invention has the following beneficial effects: The BIM-based intelligent computer room management method and system obtains real-time data in the computer room and uploads it to a cloud server, builds and trains a machine learning model for fault prediction and diagnosis based on historical data and real-time data, analyzes real-time data through the machine learning model, identifies potential faults, feeds back the identification results to the BIM system, generates a maintenance work order and notifies maintenance personnel to handle it, and can adaptively optimize the judgment criteria of abnormal data according to actual working conditions, thereby effectively reducing false alarms and missed alarms, reducing the impact of sensor noise interference and data acquisition equipment instability on system diagnostic performance, improving the credibility of equipment operating parameters, quickly locating potential faults and generating maintenance tasks, significantly reducing maintenance costs and workloads, and being able to discover abnormal trends earlier, providing a reliable basis for equipment maintenance, effectively avoiding the occurrence of major faults, enhancing the stability and reliability of intelligent computer room management, and solving the problem of false alarms or missed alarms caused by the inaccurate identification of abnormal data by the algorithm model in the fault prediction and diagnosis process in the prior art. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a flow chart of the method of the present invention; Figure 2 This is a schematic diagram of the connection of the system modules of the present invention. DETAILED DESCRIPTION

[0017] 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.

[0018] like Figure 1 As shown, the present invention provides a technical solution: a BIM-based intelligent computer room management method, the method comprising: S1, obtain the real-time data in the computer room and upload it to the cloud server; S2. Based on historical data and real-time data, build and train a machine learning model for fault prediction and diagnosis, including combining historical data with real-time data to generate time series data and denoising the time series. The specific formula is: ; Among them, S t represents the state parameters of the equipment in the computer room at the current moment after exponential smoothing, α represents the importance weight of the current real-time data, t represents time, and x trepresents the real-time data collected from the sensors in the computer room, S t－1 Indicates the state parameters of the previous moment after smoothing; The smoothed data is fed into a machine learning model to train a model for fault prediction, which is then used to analyze new data in real time to detect potential fault trends. S3, analyze real-time data through machine learning models to identify potential failures; S4. Feedback the identification results to the BIM system, generate a maintenance work order and notify the maintenance personnel to handle it.

[0019] In this implementation, by combining BIM technology and machine learning algorithms, intelligent management of the operating status of the equipment in the computer room is achieved. Specifically: First, in step S1, the key data in the computer room is collected in real time through the sensor network and uploaded to the cloud server to provide a basis for subsequent data analysis; in step S2, the time series is constructed using historical data and real-time data, and the data is denoised by exponential smoothing to remove random interference, thereby generating more representative equipment status parameters. Based on these time series data, a machine learning model is trained for the prediction of fault trends; in step S3, the model uses real-time input data to analyze the operating status of the current equipment and detect possible faults; finally, through step S4, the analysis results are fed back to the BIM system to generate a maintenance work order to achieve automated maintenance scheduling. Through this series of processes, early prediction and rapid response to equipment failures are achieved, thereby effectively reducing the maintenance cost of the computer room and improving the reliability and stability of equipment operation. By combining historical data and real-time data, as well as training and optimizing machine learning models, this method can accurately predict potential failures of computer room equipment. The data analysis results are automatically fed back to the BIM system and a maintenance work order is generated, which simplifies the traditional manual maintenance process, improves maintenance efficiency, and effectively improves the accuracy of status parameters, thereby enhancing the prediction effect of the model, reducing downtime and maintenance costs caused by equipment failures, and improving resource utilization.

[0020] S1 includes real-time data collection from sensors, including temperature, humidity, current, and voltage, smoothing the data, and reducing instantaneous abnormal values ​​caused by jitter of the collection equipment. The specific formula is: ; Among them, M t Indicates the smoothed value of the current operating status of the equipment in the computer room, w i Represents the real-time data points collected from the sensors in the computer room, N represents the number of data points considered when calculating the moving average, t represents time, and i represents the index of the sampling moment.

[0021] This implementation method provides detailed expansion of the data collection and preprocessing steps: First, the environmental parameters (such as temperature and humidity), power parameters (such as current and voltage) and equipment status parameters are collected through the sensor network in the computer room. Real-time data is transmitted to the cloud server through the acquisition module of the BIM system. In order to reduce instantaneous abnormal values ​​caused by equipment jitter, environmental interference, etc., the sliding average method is used to smooth the collected data. Specifically, the data of the first N sampling points are weighted and summed in chronological order to obtain the smoothed state value M t . In this way, the influence of random noise is eliminated, making the sensor data more representative and providing high-quality input data for subsequent fault prediction. The processed data goes through a standardization process to meet the subsequent machine learning model input requirements, thereby achieving efficient docking between data and models. It can effectively eliminate instantaneous outliers in the acquisition process and improve the smoothness and reliability of the data. Through the joint acquisition and processing of multiple parameters (such as temperature, humidity, current, and voltage), it can fully reflect the operating status of the equipment in the computer room and avoid the errors that may be caused by single parameter monitoring. Data preprocessing improves the quality of input data, thereby enhancing the performance and accuracy of subsequent machine learning models in fault prediction. Combined with the acquisition results of multiple types of sensors, it can adapt to different computer room environments and meet the status monitoring needs in complex scenarios.

[0022] S3 includes extracting abnormal features from historical fault data, forming a template vector, and calculating the difference between real-time data and historical fault templates. The specific formula is: ; Among them, d represents the difference between real-time data and historical fault templates, n represents the characteristic number of equipment room operation status, x j represents the specific value of the real-time data collected from the sensor on the jth feature, y j It represents the specific value of the template value extracted from the historical fault data on the jth feature, where j represents the index of the data feature; Set the threshold T h , if the difference between the real-time data and the historical fault template is less than the threshold T h , the current data is judged as a potential fault, otherwise it is judged as a normal state.

[0023] In this implementation, the feature comparison between historical fault data and real-time data is used to achieve accurate analysis of the equipment status in the computer room: Based on historical fault data, extract the key features of the equipment when the fault occurs (such as abnormal changes in temperature, current, and voltage). Organize these features into a template vector y={y1, y2,…, yn} to identify typical fault modes. Collect real-time data x={x1, x2,…, xn} through sensors, compare it with the historical fault template feature by feature, and calculate the difference value d by the absolute value sum method. This difference value can quantitatively reflect the similarity between the current equipment state and the historical fault state. Set the threshold Th as the standard for state judgment. If the difference value d is lower than the threshold, it means that the current equipment state is close to the historical fault state and should be judged as a potential fault; otherwise, the equipment is judged to be in normal operation. The judgment result will be fed back to the BIM system to trigger further maintenance or state monitoring processes. The fault template generated based on historical data can efficiently detect potential abnormal conditions of equipment and significantly improve the accuracy of fault prediction. The difference calculation method can flexibly adapt to different types of equipment parameters and fault characteristics, making the method highly versatile in a variety of scenarios. Through the rapid comparison of real-time data and historical templates, instant judgment of the equipment operating status is achieved, reducing delays. The judgment method combining difference values ​​with thresholds can effectively reduce misjudgments caused by accidental fluctuations, enhance the reliability of fault detection, discover potential faults in a timely manner and take measures, which can avoid the expansion of equipment damage and reduce maintenance and downtime costs.

[0024] S4 includes calculating the marginal utility of maintenance tasks, sorting tasks according to the size of marginal utility, giving priority to tasks with the highest utility, generating maintenance work orders from the sorted task list, and sending them to maintenance personnel through the BIM system. The specific formula for calculating the marginal utility of maintenance tasks is: ; Among them, U represents the utility value of the maintenance task, ΔD represents the improvement of the health score of the equipment operation status after performing the maintenance task, and ΔR represents the total amount of resources required to complete the maintenance task.

[0025] This implementation method uses marginal utility analysis to prioritize maintenance tasks, thereby optimizing the maintenance efficiency of equipment in the computer room: The BIM system generates maintenance tasks based on the equipment operating status and fault diagnosis results, and calculates the marginal utility value U of each maintenance task in combination with the health scoring system. The ΔD in the calculation formula reflects the degree of improvement of the equipment operating status by the maintenance task, and ΔR represents the consumption of resources such as manpower, material resources and time required to complete the task. Sort all maintenance tasks from high to low according to the marginal utility value. The higher the utility value, the higher the priority of the task, which means that the task is more efficient in resource utilization. According to the sorting results, an optimized task list is generated, and a maintenance work order is automatically generated and sent to the corresponding maintenance personnel through the BIM system. The work order content includes the specific description of the maintenance task, priority, required resources and expected completion time. After the maintenance is completed, the system records the actual completion status, and updates the health score and resource usage of the equipment, providing real-time data support for the utility calculation of subsequent tasks. By calculating the marginal utility value, priority is given to tasks that can improve the equipment's operating status the most and consume the least resources, which greatly improves resource utilization efficiency. Task allocation sorted by utility ensures that the most important maintenance work is completed first, avoiding waste of resources and inefficient maintenance. Task completion status and equipment health scores are recorded in real time, and the order and priority of subsequent tasks are adjusted dynamically to ensure continuous optimization of maintenance strategies. Based on data analysis and scoring mechanisms, the subjectivity and irrationality that may exist in traditional maintenance scheduling are avoided, and the scientific nature and accuracy of maintenance decisions are improved.

[0026] The real-time data in the computer room in S1 includes the temperature, current, humidity and voltage of the equipment in the computer room. In this implementation, the system collects multiple key data of the equipment in the computer room in real time through the sensor network, including temperature, current, humidity and voltage. These parameters reflect the operating status and environmental conditions of the equipment in the computer room: Temperature data is used to monitor the thermal status of the equipment to prevent damage to the equipment due to overheating. Current and voltage data are used to analyze the power supply and identify possible electrical faults such as short circuits or overloads. Humidity data is used to detect whether the ambient humidity in the equipment room is within a safe range to prevent the electrical parts of the equipment from being damp or corroded due to excessive humidity.

[0027] The collected data is transmitted to the cloud server through sensors and integrated with the BIM system for analysis and processing.

[0028] The above multi-dimensional data are pre-processed (such as denoising, smoothing, etc.) to provide high-quality input for subsequent condition monitoring and fault prediction.

[0029] The data is comprehensively analyzed through machine learning models to identify equipment operating trends and potential failure risks. The diversity of data ensures that the model can more comprehensively evaluate the equipment status, thereby improving the accuracy and reliability of the analysis.

[0030] By collecting four types of data, namely temperature, current, humidity and voltage, the operating status of the equipment in the computer room and its environmental conditions can be fully monitored. The joint analysis of different parameters helps to identify potential compound faults (such as high temperature and overload at the same time), thereby improving the accuracy of fault prediction. Humidity monitoring can effectively prevent the impact of environmental conditions on equipment and ensure the long-term safe and stable operation of equipment. Real-time collection and transmission of multiple parameters realizes efficient management of data flow and provides basic support for other functions in the BIM system (such as task allocation).

[0031] The importance weight α of the current real-time data in S2 ranges from 0.1 to 0.9. In this embodiment, by setting a value range (0.1 to 0.9) for the importance weight α of the current real-time data, the relative influence weights of historical data and real-time data in the equipment status analysis are controlled: when the α value is large, the real-time data occupies a larger weight, which is suitable for scenarios where the environment or equipment status changes rapidly; when the α value is small, the historical data occupies a larger weight, which is suitable for scenarios where the status is stable. The value range of α is limited to 0.1 to 0.9 to ensure that both real-time data and historical data have a certain influence. The lower limit of the value range (0.1) avoids completely ignoring the real-time data; the upper limit (0.9) ensures that the historical data has a retention effect on the smoothing result, thereby improving the robustness of the prediction. In practical applications, the α value is dynamically adjusted according to the speed of change of the equipment status or the stability of the computer room environment. For example, when the equipment operating status changes frequently, α is increased to quickly respond to real-time data; when the equipment status is stable, α is reduced to enhance the reliance on historical trends. By setting the value range of α, it is possible to adapt to the equipment status analysis needs in different scenarios and realize flexible data weight allocation. The α value range (0.1 to 0.9) ensures a reasonable balance between real-time data and historical data in smoothing, effectively improving the accuracy and stability of equipment status parameter calculations. Adjusting the α value according to real-time operation requirements helps to achieve the best balance between rapid response and long-term trend analysis. By limiting the value range of α, we can avoid over-reliance on real-time data or historical data due to excessively high or low weight settings, thereby ensuring the reliability of equipment status analysis results.

[0032] The calculation formula for the number of data points N considered when calculating the sliding average in S1 is: ; Where N is the number of data points considered when calculating the sliding average, L is the time range covered by the sliding average, and Δt is the time interval between each sampling of the sensor. Represents the ceiling operator.

[0033] In this embodiment, the required number of data points N is dynamically determined based on the time coverage of the sliding average and the sensor sampling interval to optimize the data smoothing effect: L is a pre-set sliding average time window, which is used to define the length of historical time that needs to be covered during the data smoothing process. For example, if you want the smoothed data to reflect the state within the last minute, L can be set to 60 seconds. Δt is the time interval between each sampling of the sensor, reflecting the frequency of data collection. The higher the sampling frequency, the smaller Δt, and the more data points are collected per unit time. Through the formula , calculate the number of historical data points N required for the sliding average. The rounding operation ensures that N is an integer, so as to adapt to the actual data processing needs. The size of N determines the smoothness of the sliding average: a larger N (corresponding to a larger L or a smaller Δt) can better eliminate random noise, but may cause smoothing delays. A smaller N can better reflect real-time performance, but has a weaker ability to suppress random fluctuations. Dynamically adjust L and Δt according to different application scenarios to strike a balance between the real-time and smoothness of the data. For example, in a fault detection scenario, a shorter L and a higher sampling frequency may be required to ensure a fast response. Automatically calculate N according to the sliding average time range and sampling interval to ensure that the method adapts to the needs of different application scenarios. Reasonable setting of N can optimize the sliding average effect, suppress random noise, and maintain the sensitivity of real-time data. By calculating the minimum number of data points required, excessive data redundancy is avoided and computing resources are saved. Adapt to the needs of high-frequency sampling to ensure that the system can quickly process real-time data and improve the timeliness of fault detection and status monitoring.

[0034] After the maintenance task is executed in S4, the calculation formula for the improvement ΔD of the health score of the equipment operation status is: ; Among them, ΔD represents the improvement of the health score of the equipment operation status after the maintenance task is performed, D after represents the health score of the equipment after completing the maintenance task, D before Indicates the health score of the equipment before the maintenance task begins.

[0035] In this implementation, the improvement of the health score before and after the equipment maintenance is calculated by comparing the health score, and the improvement effect of the maintenance task on the equipment status is quantified: the health score is a comprehensive indicator used to quantify the operating status of the equipment, usually calculated based on multi-parameter data (such as temperature, current, humidity, voltage, etc.). The score range can be 0 to 100, and the higher the value, the healthier the equipment status. Before the maintenance task begins, the system calculates the health score of the current equipment based on the real-time data and historical data collected by the sensor, recorded as D before After completing the maintenance task, the system evaluates the equipment again and obtains the latest health score, which is recorded as D after By formula , calculate the actual improvement of the maintenance task on the health status of the equipment. ΔD>0 means that the maintenance task has achieved positive results; ΔD=0 means that the maintenance task has no significant improvement on the equipment status; if ΔD<0, there may be improper maintenance or other equipment failure problems. The calculation results can be used to evaluate the priority ranking of maintenance tasks (related to marginal utility calculation) and the optimization of subsequent maintenance strategies. By calculating the improvement ΔD of the health score, the improvement effect of the maintenance task on the equipment status can be intuitively quantified, providing a reliable basis for task evaluation. Dynamically adjust the priority of maintenance tasks according to the improvement in the score to improve resource utilization efficiency. If the score does not improve or even decreases after maintenance, the system can further analyze the reasons and locate potential failures or improper maintenance problems. The combined use of health scores and improvements provides data support for equipment management in the BIM system and improves the scientific nature of decision-making. The score calculation results can be used as the basis for subsequent maintenance task effect evaluation and process improvement to form a closed-loop management.

[0036] The utility value U of the maintenance task in S4 is calculated and adjusted. The specific formula is: U′=U×(1+E); Among them, U′ represents the utility value of the adjusted maintenance task, U represents the utility value of the maintenance task, and E represents the task urgency factor.

[0037] In this embodiment, the original utility value U is adjusted in combination with the urgency factor E of the maintenance task, thereby optimizing the priority ranking of the tasks: The urgency factor E is a dimensionless parameter, usually determined by the time sensitivity of the task, the severity of the fault, or the criticality of the equipment operation. The value range of E can be dynamically set according to the specific application scenario. For example, E=0 means that the task has no urgent demand; E>0 means that the task has a certain urgency. The larger the value, the higher the urgency.

[0038] After calculating the original utility value U, it is adjusted using the formula U′=U×(1+E): When E=0, U′=U means that the urgency factor has no effect on the utility value; when E>0, U′>U, which means that the emergency task obtains a higher utility value to increase its priority. The maintenance tasks are prioritized according to the adjusted utility value U′, and the maintenance work order is generated for the highest priority task and sent to the maintenance personnel through the BIM system. The introduction of the urgency factor E enables the system to dynamically adjust the priority according to the time sensitivity of the task and the severity of the fault, thereby improving the flexibility of scheduling. By adjusting the utility value U′, it is ensured that emergency tasks are executed first when resources are limited, and the configuration of maintenance resources is optimized. When a sudden equipment failure occurs or the operation of key equipment is affected, the urgency factor can quickly increase the priority of related tasks to ensure timely response. Combining utility value with urgency provides a more scientific decision-making basis for task sorting in a complex computer room environment.

[0039] A BIM-based intelligent computer room management system is also provided, which is used to implement the steps of the BIM-based intelligent computer room management method. The system includes: Data acquisition module, used to obtain real-time data in the computer room and upload it to the cloud service; A model building module connected to the data acquisition module, used to build and train a machine learning model for fault prediction and diagnosis based on historical data and real-time data; A fault analysis module connected to the data acquisition module and the model building module is used to analyze real-time data through a machine learning model to identify potential faults; The task management module connected to the fault analysis module is used to feed back the identification results to the BIM system, generate a maintenance work order and notify the maintenance personnel to handle it.

[0040] The data acquisition module collects the operating data of the computer room environment and equipment in real time through the sensor network and uploads it to the cloud service. The integrity and accuracy of the data are ensured through high-frequency sampling and preprocessing technology. The model building module uses historical data to train the fault prediction model, and dynamically updates the model in combination with real-time data. Use machine learning algorithms (such as time series analysis or deep learning) to identify trends and potential anomalies in the operating status of the equipment. The fault analysis module receives real-time data from the data acquisition module in real time, analyzes it in combination with the trained model, detects potential fault trends and provides diagnostic results. The fault results include information such as equipment failure probability, type, and possible causes, and are uploaded to the task management module. The task management module generates maintenance work orders based on the fault analysis results, giving priority to urgent or high-impact tasks. The work order displays detailed information of the fault through the BIM system, including the 3D location of the faulty equipment, maintenance steps and required resources. At the same time, the work order will be sent to the maintenance personnel through the notification unit for processing to ensure timely maintenance. The system modules have clear division of labor, and data acquisition, model building, fault analysis and task management each perform their duties to ensure functional independence and collaborative optimization. Combining real-time data with historical data, the system can quickly and intelligently detect faults and automatically generate maintenance tasks, significantly improving maintenance efficiency. By combining fault diagnosis with the BIM system, 3D visualization technology can be used to intuitively display the status of equipment in the computer room and the location of faults, making it easier for maintenance personnel to accurately locate problems. The model building module can dynamically update the fault prediction model based on real-time data, improving the system's ability to adapt to complex scenarios. The task management module ensures the rationality and efficiency of resource allocation by adjusting priority sorting and urgency factors.

[0041] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A BIM-based intelligent computer room management method, characterized in that: The method comprises: S1, obtain the real-time data in the computer room and upload it to the cloud server; S2. Based on historical data and real-time data, build and train a machine learning model for fault prediction and diagnosis, including combining historical data with real-time data to generate time series data and denoising the time series. The specific formula is: ; Among them, S t represents the state parameters of the equipment in the computer room at the current moment after exponential smoothing, α represents the importance weight of the current real-time data, t represents time, and x t represents the real-time data collected from the sensors in the computer room, S t－1 Indicates the state parameters of the previous moment after smoothing; The smoothed data is fed into a machine learning model to train a model for fault prediction, which is then used to analyze new data in real time to detect potential fault trends. S3, analyze real-time data through machine learning models to identify potential failures; S4. Feedback the identification results to the BIM system, generate a maintenance work order and notify the maintenance personnel to handle it.

2. According to the BIM-based intelligent computer room management method of claim 1, it is characterized by: The S1 includes: Collect data from sensors in real time, including temperature, humidity, current, and voltage, smooth the data, and reduce instantaneous abnormal values ​​caused by jitter of the collection equipment. The specific formula is: ; Among them, M t Indicates the smoothed value of the current operating status of the equipment in the computer room, w i Represents the real-time data points collected from the sensors in the computer room, N represents the number of data points considered when calculating the moving average, t represents time, and i represents the index of the sampling moment.

3. The BIM-based intelligent computer room management method according to claim 1, characterized in that: The S3 includes: Extract abnormal features from historical fault data to form a template vector, and calculate the difference between real-time data and historical fault templates. The specific formula is: ; Among them, d represents the difference between real-time data and historical fault templates, n represents the characteristic number of equipment room operation status, x j represents the specific value of the real-time data collected from the sensor on the jth feature, y j It represents the specific value of the template value extracted from the historical fault data on the jth feature, where j represents the index of the data feature; Set the threshold T h , if the difference between the real-time data and the historical fault template is less than the threshold T h , the current data is judged as a potential fault, otherwise it is judged as a normal state.

4. The BIM-based intelligent computer room management method according to claim 1, characterized in that: The S4 includes: Calculate the marginal utility of maintenance tasks, sort the tasks according to the size of marginal utility, give priority to the tasks with the highest utility, generate maintenance work orders from the sorted task list, and send them to maintenance personnel through the BIM system. The specific formula for calculating the marginal utility of maintenance tasks is: ; Among them, U represents the utility value of the maintenance task, ΔD represents the improvement of the health score of the equipment operation status after performing the maintenance task, and ΔR represents the total amount of resources required to complete the maintenance task.

5. The BIM-based intelligent computer room management method according to claim 1, characterized in that: The real-time data in the computer room in S1 includes the temperature, current, humidity and voltage of the equipment in the computer room.

6. The BIM-based intelligent computer room management method according to claim 1, characterized in that: The importance weight α of the current real-time data in S2 ranges from 0.1 to 0.

9.

7. The BIM-based intelligent computer room management method according to claim 2 is characterized in that: The calculation formula for the number of data points N considered when calculating the sliding average in S1 is: ; Where N is the number of data points considered when calculating the sliding average, L is the time range covered by the sliding average, and Δt is the time interval between each sampling of the sensor. Represents the ceiling operator.

8. The BIM-based intelligent computer room management method according to claim 1, characterized in that: After the maintenance task is performed in S4, the calculation formula for the improvement ΔD of the health score of the equipment operation status is: ; Among them, ΔD represents the improvement of the health score of the equipment operation status after the maintenance task is performed, D after represents the health score of the equipment after completing the maintenance task, D before Indicates the health score of the equipment before the maintenance task begins.

9. The BIM-based intelligent computer room management method according to claim 1, characterized in that: The utility value U of the maintenance task in S4 is calculated and adjusted, and the specific formula is: U′=U×(1+E); Among them, U′ represents the utility value of the adjusted maintenance task, U represents the utility value of the maintenance task, and E represents the task urgency factor.

10. A BIM-based intelligent computer room management system, used to implement the steps of the BIM-based intelligent computer room management method according to any one of claims 1 to 9, characterized in that: The system comprises: Data acquisition module, used to obtain real-time data in the computer room and upload it to the cloud service; A model building module connected to the data acquisition module, used to build and train a machine learning model for fault prediction and diagnosis based on historical data and real-time data; A fault analysis module connected to the data acquisition module and the model building module is used to analyze real-time data through a machine learning model to identify potential faults; The task management module connected to the fault analysis module is used to feed back the identification results to the BIM system, generate a maintenance work order and notify the maintenance personnel to handle it.

Images