Machine room asset operation and maintenance management system based on dynamic environment data analysis

Abstract

The present application relates to the technical field of machine room operation and maintenance management and dynamic environment monitoring data analysis, in particular to a machine room asset operation and maintenance management system based on dynamic environment data analysis, comprising: obtaining asset core temperature data, asset real-time power consumption data, micro-environment temperature data, business weight data, historical use time length data, baseline trend and scrap age of the target asset; aligning, processing and calculating temperature difference and dynamic thermal impedance index, and extracting degradation trend; comparing the baseline trend to determine the penalty coefficient, and converting the equivalent loss of the historical use time length to obtain the equivalent service age; calculating the dispatch priority score based on the equivalent service age, the scrap age, the business weight data and the degradation trend, and generating operation and maintenance dispatch instructions; obtaining an execution completion signal, comparing the dynamic thermal impedance index before and after execution, and correcting the equivalent service age or generating a work order upgrade instruction; the present application can realize asset operation and maintenance management closer to the real loss state.

Description

Technical Field

[0001] This invention relates to the field of data center operation and maintenance management and environmental monitoring data analysis technology, specifically a data center asset operation and maintenance management system based on environmental monitoring data analysis. Background Technology

[0002] With the continuous expansion of data center scale and the high-density deployment of assets such as servers and storage nodes, the correlation between the operational status of assets within the data center and business continuity is becoming increasingly close. Especially in scenarios that carry core businesses, changes in temperature, power consumption, and local microenvironment can directly affect the aging speed of assets, maintenance timing, and allocation of operation and maintenance resources. How to more accurately identify and manage the health of assets based on the dynamic environmental data during data center operation has become an important problem that needs to be solved in the field of data center asset operation and maintenance.

[0003] Traditional data center asset operation and maintenance management currently relies mainly on the following methods: conducting manual inspections according to fixed inspection cycles, triggering alarms based on static temperature thresholds, calculating lifespan based on the natural usage time in the asset ledger, and manually assigning tasks based on business importance.

[0004] However, fixed inspections, static temperature alarms, natural age statistics, and manual dispatch all have certain shortcomings. For example, fixed inspections are difficult to detect assets that are undergoing continuous and slow heat dissipation degradation in a timely manner, leading to the long-term accumulation of hidden problems; static temperature threshold alarms are mainly for instantaneous over-limit scenarios, and lack an effective identification and quantification mechanism for assets that have already shown heat dissipation performance degradation but have not reached the alarm temperature; calculating lifespan based on natural usage time is difficult to reflect the actual wear and tear differences under different operating environments and heat dissipation conditions; and manual dispatch is easily affected by the lack of a unified quantitative basis between business weight, equipment aging degree, and current degradation trend, resulting in problems such as mis-dispatch, missed dispatch, and insufficient operation and maintenance closed-loop verification. Summary of the Invention

[0005] The purpose of this invention is to provide a data center asset operation and maintenance management system based on environmental data analysis, and to solve the following technical problems:

[0006] By integrating data from the data center's environmental conditions, asset lifespan data, and business impact data into a single scheduling logic, the hidden losses caused by long-term sub-optimal heat dissipation environments are quantified and incorporated into the asset lifespan model. This avoids missed or over-inspections caused by relying solely on static temperature thresholds or fixed inspection cycles, thereby achieving asset operation and maintenance management that more closely reflects the actual wear and tear of equipment and more accurate allocation of operation and maintenance tasks.

[0007] The objective of this invention can be achieved through the following technical solutions:

[0008] The data center asset operation and maintenance management system based on environmental data analysis includes:

[0009] The data acquisition module is used to acquire the target asset's core temperature data, real-time power consumption data, micro-environment temperature data, business weight data, historical usage duration data, as well as the system's preset baseline trend and preset scrap age.

[0010] The feature calculation module communicates with the data acquisition module and is used to align the core temperature data of the asset, the micro-environment temperature data and the real-time power consumption data of the asset and obtain the dynamic thermal resistance index. It also extracts the degradation trend of the dynamic thermal resistance index based on a preset time window.

[0011] The age conversion module, connected to the feature calculation module, is used to compare the degradation trend with the baseline trend, determine the penalty coefficient based on the comparison result, and perform equivalent loss conversion on the historical usage time data based on the penalty coefficient to obtain the equivalent service age of the target asset.

[0012] The scheduling generation module is connected to the age conversion module and the feature calculation module respectively. It is used to calculate the scheduling priority score based on the equivalent service age, scrapping age, business weight data and deterioration trend, and generate operation and maintenance scheduling instructions based on the scheduling priority score.

[0013] The feedback evaluation module, connected to the scheduling generation module and the age conversion module, is used to obtain the execution completion signal of the corresponding operation and maintenance scheduling instruction, and to compare the historical dynamic thermal impedance index before execution with the current dynamic thermal impedance index after execution. Based on the comparison results, the equivalent service age is corrected, or a work order upgrade instruction is generated.

[0014] Furthermore, methods for obtaining core asset temperature data, real-time asset power consumption data, and micro-environment temperature data of the target asset include:

[0015] According to the preset sampling frequency, the initial core temperature data and initial power consumption data are read from the underlying management interface of the target asset.

[0016] Simultaneously read initial ambient temperature data from the environmental sensors of the corresponding target asset;

[0017] A sliding window is constructed with a preset time length, and the initial core temperature data, initial power consumption data and initial ambient temperature data are timestamped within the sliding window.

[0018] Remove abrupt changes in data that exceed the preset upper and lower limits within the sliding window, and obtain core asset temperature data, real-time asset power consumption data, and microenvironment temperature data.

[0019] Furthermore, methods for extracting the degradation trend of the dynamic thermal resistance index based on a preset time window include:

[0020] Extract all dynamic thermal resistance indices within a preset time window and construct a dynamic thermal resistance index sequence.

[0021] Trend fitting is performed on the dynamic thermal resistance index sequence to obtain the linear slope corresponding to the dynamic thermal resistance index sequence, and the linear slope is used as the deterioration trend.

[0022] Furthermore, methods for obtaining the equivalent service life of the target asset by performing equivalent depreciation conversion on historical usage time data based on penalty coefficients include:

[0023] Determine whether the degradation trend is greater than the baseline trend;

[0024] If the degradation trend is less than or equal to the baseline trend, the runtime corresponding to the preset time window will be directly added to the historical usage duration data to obtain the equivalent service life.

[0025] If the degradation trend is greater than the baseline trend, then assess the degree to which the degradation trend exceeds the baseline trend.

[0026] By querying the preset mapping table between the degree of exceeding the standard and the penalty coefficient, the corresponding penalty coefficient is obtained, and the runtime corresponding to the preset time window is converted based on the penalty coefficient to obtain the acceleration loss time.

[0027] Accelerated wear and tear duration is added to historical usage duration data to obtain equivalent service life.

[0028] Furthermore, the method for calculating scheduling priority scores and generating operation and maintenance scheduling instructions based on the scheduling priority scores includes:

[0029] Age approximation rate is obtained based on equivalent service life and scrapping age;

[0030] By integrating age-approach rate, business weight data, and degradation trend, a scheduling priority score is calculated and obtained.

[0031] Obtain the preset scheduling threshold and compare the scheduling priority score with the scheduling threshold;

[0032] If the scheduling priority score is less than the scheduling threshold, no instruction will be generated.

[0033] If the scheduling priority score is greater than or equal to the scheduling threshold, an operation and maintenance scheduling instruction containing the target asset identifier and maintenance recommendations is generated.

[0034] Furthermore, methods for obtaining and comparing the historical dynamic thermal resistance index before execution with the current dynamic thermal resistance index after execution, and then adjusting the equivalent service life based on the comparison results, or generating work order upgrade instructions, include:

[0035] In response to the execution completion signal, the dynamic thermal resistance index of the target asset is recalculated within a preset time period after execution and used as the current dynamic thermal resistance index.

[0036] Extract the dynamic thermal impedance index when generating operation and maintenance scheduling instructions, and use it as the historical dynamic thermal impedance index.

[0037] Determine whether the current dynamic thermal resistance index has fallen back to the preset reference thermal resistance range;

[0038] If the current dynamic thermal resistance index falls back to the reference thermal resistance range, the operation and maintenance is deemed effective, and the preset penalty compensation value is deducted from the equivalent service life.

[0039] If the current dynamic thermal resistance index does not fall back to the reference thermal resistance range, the operation and maintenance is deemed invalid, a work order escalation instruction is generated, and the work order escalation instruction is transferred to the preset expert processing queue.

[0040] Furthermore, it also includes an abnormal circuit breaker module, the methods executed by which the abnormal circuit breaker module includes:

[0041] Calculate the rate of temperature change of the microenvironment temperature data and continuously monitor the rate of temperature change;

[0042] Determine whether the rate of temperature change exceeds the preset global fault threshold;

[0043] If the temperature change rate is less than or equal to the global fault threshold, the data processing flow of the age conversion module and the scheduling generation module will not be interfered with.

[0044] If the rate of temperature change exceeds the global fault threshold, the circuit breaker mechanism will be triggered, the calculation of the equivalent service life will be suspended, and a global physical alarm command will be output.

[0045] Once the microenvironment temperature data recovers to the preset stable range and remains within the preset buffer time, the circuit breaker mechanism is deactivated.

[0046] Furthermore, it also includes an asset audit module, the methods of which include:

[0047] Obtain real-time power consumption data and dynamic thermal resistance index of the target asset, and obtain the business traffic data of the target asset from the business gateway;

[0048] Determine whether the real-time power consumption data of the asset is greater than or equal to the preset energy consumption threshold;

[0049] Determine whether the dynamic thermal resistance index is less than or equal to the preset thermal resistance threshold.

[0050] Determine whether the business traffic data is less than the preset traffic threshold;

[0051] If the real-time power consumption data of the asset is greater than the energy consumption threshold, the dynamic thermal resistance index is less than the thermal resistance threshold, and the business traffic data is less than the traffic threshold, then the target asset will be marked as an inefficient idle asset and an asset audit report will be generated; otherwise, the current marking status of the target asset will be maintained.

[0052] Compared with existing technologies, this invention acquires core asset temperature data, real-time asset power consumption data, micro-environment temperature data, business weight data, historical usage duration data, baseline trends, and scrap age, and performs timestamp alignment, sliding window cleaning, and mutation removal on multi-source data. This enables the transformation of originally asynchronous and glitch-containing dynamic environmental data into stable analytical input, reducing the interference of collection misalignment and instantaneous anomalies on the judgment results.

[0053] In the state characterization stage, this invention calculates the temperature difference between the core temperature and the microenvironment temperature and constructs a dynamic thermal resistance index by combining it with real-time power consumption. Then, based on a preset time window, the least squares method is used to extract the degradation trend. This can transform single-point temperature monitoring into the identification of the long-term direction and rate of change of heat dissipation performance, thereby improving the ability to discover assets with slow heat dissipation degradation. Attached Figure Description

[0054] Other features, objects, and advantages of the invention will become more apparent from the following detailed description of non-limiting embodiments with reference to the accompanying drawings:

[0055] Figure 1 This is a schematic diagram of the modules of the data center asset operation and maintenance management system based on environmental data analysis provided in the embodiments of this application. Detailed Implementation

[0056] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0057] Please see Figure 1 A data center asset operation and maintenance management system based on environmental data analysis includes:

[0058] The data acquisition module is used to acquire the target asset's core temperature data, real-time power consumption data, micro-environment temperature data, business weight data, historical usage duration data, as well as the system's preset baseline trend and preset scrap age.

[0059] The feature calculation module communicates with the data acquisition module and is used to align the core temperature data of the asset, the micro-environment temperature data and the real-time power consumption data of the asset and obtain the dynamic thermal resistance index. It also extracts the degradation trend of the dynamic thermal resistance index based on a preset time window.

[0060] The age conversion module, connected to the feature calculation module, is used to compare the degradation trend with the baseline trend, determine the penalty coefficient based on the comparison result, and perform equivalent loss conversion on the historical usage time data based on the penalty coefficient to obtain the equivalent service age of the target asset.

[0061] The scheduling generation module is connected to the age conversion module and the feature calculation module respectively. It is used to calculate the scheduling priority score based on the equivalent service age, scrapping age, business weight data and deterioration trend, and generate operation and maintenance scheduling instructions based on the scheduling priority score.

[0062] The feedback evaluation module, connected to the scheduling generation module and the age conversion module, is used to obtain the execution completion signal of the corresponding operation and maintenance scheduling instruction, and to compare the historical dynamic thermal impedance index before execution with the current dynamic thermal impedance index after execution. Based on the comparison results, the equivalent service age is corrected, or a work order upgrade instruction is generated.

[0063] This embodiment provides a data center asset operation and maintenance management mechanism based on environmental data analysis. Specifically, the mechanism is deployed in the asset management platform of the operator's core data center. The platform is connected to the underlying management interface of the server, the cabinet environment sensor, the business management system and the work order system, which are used to transform environmental data such as temperature and power consumption, which were originally only used for alarms, into the basis for asset depreciation and operation and maintenance scheduling.

[0064] Specifically, the system uses a data acquisition module to collect multi-source data of the target asset; the target asset can be a rack server or a storage node with an independent heat dissipation module.

[0065] Taking the server in rack A03-12U as an example, the system periodically obtains the core temperature data of the target asset, the real-time power consumption data of the asset, and the micro-environmental temperature data at the air inlet of the same rack. At the same time, it also obtains the business weight data of the target asset in the business system. For example, the core database server is assigned a value of 5, and the ordinary log node is assigned a value of 1.

[0066] Combined with historical usage data in the asset ledger, such as 38 months of operation, and the system's preset baseline trend and default age, such as a baseline trend of 0.0008 and a default age of 72 months;

[0067] The preset baseline trend and the preset scrap age are obtained by calibrating the historical operating sample dataset of the target asset or the equipment's factory technical specifications. The feature calculation module aligns the multi-source data that are not completely synchronized in time and constructs a dynamic thermal resistance index accordingly.

[0068] Specifically, if at a certain moment the core temperature of the asset is 65℃, the microenvironment temperature is 25℃, and the real-time power consumption of the asset is 200W, then the temperature difference at that moment is 40℃, and the dynamic thermal resistance index can be denoted as: That is, 0.20; if the core temperature of the asset is 67℃, the micro-environment temperature is 26℃, and the real-time power consumption of the asset is 205W, then we get approximately 0.20.

[0069] The system does not use data from a single moment. Instead, it generates a dynamic thermal resistance index sample every hour according to a preset time window, such as the last 7 days. The system then further extracts the degradation trend of the sequence to characterize whether the heat dissipation status continues to deteriorate.

[0070] After obtaining the deterioration trend, the age conversion module no longer simply accumulates the asset's useful life by natural months, but determines the penalty coefficient based on the degree of deviation of the deterioration trend from the baseline trend, thereby calculating the equivalent service age.

[0071] This can be understood as follows: after running for 1 day, equipment with a dynamic thermal resistance index within the preset benchmark range is equivalent to having its lifespan increased by 1 day, while equipment with a heat dissipation degradation trend greater than the preset threshold may have its lifespan increased by 1.3 days or 1.5 days. This allows the hidden physical losses caused by long-term heat dissipation-limited environments to be incorporated into the asset lifespan model.

[0072] The scheduling generation module further calculates the scheduling priority score based on the equivalent service life, the scrapping age, the business weight data, and the deterioration trend.

[0073] To facilitate understanding, a simplified example can be used: If a server's equivalent service life is 48 months and its obsolescence age is 72 months, then the age approximation rate is... ,Right now The business weight data is 5.

[0074] If the absolute value of the degradation trend is 0.015, then the scheduling priority score can be: , received ;

[0075] Then compare it with the preset scheduling threshold. If the threshold is reached or exceeded, an operation and maintenance scheduling instruction is generated. The instruction must include at least the target asset identifier, priority, and maintenance suggestions, such as performing targeted surface dust removal maintenance and checking for abnormal fan speed.

[0076] The feedback evaluation module forms a closed loop after the work order is completed; that is, after the maintenance personnel complete the processing and send back the execution completion signal, the system will re-collect the current dynamic thermal resistance index of the target asset and compare it with the historical dynamic thermal resistance index recorded before the work order was generated.

[0077] If the comparison results show that the current dynamic thermal resistance index has fallen back to the preset reference thermal resistance range, it indicates that the heat dissipation channel has been improved and the equivalent service life can be corrected.

[0078] If the current dynamic thermal resistance index does not drop, it means that the current processing has not addressed the root cause of the fault. At this time, a work order escalation instruction is generated and the case is transferred to a higher-level expert processing queue.

[0079] As an anomaly handling mechanism, if real-time power consumption data of an asset is missing within a certain sampling period, the system does not directly calculate the dynamic thermal impedance index at that moment, but marks that moment as an invalid sample and waits for the next sampling point to complete it.

[0080] If the number of consecutive missing points reaches a preset number, such as three consecutive missing sampling points, the current round of degradation trend update for the asset will be suspended to avoid misjudgment due to an empty denominator or excessive error.

[0081] For example, if business weight data is temporarily unavailable, the default weight of the business group where the target asset is located can be used as a temporary value, and then written back and corrected after the business system is restored; for another example, if the completion signal is delayed, the system can remain in the pending evaluation state and not trigger the logic to correct the equivalent service life.

[0082] In the core computer room of a certain provincial operator, the database server numbered 109 has been running continuously for about 41 months.

[0083] The system detected that the server's dynamic thermal resistance index slowly increased from 0.18 to 0.24 in the past 7 days. Although the ambient temperature of the computer room remained at around 25°C and did not trigger the traditional high temperature alarm, the equivalent service life has increased significantly faster than the natural time.

[0084] Based on this, the platform generates a targeted maintenance work order in advance for surface dust removal and heat dissipation module inspection; within 24 hours after the maintenance is completed, the server's current dynamic thermal resistance index drops back to 0.19, the system determines it as effective maintenance, and makes a compensatory deduction on the equivalent service life.

[0085] The purpose of this step is to integrate data center environmental data, asset lifespan data, and business impact data into the same scheduling logic, avoiding missed or over-inspections caused by relying solely on static temperature thresholds or fixed inspection cycles, thereby achieving asset operation and maintenance management that is closer to the actual wear and tear.

[0086] In a preferred embodiment of the present invention, the method for obtaining the core temperature data, real-time power consumption data, and microenvironment temperature data of the target asset includes:

[0087] According to the preset sampling frequency, the initial core temperature data and initial power consumption data are read from the underlying management interface of the target asset.

[0088] Simultaneously read initial ambient temperature data from the environmental sensors of the corresponding target asset;

[0089] A sliding window is constructed with a preset time length, and the initial core temperature data, initial power consumption data and initial ambient temperature data are timestamped within the sliding window.

[0090] Remove abrupt changes in data that exceed the preset upper and lower limits within the sliding window, and obtain core asset temperature data, real-time asset power consumption data, and microenvironment temperature data;

[0091] This embodiment provides a data acquisition and cleaning mechanism. Specifically, in the aforementioned data center asset management scenario, if data from different sources is directly sent into subsequent calculations as is, the following technical problems will arise: the sampling period of the data reported by the server's underlying interface and the environmental sensors is not consistent, and instantaneous abnormal values ​​will be generated by device restarts, sensor signal jumps, etc. If the data is not cleaned first, the short-term abnormality will be mistakenly regarded as heat dissipation degradation.

[0092] Therefore, this embodiment introduces time alignment and mutation removal processing during the acquisition phase; the system can perform sampling at a preset frequency, for example, every 5... Once, the initial core temperature data and initial power consumption data are read from the server's underlying management interface;

[0093] At the same time, the system reads the initial ambient temperature data from the environmental sensor at the front of the server rack. Considering that the timestamps of different devices may have a second-level deviation, the system constructs a sliding window with a preset time length, such as 1 hour, and maps the three types of raw data onto a unified time axis within the window.

[0094] To illustrate this more clearly, the alignment process can be illustrated with the following specific example: Suppose that within a certain 1-hour window, the core temperature is reported at 10:00, 10:05, and 10:10.

[0095] The power consumption reporting time is 10:01, 10:06, and 10:11; the ambient temperature reporting time is 09:59, 10:04, and 10:09; the system can be divided into three alignment points of 10:00, 10:05, and 10:10 according to the nearest neighbor principle or by a fixed step size, and the system will match the closest power consumption record and the closest ambient temperature record for each alignment point.

[0096] For example, the alignment point of 10:05 can be selected with a core temperature of 66℃, a power consumption of 208W, and an ambient temperature of 26℃; this avoids the amplification of errors caused by the time misalignment of the three types of data.

[0097] After alignment is completed, the system then performs mutation removal on the samples in the window. The mutations here are not normal fluctuations with amplitudes less than the preset threshold caused by changes in normal business load, but data that clearly do not conform to the physical laws of operation, such as the core temperature suddenly jumping from 66°C to 5°C, the power consumption dropping from 210W to 0W instantly without the asset being shut down, and the ambient temperature jumping from 25°C to 48°C within 1 minute and then quickly dropping back.

[0098] The system presets upper and lower limit ranges for values ​​and a threshold for the rate of change. If a sample exceeds the limit, it is marked as invalid and will not participate in the subsequent calculation of the dynamic thermal impedance index. For the eliminated position, the system can retain the previous valid value, use the window mean to fill the gap, or directly retain the gap and skip it in the subsequent calculation. The specific settings can be set according to the deployment requirements.

[0099] As an exception handling mechanism, if the number of valid samples in a certain window is insufficient, for example, 1 hour, and there is only 1 valid sample left in the window, while the system requires at least 3 valid samples to generate a stable value, then the window is marked as insufficient samples, and the data is only cached and not output to the downstream.

[0100] If the underlying interface and environmental sensors lose connection at the same time, the asset will be marked as having an abnormal acquisition and the acquisition link will be checked instead of continuing to use expired data. If the mutation samples are too concentrated, for example, more than half of the samples in the same window are removed, the system can determine that the quality of the window is unreliable and mark the data in the sliding time window as invalid samples to prevent a small number of residual samples from misleading subsequent trend judgments.

[0101] During the continuous monitoring of the aforementioned server 109, the server experienced a brief restart during a window from 10:00 to 11:00 on a certain day; at the moment of restart, the underlying interface reported a core temperature of 7°C and a power consumption of 3W, while the ambient temperature remained at 25°C.

[0102] The system identified that the data was significantly outside the normal operating range of the server, so it removed it from the window and only retained the data from the stable phase before and after the restart to form the core temperature data of the asset, the real-time power consumption data of the asset, and the micro-environment temperature data.

[0103] The purpose of this step is to provide input data with the same time base and quality level for subsequent thermal impedance calculations, thereby reducing errors introduced by asynchronous acquisition and high-frequency discrete interference data, and thus achieving a more stable and reproducible characterization of the device's heat dissipation status.

[0104] In a preferred embodiment of the present invention, the method for extracting the degradation trend of the dynamic thermal resistance index based on a preset time window includes:

[0105] Extract all dynamic thermal resistance indices within a preset time window and construct a dynamic thermal resistance index sequence.

[0106] Trend fitting is performed on the dynamic thermal resistance index sequence to obtain the linear slope corresponding to the dynamic thermal resistance index sequence, and the linear slope is used as the deterioration trend.

[0107] This embodiment provides a degradation trend extraction mechanism. Specifically, after data alignment and cleaning are completed, although the system can obtain the dynamic thermal resistance index at each time point, if the judgment is made based on a single point value, another defect will still be exposed: a single business peak, backup task, or short-term load switching may cause the index to rise temporarily. If such transient fluctuations are directly regarded as degradation, it will bring too many false alarms.

[0108] Therefore, this embodiment introduces a time window-based sequence slope calculation method, which does not directly look at the absolute value at a certain moment, but focuses on the direction and rate of change over a period of time;

[0109] The system extracts all dynamic thermal impedance indices within a preset time window, forms a sequence arranged chronologically, and then uses the least squares method to calculate the linear slope of the sequence, taking this slope as the degradation trend.

[0110] For ease of understanding, the following simplified example can be used; assume that the dynamic thermal impedance indices obtained for the last 5 sampling days are 0.18, 0.19, 0.21, 0.22, and 0.24, respectively;

[0111] If the time sequence is recorded as 1, 2, 3, 4, 5, then the sequence as a whole shows an upward trend; the system fits a straight line using the least squares method to obtain a positive slope, for example, about 0.015. The larger this value is, the faster the thermal resistance increases, that is, the more obvious the degradation of the heat dissipation channel is.

[0112] Conversely, if the sequence is 0.24, 0.22, 0.21, 0.20, 0.19, the slope is negative, indicating that the overall heat dissipation has improved after maintenance or load changes.

[0113] The reason for using a linear slope instead of a simple first-to-last difference is that the first-to-last difference is easily affected by the random bias of the first and last samples, while the least squares method uses all samples within the entire window and can more stably reflect the long-term direction of change.

[0114] For example, in the sequence 0.18, 0.23, 0.19, 0.24, 0.20, the difference between the first and last values ​​is less than the preset fluctuation threshold, but the fluctuation amplitude of the middle data is greater than the preset jump threshold. By using the fitting slope, it is possible to more objectively distinguish between an overall upward trend and a signal jump without a clear direction.

[0115] As an exception handling mechanism, if the number of samples in the window is less than the preset minimum value, such as less than 5 samples, the system will not output the slope, but will only retain the trend undetermined mark.

[0116] If the sample is sufficient but the time interval is not continuous, such as a large segment of data is missing in the middle, the actual timestamp can be used for fitting to avoid compressing long periods of missing data into short-term changes.

[0117] If the absolute value of the fitting result is less than the preset minimum parameter, such as less than a noise tolerance threshold, the system can regard it as approximately stable and will not enter the accelerated depreciation process.

[0118] During seven consecutive days of monitoring of server 109, the system calculated an average dynamic thermal resistance index every hour, resulting in a total of 168 samples.

[0119] The slope after fitting remained positive and was significantly higher than the baseline trend of this server model under historical health conditions. Although the air supply temperature in the data center did not exceed the traditional alarm line during these 7 days, the heat dissipation degradation of this server has been quantified as a clear upward trend line.

[0120] The purpose of this step is to transform scattered instantaneous thermal states into comparable long-term degradation trends, reduce the interference of occasional load fluctuations on scheduling results, and thus achieve a state assessment that is closer to the actual aging process of the equipment.

[0121] In a preferred embodiment of the present invention, a method for obtaining the equivalent service life of a target asset by performing equivalent depreciation conversion on historical usage duration data based on a penalty coefficient includes:

[0122] Determine whether the degradation trend is greater than the baseline trend;

[0123] If the degradation trend is less than or equal to the baseline trend, the runtime corresponding to the preset time window will be directly added to the historical usage duration data to obtain the equivalent service life.

[0124] If the degradation trend is greater than the baseline trend, then assess the degree to which the degradation trend exceeds the baseline trend.

[0125] By querying the preset mapping table between the degree of exceeding the standard and the penalty coefficient, the corresponding penalty coefficient is obtained, and the runtime corresponding to the preset time window is converted based on the penalty coefficient to obtain the acceleration loss time.

[0126] Accelerated wear and tear duration is added to historical usage duration data to obtain equivalent service life.

[0127] This embodiment provides an equivalent service life conversion mechanism; specifically, if the natural time accumulation method in the asset ledger is still used after the deterioration trend has been obtained, new defects will be exposed: both servers with good heat dissipation and servers that have been in a state of heat dissipation degradation for a long time are counted as having the same age after running for 1 month, which does not conform to the actual wear and tear.

[0128] Therefore, this embodiment compares the degradation trend with the baseline trend and introduces a penalty coefficient based on the degree of exceeding the standard, converting the differences in environment and heat dissipation status into age differences.

[0129] Specifically, if the degradation trend corresponding to a certain time window is less than or equal to the baseline trend, it indicates that the device is operating within the normal healthy degradation range, and the duration of this window is directly added to the historical usage time.

[0130] If the degradation trend is greater than the baseline trend, it indicates that the device is aging faster within this window. The excess multiple should be calculated first, then the penalty coefficient should be obtained from the table, and the duration of accelerated wear should be calculated accordingly.

[0131] This can be illustrated with a simplified data example; assuming a server has a historical usage duration of 38 months, the current statistical window is 1 month, the baseline trend is 0.010, and the actual measured degradation trend in this round is 0.016, then the exceedance multiple can be calculated as follows: Calculation, obtained ;

[0132] After querying the mapping table, if the excess multiple is between 0.5 and 1.0, the corresponding penalty coefficient is 1.4. Therefore, although the window ran naturally for one month, the acceleration loss time was recorded as [missing information]. ,Right now Months; after accumulation, the equivalent service life of the asset becomes Months, not Months;

[0133] Here's another comparison: If the degradation trend of another server in this round is 0.009, which does not exceed the baseline trend of 0.010, it indicates that no additional wear and tear has occurred. This window runs for one month, increasing only by one month, and the equivalent service life is... Months become Months;

[0134] Therefore, it can be seen that the mechanism does not accelerate depreciation indiscriminately, but rather adjusts the age of the equipment conditionally based on its current condition.

[0135] The mapping table can be implemented using discrete intervals; for example, if the excess multiple is less than or equal to 0.2, the corresponding penalty coefficient is 1.1; if it is greater than 0.2 and less than or equal to 0.5, the corresponding penalty coefficient is 1.2.

[0136] Values ​​greater than 0.5 and less than or equal to 1.0 correspond to 1.4; values ​​greater than 1.0 correspond to 1.8. A lookup table is used instead of a continuous complex function, which facilitates engineering deployment and allows different data centers to be adjusted based on their operational experience.

[0137] As an anomaly handling mechanism, if the baseline trend is configured to 0 or close to 0, direct division will cause an anomaly. In this case, the system can enable the minimum baseline protection value, for example, limit the lower limit of the baseline to 0.001, and then participate in the calculation of the excess multiple.

[0138] If an asset is a newly launched device and its historical usage time is less than a complete statistical period, an initial equivalent service life can be established based on the actual number of days it has been running. Once the accumulated sample is sufficient, the complete conversion process can be initiated.

[0139] If the corresponding interval cannot be found in the mapping table, the penalty coefficient of the nearest interval can be used, or the default coefficient of 1.0 can be rolled back to avoid interrupting the entire process.

[0140] After the thermal impedance of server 109 increased for 7 consecutive days, the system used these 7 days as a statistical window for calculation. Since its degradation trend was higher than the healthy baseline of the same model, the system did not simply add it up as a natural 7 days, but instead converted it into an accelerated wear-out period of 9.8 days through a mapping table.

[0141] In this way, the server's equivalent service life will approach the maintenance and retirement threshold earlier in the asset life record, thus triggering intervention in advance.

[0142] The purpose of this step is to quantify the unrecorded physical losses caused by prolonged abnormal heat dissipation, so as to avoid the discrepancy between the registered service life and the actual equipment condition, thereby achieving a service life assessment that is more consistent with the actual wear and tear of the equipment.

[0143] In a preferred embodiment of the present invention, the method for calculating a scheduling priority score and generating an operation and maintenance scheduling instruction based on the scheduling priority score includes:

[0144] Age approximation rate is obtained based on equivalent service life and scrapping age;

[0145] By integrating age-approach rate, business weight data, and degradation trend, a scheduling priority score is calculated and obtained.

[0146] Obtain the preset scheduling threshold and compare the scheduling priority score with the scheduling threshold;

[0147] If the scheduling priority score is less than the scheduling threshold, no instruction will be generated.

[0148] If the scheduling priority score is greater than or equal to the scheduling threshold, an operation and maintenance scheduling instruction containing the target asset identifier and maintenance recommendations is generated.

[0149] This embodiment provides an operation and maintenance scheduling generation mechanism; specifically, after obtaining the equivalent service life, if work orders are dispatched only based on the age, a defect in actual operation and maintenance will be exposed: some assets that are close to being scrapped but carry low-priority services may over-consume the preset operation and maintenance hardware resources.

[0150] Conversely, some critical assets that are deteriorating rapidly may be delayed if their equivalent service life has not reached the preset scrapping threshold.

[0151] Therefore, in this embodiment, the age approximation rate, business weight data, and deterioration trend are combined to form a scheduling priority score, and whether the scheduling priority score reaches the preset scheduling threshold is used as the basis for order dispatch.

[0152] Specifically, the equivalent service life is first divided by the scrapping age to obtain the age approximation rate; then the absolute value of the degradation trend is extracted and multiplied with the business weight data to form a unified scheduling priority score.

[0153] For ease of understanding, a simplified example can be provided; assuming that the equivalent service life of asset A is 54 months and the scrapping age is 72 months, then the age approximation rate is 0.75; the business weight data is 5;

[0154] The absolute value of the degradation trend is 0.018; therefore, the scheduling priority score is... That is, 0.0675; if the preset scheduling threshold is 0.05, then because the scheduling priority score is greater than the scheduling threshold, an operation and maintenance scheduling instruction containing the target asset identifier and maintenance suggestions is generated for asset A.

[0155] Looking at asset B, its equivalent service life is also 54 months, but its business weight data is only 1, and its absolute value of the degradation trend is 0.010. Therefore, its scheduling priority score is 0.0075. At this time, the scheduling priority score is less than the scheduling threshold, and the system does not generate instructions. Thus, the system realizes the differentiated allocation of limited operation and maintenance resources.

[0156] The absolute value of the degradation trend is introduced here because engineering requires a unified approach to handling the magnitude information of positive deterioration and negative improvement.

[0157] If the trend value of an asset is negative, it usually means that the heat dissipation is improving. Although its absolute value can be used in calculations, an additional directional threshold can be added in actual configuration. For example, instructions can only be generated when the degradation trend is positive and the scheduling priority score is greater than or equal to the scheduling threshold. This can prevent equipment that is improving from being mistakenly assigned orders due to large amplitude.

[0158] Maintenance recommendations can be generated using rule matching; for example, if the dynamic thermal resistance index rises over a long period but the asset's real-time power consumption data remains stable, the maintenance recommendation could be to prioritize surface dust removal and duct inspection.

[0159] If the real-time power consumption data of the asset rises abnormally, the maintenance recommendation is to check the fans, heat dissipation modules and abnormal load processes; if the asset has been generated with maintenance scheduling instructions multiple times and the effect is not good, the maintenance recommendation is to arrange for expert review or replacement of spare parts.

[0160] As an exception handling mechanism, if business weight data is missing, the default level in the business system can be used, such as level 3; if the obsolescence age is not configured, it can be rolled back to the common obsolescence age of the target asset category, such as 72 months for standard servers and 84 months for storage devices.

[0161] If the scheduling priority score falls right around the scheduling threshold, a hysteresis interval can be set. For example, if the scheduling threshold is 0.05, if the score is between 0.048 and 0.052, instructions will not be generated immediately, but will wait for confirmation in the next window to prevent frequent triggering and cancellation.

[0162] If there are already incomplete work orders for the same target asset within a short period of time, new operation and maintenance scheduling instructions can be processed together instead of creating multiple similar instructions repeatedly.

[0163] In the same core computer room, the target asset numbered server 109 carries the core database. Its business weight data is 5. Although it has not yet triggered a high temperature alarm, because its equivalent service life is approaching its scrap age and the deterioration trend continues to rise, the scheduling priority score calculated by the system is greater than the scheduling threshold. Therefore, the system generates an operation and maintenance scheduling instruction to the work order system, which includes the target asset identifier, high priority, and suggestions for surface dust removal and maintenance and checking the heat dissipation module.

[0164] In contrast, a regular server hosting the test environment also showed slight degradation, but its scheduling priority score was less than the scheduling threshold, so it did not generate any instructions, only kept it under observation, and did not immediately occupy on-site maintenance resources.

[0165] The purpose of this step is to incorporate asset lifespan risk, business impact, and current rate of degradation into the dispatch decision, thereby avoiding mis-dispatch or missed dispatch driven by a single indicator and achieving more accurate allocation of operation and maintenance tasks.

[0166] In a preferred embodiment of the present invention, the method of obtaining the historical dynamic thermal impedance index before execution and the current dynamic thermal impedance index after execution and comparing them, and correcting the equivalent service life based on the comparison results, or generating a work order upgrade instruction includes: in response to the execution completion signal, recalculating the dynamic thermal impedance index of the target asset within a preset time period after execution, and using it as the current dynamic thermal impedance index.

[0167] Extract the dynamic thermal impedance index when the operation and maintenance scheduling command is generated, and use it as the historical dynamic thermal impedance index; determine whether the current dynamic thermal impedance index has fallen back to the preset reference thermal impedance range.

[0168] If the current dynamic thermal resistance index falls back to the reference thermal resistance range, the operation and maintenance is deemed effective, and the preset penalty compensation value is deducted from the equivalent service life.

[0169] If the current dynamic thermal resistance index does not fall back to the reference thermal resistance range, the operation and maintenance is deemed invalid, a work order escalation instruction is generated, and the work order escalation instruction is transferred to the preset expert processing queue.

[0170] This embodiment provides a feedback evaluation and work order escalation mechanism. Specifically, after dispatching a work order and completing on-site maintenance, if the system does not quantitatively verify the maintenance results, another defect will be exposed: although the work order shows that it has been completed, the on-site handling may have only performed external physical cleaning and did not actually improve the heat dissipation path. If the equipment is still regarded as restored, it will cause the system to reduce the monitoring frequency or stop the monitoring work.

[0171] Therefore, this embodiment introduces a before-and-after comparison evaluation after receiving the execution completion signal; the system recalculates the current dynamic thermal resistance index of the asset within a preset time period, and retrieves the historical dynamic thermal resistance index recorded when the work order was generated, and determines whether the index has fallen back to the preset reference thermal resistance range after maintenance.

[0172] If the service life drops, the maintenance is deemed effective, and the equivalent service life is adjusted accordingly; if the service life does not drop, the work order is automatically escalated and the task is assigned to an expert queue.

[0173] To facilitate understanding, a specific processing example is provided below; assuming that the dynamic thermal resistance index of server 109 before dispatching the order is 0.24, and the system sets the reference thermal resistance range for this type of server to be 0.17 to 0.20;

[0174] After the maintenance personnel complete the surface dust removal maintenance, the system will recalculate the average index within the next 24 hours. If the result is 0.19, it means that the server has returned to the healthy range and the maintenance is effective. The system can deduct a preset compensation value from the current equivalent service life, such as deducting 0.3 months, to reflect the life restoration effect brought about by this maintenance.

[0175] If the recalculated value is 0.23, which is still higher than the upper limit of the baseline range, it indicates that the root cause of the fault is not solely due to surface dust. The system should generate an upgrade work order, such as prompting to check for fan aging, thermal conductive material failure, or localized overheating of the motherboard.

[0176] The preset compensation value can be fixed or set in stages; for example, if the current index falls by more than 20%, 0.5 months will be deducted; if the fall is between 10% and 20%, 0.3 months will be deducted.

[0177] If the decline is less than 10% but has returned to the baseline range, deduct 0.1 months; this provides a more detailed reflection of the maintenance effect, rather than a simple threshold classification.

[0178] As an exception handling mechanism, if the asset undergoes business migration, power consumption mode switching, or shutdown maintenance within a preset time period after the completion signal is received, the data segment may not be comparable. The system should mark this evaluation as a delayed evaluation and resample after the equipment resumes stable operation.

[0179] If the historical dynamic thermal resistance index is lost, the index of the most recent stable window before the work order was generated can be used as a substitute; if the current index has not fallen back to the baseline range, but has decreased significantly from the historical value, for example from 0.30 to 0.22, the system can also mark it as a partial improvement and attach the improvement achieved in the upgrade work order to help the expert queue determine the subsequent priority.

[0180] After the initial maintenance of server 109 was completed, the system recalculated its dynamic thermal resistance index within the following 24 hours. The result dropped back to 0.19, entering the preset healthy range. Therefore, the maintenance was deemed effective, and the compensation value was deducted from the equivalent service life of the server.

[0181] If the index remains at 0.26 after surface dust removal and maintenance, the system will not simply shut down the process. Instead, it will automatically generate an upgrade work order and transfer it to the hardware expert queue, requiring further inspection of the fan and heat dissipation module.

[0182] The purpose of this step is to distinguish between work order completion and problem resolution, and to verify the quality of operation and maintenance with quantifiable thermal impedance fallback results, thereby achieving a truly closed-loop asset maintenance assessment.

[0183] In a preferred embodiment of the present invention, an abnormal circuit breaker module is further included. The abnormal circuit breaker module performs the following methods: calculating the temperature change rate of the microenvironment temperature data and continuously monitoring the temperature change rate; determining whether the temperature change rate is greater than a preset global fault threshold; if the temperature change rate is less than or equal to the global fault threshold, then the data processing flow of the age conversion module and the scheduling generation module is not interfered with.

[0184] If the temperature change rate exceeds the global fault threshold, the circuit breaker mechanism is triggered, the calculation of the equivalent service life is suspended, and a global physical alarm command is output. After the microenvironment temperature data recovers to the preset stable range and continues for the preset buffer time, the circuit breaker mechanism is released.

[0185] This embodiment provides an abnormal circuit breaker mechanism; specifically, the aforementioned solution is applicable to the identification of single asset or localized heat dissipation degradation, but if a data center-level global environmental failure occurs, such as precision air conditioning shutdown, cold aisle failure, or chiller unit malfunction, the applicability boundaries of the previous solution will be exposed:

[0186] At this time, the micro-environment temperature of almost all devices will rise rapidly, and the dynamic thermal resistance index and its trend will be uniformly raised by the global event. If the equivalent service life and dispatching are still calculated according to the local degradation logic, a large number of distorted depreciation results and redundant operation and maintenance scheduling instructions will be generated.

[0187] Therefore, this embodiment introduces an abnormal circuit breaker module to specifically identify the threshold-breaking characteristics of the global environment; the system continuously calculates the temperature change rate of the microenvironment temperature data, for example, by dividing the temperature difference between two adjacent sampling points by the time interval;

[0188] When the rate of change exceeds the preset global fault threshold, it indicates that the current anomaly is more likely to be an infrastructure-level fault rather than a single asset heat dissipation degradation. At this time, the circuit breaker mechanism is immediately triggered, the equivalent service life calculation and scheduling generation process is suspended, and a global physical alarm command is output to the superior monitoring or duty system.

[0189] Here is a specific example; assuming that under normal circumstances, the temperature change of the air inlet of a certain cabinet usually does not exceed 0.5℃ within 5 minutes, then the rate of change is approximately 0.1℃ / min;

[0190] The system sets the global fault threshold to 1.5℃ / min; if in a certain sampling, the temperature rises rapidly from 25℃ to 34℃ within 5 minutes, with a change rate of 1.8℃ / min, then the global fault threshold is exceeded.

[0191] At this point, the system no longer uses this window for age conversion of each server, but instead uniformly enters circuit breaker mode to avoid misrecording air conditioning failure as all assets aging rapidly at the same time.

[0192] The circuit breaker is not released immediately when the temperature drops slightly, but requires the microenvironment temperature to return to the preset stable range and remain there for a preset buffer time.

[0193] For example, if the stable range is set to 24℃ to 27℃ and the buffer time is set to 2 hours, the system will only restart the age conversion and scheduling process after the temperature returns to this range and remains there for 2 hours. This can avoid frequent opening and closing of the circuit breaker due to repeated temperature fluctuations.

[0194] As an anomaly handling mechanism, if only the temperature change rate of a few cabinets is abnormal while other cabinets remain stable, the scope of the anomaly can be used to determine whether it is a localized event.

[0195] If the abnormal range is less than the preset proportion, for example, only 2% of all cabinets, the system may not activate the global circuit breaker, but only freeze the assets in the corresponding cabinet.

[0196] If the temperature sensor itself malfunctions, causing an abnormal rate of change, the system can cross-verify with adjacent sensors to avoid false triggering. If the fuse duration is too long, such as exceeding the preset limit, the system can repeatedly push high-level physical alarms and prompt manual intervention.

[0197] During nighttime operation in the same core computer room, the chiller malfunctioned, and the front-end temperature of multiple cabinets rose from 24°C to over 33°C within 10 minutes.

[0198] The abnormal fuse module detected that the temperature change rate of most cabinets exceeded the global fault threshold at the same time. It immediately suspended the calculation of the equivalent service life of all servers, stopped generating regular maintenance work orders, and instead issued a data center-level physical environment abnormality alarm to the duty system.

[0199] The platform resumed its original calculation process only after the air conditioning system was restored and the ambient temperature remained stable at around 25°C for two consecutive hours.

[0200] The purpose of this mechanism is to clearly distinguish between local asset degradation and global infrastructure failure, and to avoid outputting distorted depreciation results and erroneous order assignments under extreme operating conditions, thereby achieving secure isolation between data center asset management logic and physical alarm logic.

[0201] In a preferred embodiment of the present invention, an asset audit module is further included. The method performed by the asset audit module includes: obtaining real-time power consumption data and dynamic thermal resistance index of the target asset, and obtaining service traffic data of the target asset from the service gateway; determining whether the real-time power consumption data of the asset is greater than or equal to a preset energy consumption threshold.

[0202] Determine whether the dynamic thermal resistance index is less than or equal to the preset thermal resistance threshold; determine whether the business flow data is less than the preset flow threshold.

[0203] If the real-time power consumption data of the asset is greater than the energy consumption threshold, the dynamic thermal resistance index is less than the thermal resistance threshold, and the business traffic data is less than the traffic threshold, then the target asset will be marked as an inefficient idle asset and an asset audit report will be generated; otherwise, the current marking status of the target asset will be maintained.

[0204] This embodiment provides an asset auditing mechanism; specifically, in the aforementioned operation and maintenance scheduling link, the system has continuously monitored asset power consumption, dynamic thermal impedance, and business-related information.

[0205] If this data is only used for fault prevention and not further used for asset inventory, the following technical defects will exist: There are often some devices in the data center that are still generating power consumption and have normal heat dissipation, but hardly carry any business; such devices may not trigger maintenance alarms, but will occupy rack space and power hardware resources for a long time.

[0206] Therefore, this embodiment adds an asset audit module based on the collected data to identify inefficient and idle assets; the system obtains the real-time power consumption data, dynamic thermal resistance index and business traffic data obtained from the business gateway of the target asset, and compares them with the preset energy consumption threshold, thermal resistance threshold and traffic threshold respectively.

[0207] If the asset's real-time power consumption data is greater than the energy consumption threshold, the dynamic thermal resistance index is less than the thermal resistance threshold, and the business traffic data is less than the traffic threshold, then it means that the target asset is not generating power consumption due to abnormal heat dissipation, nor is it increasing power due to busy business. It is more likely to be an inefficient asset that has been running for a long time but lacks actual load. Therefore, it is marked as an inefficient idle asset and an asset audit report is generated.

[0208] The following is a specific example for illustration; assuming that the current real-time power consumption data of server C is 180W, the preset energy consumption threshold is 150W; the dynamic thermal resistance index is 0.12, the preset thermal resistance threshold is 0.18; the service traffic data is 2Mbps, and the preset traffic threshold is 10Mbps;

[0209] Since 180W is greater than 150W, 0.12 is less than 0.18, and 2Mbps is less than 10Mbps, the system marks server C as an inefficient idle asset when all three conditions are met simultaneously.

[0210] The low dynamic thermal resistance index indicates that its heat dissipation performance has not deteriorated significantly, so the high power consumption is not an abnormal performance caused by overheating; while the service traffic data has been low for a long time, indicating that the energy consumed has not been converted into corresponding service output.

[0211] The generated asset audit report can include the target asset identifier, the cabinet it belongs to, the duration of continuous low traffic, the average real-time power consumption data of the asset, the dynamic thermal impedance level, and the recommended disposal method, such as business verification, merger and shutdown or migration and recycling; this can provide a direct basis for the asset department or business department to conduct subsequent inventory without the need to organize a large-scale manual investigation.

[0212] As an exception handling mechanism, if business traffic data is missing, the system should not directly determine that the asset is idle based solely on real-time power consumption data and dynamic thermal resistance index, but should first mark it as pending audit confirmation.

[0213] If a certain type of background task server naturally exhibits low traffic but high computing power consumption characteristics, an asset role whitelist can be introduced to avoid misjudgment.

[0214] If a server is briefly in a business switching window and business traffic data drops sharply but the duration is less than the preset observation period, such as less than 24 hours, it will not be marked immediately. Instead, it will be required to continuously meet three conditions for a preset duration before an asset audit report is output.

[0215] During continuous operation in the same core data center, in addition to thermal degradation scheduling of high-business-weight production servers such as server 109, the system also found that another legacy server maintained a power consumption of about 170W for a long time, and the dynamic thermal resistance index was stable at 0.11, indicating good heat dissipation. However, the business traffic data from the business gateway was almost zero, and this state lasted for more than 14 days.

[0216] The asset audit module marks it as an inefficient idle asset and generates an asset audit report, prompting the asset administrator to check whether it is a long-term idle and unprocessed faulty device;

[0217] The purpose of this mechanism is to identify inefficient and idle assets by reusing the results of joint analysis of environmental and business operations without adding additional physical hardware nodes, thereby achieving integrated linkage between operation and maintenance management and asset auditing.

[0218] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the claims of this invention should still fall within the patent coverage of this invention.

Claims

1. A data center asset operation and maintenance management system based on environmental data analysis, characterized in that, include: The data acquisition module is used to acquire the target asset's core temperature data, real-time power consumption data, micro-environment temperature data, business weight data, historical usage duration data, as well as the system's preset baseline trend and preset scrap age. The feature calculation module is communicatively connected to the data acquisition module. It is used to align the core temperature data of the asset, the microenvironment temperature data and the real-time power consumption data of the asset and obtain the dynamic thermal resistance index. It also extracts the degradation trend of the dynamic thermal resistance index based on a preset time window. An age conversion module, connected to the feature calculation module, is used to compare the degradation trend with the baseline trend, determine the penalty coefficient based on the comparison result, and perform equivalent loss conversion on the historical usage time data based on the penalty coefficient to obtain the equivalent service age of the target asset. The scheduling generation module is connected to the age conversion module and the feature calculation module respectively. It is used to calculate the scheduling priority score based on the equivalent service age, the scrapping age, the business weight data and the degradation trend, and generate operation and maintenance scheduling instructions based on the scheduling priority score. The feedback evaluation module, connected to the scheduling generation module and the age conversion module, is used to obtain the execution completion signal of the corresponding operation and maintenance scheduling instruction, obtain the historical dynamic thermal impedance index before execution and the current dynamic thermal impedance index after execution for comparison, and correct the equivalent service age based on the comparison results, or generate a work order upgrade instruction. It also includes an abnormal circuit breaker module, the method executed by the abnormal circuit breaker module includes: Calculate the rate of temperature change of the microenvironment temperature data and continuously monitor the rate of temperature change; Determine whether the temperature change rate is greater than a preset global fault threshold; If the temperature change rate is less than or equal to the global fault threshold, the data processing flow of the age conversion module and the scheduling generation module will not be interfered with. If the temperature change rate is greater than the global fault threshold, the circuit breaker mechanism is triggered, the calculation of the equivalent service life is suspended, and a global physical alarm command is output. After the microenvironment temperature data recovers to a preset stable range and remains within a preset buffer time, the circuit breaker mechanism is released. It also includes an asset audit module, the methods of which include: Obtain the real-time power consumption data and dynamic thermal resistance index of the target asset, and obtain the service traffic data of the target asset from the service gateway; Determine whether the real-time power consumption data of the asset is greater than or equal to a preset energy consumption threshold; Determine whether the dynamic thermal resistance index is less than or equal to a preset thermal resistance threshold; Determine whether the service traffic data is less than a preset traffic threshold; If the real-time power consumption data of the asset is greater than the energy consumption threshold, the dynamic thermal resistance index is less than the thermal resistance threshold, and the service traffic data is less than the traffic threshold, then the target asset is marked as an inefficient idle asset and an asset audit report is generated; otherwise, the current marking status of the target asset is maintained.

2. The data center asset operation and maintenance management system based on environmental data analysis according to claim 1, characterized in that, Methods for obtaining core asset temperature data, real-time asset power consumption data, and micro-environment temperature data of the target asset include: According to the preset sampling frequency, the initial core temperature data and initial power consumption data are read from the underlying management interface of the target asset. Simultaneously read initial ambient temperature data from the environmental sensors corresponding to the target asset; A sliding window is constructed with a preset time length, and the initial core temperature data, the initial power consumption data and the initial ambient temperature data are timestamped within the sliding window. Remove abrupt changes in data that exceed the preset upper and lower limits within the sliding window, and obtain the core temperature data of the asset, the real-time power consumption data of the asset, and the microenvironment temperature data.

3. The data center asset operation and maintenance management system based on environmental data analysis according to claim 2, characterized in that, The method for extracting the degradation trend of the dynamic thermal resistance index based on a preset time window includes: Extract all the dynamic thermal resistance indices within the preset time window to construct a dynamic thermal resistance index sequence. The dynamic thermal resistance index sequence is trend-fitted to obtain the linear slope corresponding to the dynamic thermal resistance index sequence, and the linear slope is used as the degradation trend.

4. The data center asset operation and maintenance management system based on environmental data analysis according to claim 3, characterized in that, The method for determining the penalty coefficient based on the comparison results, and then performing equivalent depreciation conversion on the historical usage duration data based on the penalty coefficient to obtain the equivalent service life of the target asset includes: Determine whether the degradation trend is greater than the baseline trend; If the degradation trend is less than or equal to the baseline trend, the runtime corresponding to the preset time window is directly added to the historical usage duration data to obtain the equivalent service life. If the degradation trend is greater than the baseline trend, then assess the degree to which the degradation trend exceeds the baseline trend. By querying a preset mapping table between the degree of exceeding the standard and the penalty coefficient, the corresponding penalty coefficient is obtained, and the running time corresponding to the preset time window is converted based on the penalty coefficient to obtain the acceleration loss time. The accelerated wear duration is accumulated into the historical usage duration data to obtain the equivalent service life.

5. The data center asset operation and maintenance management system based on environmental data analysis according to claim 4, characterized in that, The method for calculating scheduling priority scores and generating operation and maintenance scheduling instructions based on the scheduling priority scores includes: The age approximation rate is obtained based on the equivalent service age and the scrapping age. The scheduling priority score is calculated by integrating the age approximation rate, the business weight data, and the degradation trend. Obtain a preset scheduling threshold, and compare the scheduling priority score with the scheduling threshold; If the scheduling priority score is less than the scheduling threshold, no instruction is generated; If the scheduling priority score is greater than or equal to the scheduling threshold, then the operation and maintenance scheduling instruction containing the target asset identifier and maintenance suggestions is generated.

6. The data center asset operation and maintenance management system based on environmental data analysis according to claim 1, characterized in that, The method of obtaining and comparing the historical dynamic thermal impedance index before execution and the current dynamic thermal impedance index after execution, and then correcting the equivalent service life based on the comparison results, or generating a work order upgrade instruction, includes: In response to the execution completion signal, the dynamic thermal resistance index of the target asset is recalculated within a preset time period after execution and used as the current dynamic thermal resistance index. Extract the dynamic thermal impedance index when the operation and maintenance scheduling instruction is generated, and use it as the historical dynamic thermal impedance index. Determine whether the current dynamic thermal resistance index has fallen back to the preset reference thermal resistance range; If the current dynamic thermal resistance index falls back to the reference thermal resistance range, the operation and maintenance are deemed effective, and a preset penalty compensation value is deducted from the equivalent service life. If the current dynamic thermal resistance index does not fall back to the reference thermal resistance range, the operation and maintenance is deemed invalid, the work order upgrade instruction is generated, and the work order upgrade instruction is transferred to the preset expert processing queue.

Images