Intelligent monitoring method and system for building electrical system

Abstract

This application provides an intelligent monitoring method and system for building electrical systems. It includes acquiring historical monitoring flow rates of high-power electronic devices through a monitoring system; determining intraday flow rate variation curves and inter-day flow rate variation trends based on the historical monitoring flow rates; calculating the average estimated flow rate for a preset future time period based on the intraday flow rate variation curves, inter-day flow rate variation trends, and the actual monitoring flow rate of the day; determining the estimated average operating power of high-power electronic devices in a large computing building during the preset future time period based on the average estimated flow rate; and determining the optimal heat dissipation power of the forced air cooling system during the preset future time period based on the estimated average operating power, for use in controlling the operation of the forced air cooling system. This method combines the actual monitoring flow rate of the day to calculate the average estimated flow rate for a preset future time period, thereby using this estimated flow rate method to perform feedforward control of the forced air cooling system, overcoming the delay defects caused by current feedback control strategies.

Description

Technical Field

[0001] This invention relates to the field of electrical intelligence technology, and in particular to an intelligent monitoring method and system for building electrical systems. Background Technology

[0002] With the rapid development of the digital economy, the construction scale of large computing power buildings such as computing centers and communication equipment rooms is constantly expanding. These buildings are densely packed with high-power electronic devices such as servers and switches. In practical applications, these high-power electronic devices generate a lot of heat during continuous operation. If the heat cannot be dissipated in a timely and effective manner, it will directly affect the stability of their operation and may even lead to equipment failure and service interruption. Therefore, large computing power buildings such as computing centers have extremely high requirements for heat dissipation systems.

[0003] Currently, the mainstream cooling method for large computing buildings is to deploy a forced air cooling system consisting of multiple cooling fan modules (usually forming a "fan wall") within the building, thereby driving airflow to remove the heat generated during equipment operation. However, in practical applications, the cooling fan modules of the forced air cooling system itself also have high operating power, resulting in a significant waste of electrical energy during operation. Therefore, how to utilize these cooling fan modules for heat dissipation more efficiently is an important research direction.

[0004] The technical solutions disclosed in patents CN119597128A and CN110319043A mainly rely on direct monitoring and feedback of the ambient temperature of the computer room, thereby adjusting the speed or rotation angle of the cooling fans to more accurately direct the cooling airflow to high-temperature areas, aiming to improve the local heat dissipation effect. The temperature energy-saving control method for communication computer rooms disclosed in patent CN120321922A also utilizes temperature data, humidity data, and communication equipment load data from multiple areas within the communication computer room, and ultimately achieves the allocation of heat dissipation power through the energy-saving control method described in this patent.

[0005] However, the core idea of ​​the prior art disclosed in the aforementioned patents is to regulate the power, speed, or direction of the cooling fan by detecting current data such as temperature. This regulation approach belongs to a feedback control strategy, and its control action is inherently delayed and passive. For example, a temperature increase is a result of an imbalance between the heat generated by high-power electronic devices and the cooling capacity of the forced air cooling system. In this control method, regulation to increase heat dissipation power is only initiated after this temperature increase signal is detected, resulting in an unavoidable delay in the entire process. This delay is an inherent defect of the current feedback control strategy. Obviously, during this delay period, the high-power electronic device may already be under high thermal stress, which is detrimental to its long-term reliable operation. Summary of the Invention

[0006] This invention provides an intelligent monitoring method and system for building electrical systems, which solves the problem of control delay caused by feedback control strategies in the prior art.

[0007] On one hand, the present invention provides an intelligent monitoring method for building electrical systems, comprising: Historical monitoring traffic of high-power electronic devices within large computing power buildings is obtained through a monitoring system. Based on the historical monitored flow rates, determine the intraday flow rate change curve and the interday flow rate change trend; Based on the intraday flow change curve, the interday flow change trend, and the actual monitored flow on the day, the average estimated flow for a preset future time period is calculated. Based on the average estimated traffic, the estimated average operating power of high-power electronic devices in the large computing building is determined in the preset future time period. The optimal heat dissipation power of the forced air cooling system in the preset future time period is determined based on the estimated average operating power, so as to control the operation of the forced air cooling system.

[0008] Preferably, the forced air cooling system includes multiple cooling fan modules; the method further includes: The operating parameters of each of the cooling fan modules are obtained through the monitoring system. Based on the operating parameters, determine the health assessment value of each of the cooling fan modules; Based on the health assessment value of each of the cooling fan modules, the optimal heat dissipation power is allocated among the multiple cooling fan modules.

[0009] Preferably, the operating parameters include at least one of the following parameters of the cooling fan module: electrical parameters, performance parameters, reliability parameters, and physical state parameters; and, Based on the operating parameters, the health assessment value of each cooling fan module is determined. Specifically, the health assessment value of each cooling fan module is determined in the following manner: Based on each parameter in the operating parameters, calculate the individual health score corresponding to each parameter; The health assessment value of the cooling fan module is determined by weighted summation based on the individual health scores corresponding to each parameter in the operating parameters.

[0010] Preferably, based on the health assessment value of each of the cooling fan modules, the optimal heat dissipation power is allocated among the plurality of cooling fan modules, specifically including: Based on the health assessment value of each cooling fan module, the multiple cooling fan modules are divided into a main fan group and a backup fan group. Determine the total rated power of each cooling fan module in the main fan group, and determine whether the total rated power is greater than or equal to the optimal cooling power; If the total rated power is greater than or equal to the optimal heat dissipation power, then the optimal heat dissipation power is evenly distributed among the heat dissipation fan modules of the main fan group; or, If the total rated power is less than the optimal heat dissipation power, then each heat dissipation fan module in the main fan group is controlled to operate at rated power, and the power difference between the optimal heat dissipation power and the total rated power is calculated. Based on the power difference, N heat dissipation fan modules are selected from the standby fan group, and the power difference is evenly distributed among the selected M heat dissipation fan modules, where M = [power difference / (rated power × reduction factor)] + 1.

[0011] Preferably, based on the historical monitored flow rate, the intraday flow rate change curve and the interday flow rate change trend are determined, specifically including: S121. Arrange the original sequence of the historical monitored flow in time sequence according to a preset time granularity; S122. Based on the fluctuation amplitude and periodic characteristics of historical monitoring flow after time-series arrangement, determine the optimal smoothing coefficient for extracting long-term trends. S123. Using the exponentially weighted moving average algorithm and applying the optimal smoothing coefficient, iterative calculations are performed on the historical monitoring flow after time-series arrangement to obtain an initial diurnal variation trend sequence characterizing the long-term change direction. S124. Remove the initial diurnal variation trend sequence from the original sequence of the historical monitored flow to obtain the first residual sequence; S125. Perform a stationarity test on the first residual sequence. If the test passes, use the initial diurnal variation trend sequence as the diurnal flow variation trend. If the test fails, adjust the optimal smoothing coefficient and re-execute steps S123 and S124 until the new first residual sequence passes the stationarity test, and then obtain the diurnal flow variation trend. S126. Remove the daytime flow change trend from the original sequence of the historical monitored flow to obtain the second residual sequence; S127. Based on the second residual sequence, calculate the historical flow statistics at the same time points within a single natural day, wherein the historical flow statistics include the arithmetic mean, median, or mode. S128. Connect the historical flow statistics at each time point to form the intraday flow change curve that characterizes the periodic fluctuation of intraday flow.

[0012] Preferably, based on the intraday flow rate change curve, the inter-day flow rate change trend, and the actual monitored flow rate for the day, the average estimated flow rate for a preset future time period is calculated, specifically including: Based on the daytime flow variation trend, the theoretical total flow for the day is predicted using an extrapolation algorithm; Based on the intraday flow change curve, calculate the area under the first curve from midnight to the current time, the area under the second curve for a preset future time period, and the area under the third curve for the remaining time of the day. The theoretical flow rate for the day is determined based on the area under the first curve and the area under the third curve, as well as the theoretical total flow rate for the day. Based on the area under the second curve, the area under the third curve, and the area under the first curve, as well as the theoretical total flow for the day, estimate the theoretical total flow for the preset future time period; The ratio of the actual monitored flow rate to the theoretical flow rate of the day is used as a real-time correction factor. Multiply the theoretical total flow of the preset future time period by the real-time correction factor, and then divide by the duration of the preset future time period to obtain the average estimated flow of the preset future time period.

[0013] Preferably, based on the average estimated traffic flow, the estimated average operating power of high-power electronic devices within the large computing building during the preset future time period is determined, specifically including: Multiple constant traffic loads Fi at different levels are applied to the high-power electronic device, and the operating power Pi of the high-power electronic device is measured simultaneously to obtain a series of data pairs (Fi, Pi); Regression analysis was performed on the series of data pairs (Fi, Pi) to obtain the corresponding relationship between flow rate and operating power. Substitute the average estimated flow rate into the corresponding relationship to obtain the estimated average operating power.

[0014] Preferably, determining the optimal heat dissipation power of the forced air cooling system in the preset future time period based on the estimated average operating power specifically includes: Multiply the estimated average operating power by a heat dissipation efficiency coefficient to obtain the baseline heat dissipation power; Based on the real-time environmental parameters of the large computing building, a heat dissipation power correction value is calculated; the environmental parameters include at least one of temperature, humidity, air pressure and natural wind speed inside and outside the building. The optimal heat dissipation power is calculated using the heat dissipation power correction value and the baseline heat dissipation power.

[0015] Preferably, the method further includes: The high-power electronic device and the forced air cooling system are deployed in an experimental unit. Control the high-power electronic devices within the experimental unit to operate at at least two different operating power P it Below, and at each operating power P it Next, adjust the heat dissipation power P of the forced air cooling system. cool This continues until the temperature at the key temperature measurement points within the experimental unit stabilizes within the safe target temperature range, in order to obtain multiple power pairs (P... it P cool ); Through each power pair (P) it P cool P in ) cool With P it The ratio of the values ​​is used to calculate the heat dissipation efficiency coefficients corresponding to the multiple operating powers.

[0016] Secondly, the present invention provides an intelligent monitoring system for building electrical systems, comprising: The acquisition unit is used to acquire historical monitoring traffic of high-power electronic devices in large computing buildings through the monitoring system. The determining unit is used to determine the intraday flow change curve and the interday flow change trend based on the historical monitored flow. The calculation unit is used to calculate the average estimated flow for a preset future time period based on the intraday flow change curve, the interday flow change trend and the actual monitored flow on the day. The estimation unit is used to determine the estimated average operating power of high-power electronic devices in the large computing building during the preset future time period based on the average estimated traffic. The second determining unit is used to determine the optimal heat dissipation power of the forced air cooling system in the preset future time period based on the estimated average operating power, so as to control the operation of the forced air cooling system.

[0017] The intelligent monitoring method for building electrical systems provided in this application includes: acquiring historical monitoring flow rates of high-power electronic devices within a large computing building through a monitoring system; determining intraday flow rate variation curves and inter-day flow rate variation trends based on the historical monitoring flow rates; calculating the average estimated flow rate for a preset future time period based on the intraday flow rate variation curves, the inter-day flow rate variation trends, and the actual monitoring flow rate of the day; determining the estimated average operating power of the high-power electronic devices within the large computing building during the preset future time period based on the average estimated flow rate; and determining the optimal heat dissipation power of the forced air cooling system during the preset future time period based on the estimated average operating power, for use in controlling the operation of the forced air cooling system. This method obtains intraday flow rate variation curves and inter-day flow rate variation trends through historical monitoring flow rates, and then combines this with the actual monitoring flow rate of the day to calculate the average estimated flow rate for a preset future time period. This estimated flow rate method enables feedforward control of the forced air cooling system, which is more proactive, overcomes the delay defects of current feedback control strategies, and is more conducive to the long-term effective operation of high-power electronic devices. Attached Figure Description

[0018] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0019] Figure 1 A flowchart illustrating the intelligent monitoring method for building electrical systems provided by this invention; Figure 2 The present invention provides a flowchart illustrating the process for determining the intraday flow rate change curve and the interday flow rate change trend in the intelligent monitoring method for building electrical systems. Figure 3 The present invention provides a schematic diagram of the process for calculating the average estimated flow rate over a preset future time period in the intelligent monitoring method for building electrical systems. Figure 4 The present invention provides a structural block diagram of an intelligent monitoring system for building electrical systems; Figure 5 A schematic diagram of the structure of the electronic device provided by the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0021] As mentioned earlier, with the development of the digital economy, the scale of large computing buildings such as computing centers and communication equipment rooms is constantly expanding, and these buildings are densely packed with high-power electronic devices such as servers and switches. These devices generate a large amount of heat during continuous operation. If this heat cannot be dissipated effectively and in a timely manner, it will affect operational stability and may even lead to equipment failure and service interruption. Currently, mainstream heat dissipation methods rely on real-time monitoring of data such as ambient temperature, humidity, and equipment load in the data center, and use feedback control strategies to adjust the speed or power of cooling fans. This type of method uses temperature rise as a trigger condition and has an inherent lag, only initiating a response after thermal imbalance has occurred. This results in high-power electronic devices being under high thermal stress during the delay period, which is detrimental to long-term reliable operation.

[0022] In view of this, embodiments of this application provide an intelligent monitoring method and system for building electrical systems, which can be used to solve the problems in the prior art. In this embodiment, the building electrical system is a closed-loop control system integrating data sensing, intelligent analysis, predictive decision-making, and precise execution functions. The building electrical system mainly consists of two core components: a monitoring system and a forced air cooling system. The two can communicate and work collaboratively through a data bus and control network.

[0023] The monitoring system, serving as the intelligent hub of the entire large-scale computing building, is responsible for the entire data acquisition, processing, analysis, and decision generation process. On one hand, it can collect data, such as communicating with various sensors and equipment management interfaces within the building to continuously acquire real-time traffic data. On the other hand, it can perform in-depth time-series analysis on the collected traffic data, extracting intraday traffic variation curves and inter-day traffic variation trends. Therefore, this monitoring system is the core and decision-making module of the building's electrical system. The forced-air cooling system, as the execution terminal of the building's electrical system, is typically physically composed of multiple cooling fan modules deployed within the building (such as a fan wall arranged in an array). In this application, the monitoring system controls the cooling of high-power electronic devices within the large-scale computing building. The technical solutions provided in the embodiments of this application will be described in detail later.

[0024] like Figure 1 The diagram shown is a schematic flowchart of a smart monitoring method for a building electrical system provided in an embodiment of this application. The method includes the following steps: Step S11: Obtain historical monitoring traffic of high-power electronic devices within a large computing building through a monitoring system.

[0025] In this application, the large computing building may include a computing center, a communication room, etc., and high-power electronic devices are deployed in the large computing building. These high-power electronic devices may include servers, switches, etc. Due to their high power consumption, these high-power electronic devices generate a lot of heat when they are running.

[0026] The historical monitoring traffic can refer to the business-level data throughput monitoring value generated by high-power electronic devices in large computing buildings within a preset historical period (such as the past 30 days, the past six months, etc.). This can be network layer traffic (e.g., the number of bytes received / sent via Ethernet ports), computing layer task scheduling throughput (e.g., requests processed per second (QPS), or storage layer I / O throughput (e.g., data read / write volume per second (MB / s)). In practical applications, this historical monitoring traffic can be collected by traffic probes, SNMP collectors, eBPF observation modules, or embedded sensors deployed locally on the device or at aggregation nodes, and uploaded to the monitoring system, enabling the monitoring system to acquire this historical monitoring traffic.

[0027] For example, in a supercomputing center's computer room, the monitoring system continuously acquires the RDMA network throughput of all 256 GPU servers over 30 days, with a sampling period of 1 minute, forming a raw sequence of historical monitoring traffic containing 43,200 time points. This historical monitoring traffic can fully cover the operating conditions of weekdays, weekends, and holidays, thus exhibiting typical periodicity and long-term evolution characteristics.

[0028] Step S12: Based on historical monitoring flow, determine the intraday flow change curve and the interday flow change trend.

[0029] The intraday flow change curve can be a periodic curve characterizing the statistical regularity of flow at the same time point within a single natural day, and can be any one of the mean curve, median curve, or mode curve; the interday flow change trend can be a non-periodic change direction reflecting the long-term growth, decline, or stable evolution of historical monitored flow on a multi-day scale, and can be any one of the monotonically upward trend, monotonically downward trend, slow fluctuation trend, or approximately horizontal trend; in this application, the intraday flow change curve and the interday flow change trend together constitute a dual decomposition expression of the historical flow time series structure, respectively characterizing the short-term periodic regularity and long-term non-periodic regularity in historical monitored flow.

[0030] In one implementation of step S12 of this application, based on the original sequence of historical monitored traffic, the data for each natural day can be aligned hourly, and then the average value of all traffic samples within the same hourly interval each day can be calculated. Finally, the 24 mean points are connected to form a daily traffic change curve. A sliding window method can be used to statistically analyze the median of the traffic value at each hour within a rolling 7-day window, and then a smooth curve is fitted as the daily traffic change curve. Of course, the median can also be replaced by the mode as the representative traffic value, thus making it suitable for business scenarios with obvious bimodal or multimodal distributions. Regarding the daily traffic change trend, this application can use an exponentially weighted moving average algorithm to extract the daily traffic change trend, where the smoothing coefficient is dynamically determined based on the decay rate and variance ratio of the autocorrelation function of the historical sequence.

[0031] It should be further noted that, considering the impact of daytime traffic flow trends on the accuracy of intraday traffic flow curves, causing shifts, therefore... Figure 2 As shown, step S12 can also be performed in the following manner to determine the intraday flow rate change curve and the interday flow rate change trend: Step S121: Arrange the original sequence of historical monitoring flow in time sequence according to the preset time granularity.

[0032] The preset time granularity can be, for example, 1 minute, 5 minutes, 15 minutes, or 1 hour. By arranging the original sequences of historical monitored traffic according to the preset time granularity, the sampling interval can be standardized and an ordered data sequence with a time index can be constructed. In practical applications, the selection of this time granularity needs to balance data resolution and computational efficiency. For example, in the scenario of large-scale computing power buildings, the preset time granularity can be 15 minutes, thereby providing a structured input basis for subsequent trend extraction and periodic analysis.

[0033] In step S121 of this application, for example, the original sequence of historical monitoring traffic can be arranged in ascending or descending order according to a preset time granularity (e.g., 15 minutes). Taking ascending order as an example, the preset time granularity can be 15 minutes, combined with the timestamp of the original sequence in the historical monitoring traffic, and arranged in ascending order. Of course, the default values ​​in the original sequence of the historical monitoring traffic can also be filled in (e.g., by interpolation) to form a continuous and equally spaced sequence.

[0034] For example, this application could obtain network inbound traffic data of a server cluster in a data center, recorded every 15 minutes for 30 consecutive days. Therefore, a total of 2880 monitoring traffic points were collected. Arranging this data in ascending order by timestamp would form a one-dimensional time-series array of length 2880. ,in Indicates the first The monitored flow rate at each time point is represented by a one-dimensional time-series array, which is the historical monitored flow rate arranged in time series. For ease of description, the historical monitored flow rate arranged in time series will be represented as an array thereafter. Furthermore, the original sequence of historical monitored traffic is represented as an array. N is the total number of time points. In this example, N is 2880.

[0035] Step S122: Based on the fluctuation amplitude and periodic characteristics of the historical monitoring flow after time-series arrangement, determine the optimal smoothing coefficient for extracting long-term trends.

[0036] The fluctuation amplitude can be used to measure the short-term instability and noise level of a data series. In practical applications, it can be calculated as the coefficient of variation of the historical monitoring flow after the time series is arranged, which is the ratio of the standard deviation to the arithmetic mean of the historical monitoring flow after the time series is arranged, and is taken as the fluctuation amplitude. For example, based on the above one-dimensional time series array... It can be calculated The standard deviation and arithmetic mean are then used to calculate the fluctuation range, which is obtained by dividing the standard deviation by the arithmetic mean.

[0037] This periodicity feature is used to quantify the regularity of data over a daily (24-hour) cycle. In practical applications, Fourier transforms can be performed on the historical monitoring flow after the time series is arranged to analyze its frequency components and calculate the proportion of signal energy at the 24-hour fundamental frequency and its main harmonic frequencies to the total signal energy, i.e., the dominant frequency energy proportion, as this periodicity feature. The higher the dominant frequency energy proportion, the more significant and stable the daily periodicity pattern is.

[0038] After obtaining the fluctuation range and periodicity characteristics of the historical monitoring flow after the time series arrangement, the optimal smoothing coefficient can be determined using these fluctuation range and periodicity characteristics. Specifically, the optimal smoothing coefficient can be obtained by querying the smoothing coefficient mapping table. For example, Table 1 shows the smoothing coefficient mapping table.

[0039] Table 1

[0040] Thus, in step S122, after obtaining the fluctuation range and periodic characteristics of the historical monitoring flow after the time sequence arrangement, the table can be looked up to obtain the corresponding optimal smoothing coefficient. In this application, the optimal smoothing coefficient α specifically refers to the ability of the exponentially weighted moving average algorithm used in the subsequent step S123 to control the rate of weight decay of historical data. The choice of the value of the optimal smoothing coefficient α directly affects the effect of trend extraction. For example, the larger the value of α, the more the algorithm emphasizes recent data and is more sensitive to short-term fluctuations; the smaller the value of α, the more the algorithm tends to smooth noise and retain the long-term evolution direction.

[0041] Step S123: Using the exponentially weighted moving average algorithm and applying the optimal smoothing coefficient, iteratively calculate the historical monitoring flow after time series arrangement to obtain the initial diurnal variation trend sequence representing the long-term change direction.

[0042] Among them, the Exponentially Weighted Moving Average (EWMA) algorithm is a filtering method that recursively calculates a weighted average value for time-series data (i.e., historical monitoring flow arranged in time series). Its mathematical expression is: t i =α× +(1 α)×t i 1, i = 1, 2, ..., N.

[0043] In this formula, For example, the i-th data point in the historical monitoring traffic after time-series arrangement, such as the one-dimensional time-series array mentioned above. The third data point is α is the optimal smoothing coefficient; t0 takes the value of By combining the optimal smoothing coefficient α with the historical monitoring flow rates arranged in time sequence, t1, t2...t can be calculated sequentially. N This constitutes the initial diurnal variation trend sequence T = {t1, t2, ..., t...} N}

[0044] Step S124: Remove the initial diurnal variation trend sequence from the original sequence of historical monitored flow to obtain the first residual sequence.

[0045] In this application, the method of removing the initial diurnal variation trend sequence from the original sequence of historical monitored flow can subtract the corresponding data points to obtain the first residual sequence.

[0046] For example, the original sequence of historical monitoring traffic is an array. The initial daily trend sequence is an array T = {t1, t2, ..., t}.N At this point, r1 can be calculated. - t1, r2= - t2……,r N = - t N Thus, the residual sequences r1, r2, ..., r are obtained. N The residual sequence r1, r2, ..., r can be... N Represented as R = {r1, r2, ..., r} N The initial diurnal variation trend sequence is removed from the original sequence of historical monitored flow in this way, thereby eliminating the part of historical monitored flow that is not explained by the trend component. Theoretically, this should mainly include intraday periodic fluctuations and random noise, thus reducing the impact of diurnal variation trend on intraday flow variation curve.

[0047] Step S125: Perform a stationarity test on the first residual sequence. If the test passes, the initial diurnal variation trend sequence is taken as the diurnal flow variation trend. If the test fails, adjust the optimal smoothing coefficient and repeat steps S123 and S124 until the new first residual sequence passes the stationarity test, thus obtaining the diurnal flow variation trend.

[0048] In this embodiment, the stationarity of the first residual sequence can be tested using the Augmented Dickey-Fullertes (ADF) test. Specifically, the first residual sequence R = {r1, r2, ..., r...} can be tested first. N A regression equation for the ADF test is established, and the lag order p is determined based on this regression equation. Hypothesis testing is then performed to determine whether the test passes. If the test passes (p-value less than 0.05), the first residual sequence R = {r1, r2, ..., r...} can be determined. N The data is stable, which indicates that the initial diurnal variation trend sequence is effective and can be used as the diurnal flow variation trend.

[0049] Conversely, if the test fails, it can be determined that the first residual sequence still contains trend or long-term fluctuation components, indicating that the currently used smoothing coefficient α may be inappropriate (usually because α is too large, resulting in insufficient trend extraction). At this time, the optimal smoothing coefficient can be adjusted (for example, reducing the value of α to capture a longer-term trend, which can be done by multiplying the value of α by 0.95). Then, steps S123 and S124 are re-executed to generate a new trend sequence and residual sequence, and the ADF test is performed again until a stationary residual sequence is obtained. At this time, the corresponding initial intraday change trend sequence can be used as the intraday flow change trend.

[0050] Step S126: Remove the daytime flow change trend from the original sequence of historical monitoring flow data to obtain the second residual sequence.

[0051] After obtaining the daily flow change trend through the above step S125, in this step S126, the daily flow change trend is further removed from the original sequence of historical monitoring flow data. The method can be the same as the above step S124, that is, subtract the corresponding data points to obtain the second residual sequence. This will not be elaborated here.

[0052] Step S127: Based on the second residual sequence, calculate the historical flow statistics at the same time points within a single natural day. The historical flow statistics include the arithmetic mean, median, or mode.

[0053] Step S128: Connect the historical flow statistics at each time point to form the intraday flow change curve that characterizes the periodic fluctuation of intraday flow.

[0054] Here we can provide a unified explanation for steps S127 and S128.

[0055] A single natural day can be a continuous 24-hour period; each identical time point refers to aligning all historical data to the same day template by hour (or finer granularity), for example, grouping all data from the 8th hour (i.e., 08:00-09:00) into one group; historical traffic statistics can be the arithmetic mean of all samples in the group, or the median which is more resistant to outlier interference, or the mode which is suitable for discrete distributions, etc.

[0056] In step S127, after obtaining the second residual sequence through step S126, the historical flow statistics at the same time point within a single natural day are calculated. For example, the arithmetic mean of all 8th hours (or the arithmetic mean of the 9th hour, etc.) can be calculated using the second residual sequence to obtain the historical flow statistics at the same time point within a single natural day. The historical flow statistics at each time point are then connected to obtain the flow change curve for that day. Obviously, the flow change curve for that day can characterize the periodic fluctuation of the flow within the day.

[0057] The intraday flow rate change curve and interday flow rate change trend obtained by this application through S121 to S12 can decouple the long-term intraday trend and intraday periodic fluctuation in the historical flow rate of high-power electronic devices, significantly improving the accuracy and robustness of future flow rate prediction. It provides a high-quality advance prediction basis for solving the problem of untimely heat dissipation control caused by thermal response lag in the prior art.

[0058] Step S13: Based on the intraday flow change curve, the intraday flow change trend, and the actual monitored flow on the day, calculate the average estimated flow for the preset future time period.

[0059] Among them, the actual monitored traffic on the day can refer to the actual traffic value of high-power electronic devices collected by the monitoring system from midnight (e.g., 0:00) to the current time (e.g., 8:00); the preset future time period can refer to the time interval set by the user in advance, which can be 15 minutes, 1 hour, 2 hours or 4 hours in the future; the average estimated traffic can refer to the expected average traffic per unit time within the future time period, which can reflect the overall business load intensity of the equipment in the future period.

[0060] In practical applications, such as Figure 3 As shown, step S13 can be implemented in the following way to calculate the average estimated flow rate for a preset future time period: Step S131: Based on the daily flow trend, predict the theoretical total flow for the day using an extrapolation algorithm.

[0061] Among them, the daily flow change trend (a sequence representing the long-term change direction) is a low-frequency component extracted from historical monitored flow that reflects the overall upward / downward / stable trend over a multi-day scale. This has been explained in the aforementioned implementation method and will not be repeated here.

[0062] The extrapolation algorithm can refer to a method for extending the daily flow trend sequence based on a time series model (such as linear regression, ARIMA, or exponential smoothing). In the embodiments of this application, the extrapolation algorithm extends the known daily trend to the current day, thereby obtaining a macro-estimate of the overall flow scale of the current day. This estimate serves as the total benchmark for subsequent area ratio allocation.

[0063] This application can, for example, extrapolate the total daily flow (the theoretical total flow) based on the linear fitting result of the daily flow variation trend, or it can calculate the theoretical total flow using a sliding window extrapolation method based on the exponentially weighted moving average series of the daily flow variation trend. Alternatively, this application can also extrapolate the predicted flow for each time period (i.e., what time of day) based on the linear fitting result of the daily flow variation trend, or calculate the predicted flow for each time period using a sliding window extrapolation method based on the exponentially weighted moving average series of the daily flow variation trend, and then sum the predicted flows for each time period to obtain the theoretical total flow for the day.

[0064] Step S132: Based on the intraday flow change curve, calculate the area under the first curve from midnight to the current time, the area under the second curve for the preset future time period, and the area under the third curve for the remaining time of the day.

[0065] The area under the first curve can refer to the integral area enclosed on the intraday flow rate change curve from 00:00 on the day to the current time; the area under the second curve can refer to the integral area enclosed on the intraday flow rate change curve from the current time to the end of the preset future time period; and the area under the third curve can refer to the integral area enclosed on the intraday flow rate change curve from the current time to 24:00 on the day.

[0066] In this application, after obtaining the intraday flow change curve, the current time can be determined, and then the integral area enclosed on the intraday flow change curve from midnight to the current time can be obtained through integration calculation, which is the area under the first curve; of course, based on the same principle, the areas under the second curve and the areas under the third curve can be calculated.

[0067] Step S133: Determine the theoretical flow rate for the day based on the area under the first curve and the area under the third curve, as well as the theoretical total flow rate for the day.

[0068] The theoretical daily flow rate refers to the flow rate theoretically generated by the high-power electronic device from midnight to the present moment. In this application, the sum of the areas under the first curve and the areas under the third curve is the total integrated area under the daily flow rate change curve, which corresponds to the theoretical total flow rate for the day. Therefore, the theoretical daily flow rate can be obtained by dividing the area under the first curve by the sum of the areas under the first curve and the third curve, and then multiplying this ratio by the theoretical total flow rate for the day. That is, the theoretical daily flow rate = theoretical total flow rate for the day × area under the first curve / (area under the first curve + area under the third curve).

[0069] Step S134: Estimate the theoretical total flow for the preset future time period based on the area under the second curve, the area under the third curve, the area under the first curve, and the theoretical total flow for the day.

[0070] As mentioned above, the sum of the areas under the first curve and the areas under the third curve is the total integrated area under the daily flow change curve, which corresponds to the theoretical total flow for that day. Therefore, the theoretical total flow for the preset future time period can be calculated by multiplying the theoretical total flow for that day by the area under the second curve and then by (the area under the first curve + the area under the third curve).

[0071] Step S135: Use the ratio of the actual monitored flow rate to the theoretical flow rate of the day as a real-time correction factor.

[0072] In this embodiment, the real-time correction factor is the ratio of the actual monitored flow rate to the theoretical flow rate on the same day. It reflects the degree of deviation between the current actual operating status (characterized by the actual monitored flow rate on the same day) and the periodic trend model (characterized by the theoretical flow rate on the same day), and can be used as a subsequent correction amount.

[0073] Step S136: Multiply the theoretical total flow of the preset future time period by the real-time correction factor, and then divide by the duration of the preset future time period to obtain the average estimated flow of the preset future time period.

[0074] In other words, the average estimated flow for the preset future time period = the theoretical total flow for the future time period × the real-time correction factor. Then, the product is divided by the duration of the preset future time period, which can be in minutes to calculate the average estimated flow per minute. Alternatively, the duration can be in hours to calculate the average estimated flow per hour.

[0075] Step S14: Based on the average estimated flow, determine the estimated average operating power of high-power electronic devices in the large computing building over a preset future time period.

[0076] In practical applications, one optional implementation of step S14 is to first apply multiple constant traffic loads Fi at different levels to the high-power electronic device, and simultaneously measure the corresponding operating power Pi of the high-power electronic device to obtain a series of data pairs (Fi, Pi). Here, the constant traffic load Fi can refer to network traffic, computing task flow, or storage I / O throughput continuously injected into the high-power electronic device under experimental conditions, with a stable amplitude and a duration sufficient to allow the device to enter a thermal steady state. The operating power Pi of the high-power electronic device can refer to the measured active power of the entire high-power electronic device or target functional unit on the power supply circuit under the corresponding constant traffic load Fi. In this application, a one-to-one mapping relationship is established between the constant traffic load Fi and the synchronously measured operating power Pi, which reflects the inherent power consumption characteristics of the high-power electronic device under specific operating conditions and power supply conditions.

[0077] After obtaining a series of data pairs (Fi, Pi), regression analysis can be performed on these pairs to fit the relationship between traffic flow and operating power. This regression analysis can refer to the process of deriving a continuous functional relationship from discrete data pairs based on statistical learning methods such as least squares, ridge regression, or multinomial fitting. For example, Fi can be used as the independent variable and Pi as the dependent variable, and the least squares method can be used to linearly fit the series of data pairs (Fi, Pi) to obtain the relationship between traffic flow and operating power. Typically, this relationship can be expressed as an explicit function of the form P = f(F), where P is the operating power, F is the traffic flow, and f(·) is the analytical expression or lookup table mapping relationship obtained through fitting. This function can be a linear function, a quadratic function, a piecewise linear function, or an S-curve function, the specific form of which is determined by the distribution characteristics of the measured data.

[0078] After obtaining the correspondence between flow rate and operating power, the average estimated flow rate obtained in step S13 can be substituted into the correspondence to obtain the estimated average operating power. Thus, through step S14, the conversion from flow rate dimension prediction to power dimension prediction can be realized, so that subsequent heat dissipation power regulation has a quantifiable and traceable input basis.

[0079] Step S15: Determine the optimal heat dissipation power of the forced air cooling system in a preset future time period based on the estimated average operating power, so as to control the operation of the forced air cooling system.

[0080] Step S15 can be achieved as follows: specifically, the estimated average operating power can be multiplied by a cooling efficiency coefficient to obtain the baseline cooling power. The cooling efficiency coefficient (CEC) is a parameter characterizing the proportional relationship between the cooling power required to maintain the stable operation of a unit of equipment operating power. Its physical meaning is the ratio between the cooling power required to maintain the stable operation of a high-power electronic device and the device's own operating power under standard environmental conditions (e.g., indoor temperature 25°C, relative humidity 50%, air pressure 101.3 kPa, no natural wind).

[0081] After obtaining the baseline heat dissipation power, a correction value for the heat dissipation power can be calculated based on the real-time environmental parameters of the large computing building. These real-time environmental parameters refer to the set of physical quantity data collected in real-time by multiple sensors deployed at key locations inside and outside the building and uploaded to the monitoring system at the moment this step is executed. These parameters can include at least one of the following: temperature, humidity, air pressure, and natural wind speed inside and outside the building. Specifically, temperature can be the measured temperature value of the air intake area, air outlet area, or surface of key equipment in the computer room; humidity can be the relative humidity of the air on the return air side of the computer room; air pressure can be the atmospheric static pressure at the building's altitude or the dynamic pressure difference within local ventilation channels; and natural wind speed can be the horizontal wind speed measured by wind speed sensors on the building's exterior facade, or the effective ventilation wind speed calculated by combining the building's orientation and wind direction angle. These environmental parameters collectively affect the thermal conductivity, convective heat transfer efficiency, and heat dissipation path resistance of the air medium, thereby altering the actual temperature rise control effect under the same heat dissipation power. In this application, the corresponding individual correction values ​​can be calculated by performing simulation calculations under standard operating conditions based on the degree to which each environmental parameter deviates from the standard operating conditions (e.g., indoor temperature 25℃, relative humidity 50%, air pressure 101.3kPa, no natural wind), and the individual correction values ​​can be summed to obtain the heat dissipation power correction value.

[0082] After obtaining the heat dissipation power correction value, the optimal heat dissipation power can be calculated using the correction value and the baseline heat dissipation power. For example, the sum of the two can be used as the optimal heat dissipation power. After obtaining the optimal heat dissipation power, it can be used to control the operation of the forced air cooling system.

[0083] It should be further explained that the heat dissipation efficiency coefficient can be obtained in the following way: Specifically, the high-power electronic device and the forced air cooling system can be deployed in an experimental unit. The experimental unit can be a physical test area independently divided in a large computing building, which has a complete power supply and distribution circuit, environmental monitoring capabilities and a controllable interface for the heat dissipation system. The experimental unit can reproduce the thermal coupling relationship under the real operating scenario, and its spatial dimensions, thermal parameters of the building envelope, and airflow organization method are consistent with or scaled proportionally to the actual building sub-area.

[0084] Then, the high-power electronic devices within the experimental unit are controlled to operate at at least two different operating power P. it Below, and at each operating power P it Adjust the heat dissipation power P of the forced air cooling system. cool This continues until the temperature at the key temperature measurement points within the experimental unit stabilizes within the safe target temperature range, thereby obtaining multiple power pairs (P... it P cool Among them, at least two different operating powers P itIt can be a discrete power point covering typical low load (e.g., 30% of rated power), medium load (e.g., 60% of rated power), and high load (e.g., 90% of rated power), which can control the high-power electronic devices in the experimental unit to operate at these discrete power points respectively, and synchronously adjust the heat dissipation power P of the forced air cooling system. cool This continues until the temperature at the key temperature measurement points within the experimental unit stabilizes within the safe target temperature range, thereby obtaining multiple power pairs (P... it P cool ).

[0085] After obtaining multiple power pairs (P) it P cool After that, the first method can be achieved through each power pair (P) it P cool P in ) cool With P it The first method calculates the heat dissipation efficiency coefficient corresponding to each of the multiple operating power ratios. The second method further calculates the average of each ratio as the final heat dissipation efficiency coefficient. Therefore, for the aforementioned method of multiplying the estimated average operating power by a heat dissipation efficiency coefficient to obtain the baseline heat dissipation power, we can first determine the heat dissipation efficiency coefficient corresponding to the estimated average operating power (in the case of heat dissipation efficiency coefficients corresponding to multiple operating powers). For example, we can obtain the heat dissipation efficiency coefficient corresponding to the estimated average operating power through interpolation, and then multiply the estimated average operating power by the heat dissipation efficiency coefficient to obtain the baseline heat dissipation power.

[0086] The intelligent monitoring method for building electrical systems provided in this application includes: acquiring historical monitoring flow rates of high-power electronic devices within a large computing building through a monitoring system; determining intraday flow rate variation curves and inter-day flow rate variation trends based on the historical monitoring flow rates; calculating the average estimated flow rate for a preset future time period based on the intraday flow rate variation curves, the inter-day flow rate variation trends, and the actual monitoring flow rate of the day; determining the estimated average operating power of the high-power electronic devices within the large computing building during the preset future time period based on the average estimated flow rate; and determining the optimal heat dissipation power of the forced air cooling system during the preset future time period based on the estimated average operating power, for use in controlling the operation of the forced air cooling system. This method obtains intraday flow rate variation curves and inter-day flow rate variation trends through historical monitoring flow rates, and then combines this with the actual monitoring flow rate of the day to calculate the average estimated flow rate for a preset future time period. This estimated flow rate method enables feedforward control of the forced air cooling system, which is more proactive, overcomes the delay defects of current feedback control strategies, and is more conducive to the long-term effective operation of high-power electronic devices.

[0087] It should be further explained that after obtaining the optimal heat dissipation power of the forced air cooling system in a preset future time period through the above-mentioned step S15, considering that the forced air cooling system includes multiple cooling fan modules, the distribution method of the optimal heat dissipation power among the various cooling fan modules can be further adopted in the following ways. Specifically, one method is to directly distribute the optimal heat dissipation power evenly among the various cooling fan modules. Another distribution method is to first obtain the operating parameters of each cooling fan module through the monitoring system, then determine the health assessment value of each cooling fan module based on the operating parameters, and then distribute the optimal heat dissipation power among the multiple cooling fan modules based on the health assessment value of each cooling fan module. The second power distribution method, because it further integrates the cooling fan modules, can further improve the service life of the cooling fan modules.

[0088] In practical applications, the operating parameters of the cooling fan module may include at least one of the following parameters: electrical parameters, performance parameters, reliability parameters, and physical state parameters.

[0089] Among them, electrical parameters can refer to operating parameters that characterize the electrical input and conversion characteristics of the cooling fan module, such as voltage, current, power factor, or input power. These electrical parameters are used to reflect whether the electrical operating state of the cooling fan module in the power supply circuit is within the rated tolerance range. Performance parameters can refer to operating parameters that characterize the aerodynamic output capability of the cooling fan module, such as speed, air volume, static pressure, or wind speed. These performance parameters are used to reflect the actual ability of the cooling fan module to complete the heat exchange task under the current operating conditions. Reliability parameters can refer to operating parameters that characterize the long-term operating stability and failure tendency of the cooling fan module, such as the cumulative number of start-stop cycles, the number of historical fault alarms, the mean time between failures (MTBF), or the distribution characteristics of the fault interval. These reliability parameters are used to reflect the failure risk level exhibited by the cooling fan module in the past operating cycle. Physical state parameters can refer to operating parameters that characterize the interaction state between the mechanical structure of the cooling fan module and the environment, such as vibration amplitude, noise intensity, bearing temperature, or shell surface temperature. These physical state parameters are used to reflect whether there are abnormal wear, loosening, or heat accumulation phenomena in the internal mechanical components of the cooling fan module.

[0090] In this application, determining the health assessment value of each cooling fan module based on the operating parameters can specifically include determining the health assessment value of each cooling fan module in the following manner: First, based on each parameter in the operating parameters, calculate the individual health score corresponding to each parameter; then, based on the individual health scores corresponding to each parameter in the operating parameters, determine the health assessment value of the cooling fan module by weighted summation.

[0091] The individual health score refers to a quantified value assigned to each specific parameter in the operating parameters, normalized to the interval [0, 1]. This individual health score can be a real number between 0.0 and 1.0. For example, for electrical parameters, this application can map the input power in the electrical parameters to an individual health score in the interval [0, 1] based on a preset threshold interval and membership function. Similarly, for performance parameters, the individual health score can be calculated based on the deviation ratio between the speed and its rated speed, combined with a nonlinear decay function. Furthermore, this application can also map the number of historical fault alarms in the reliability parameters to the interval [0, 1] after logarithmic transformation to generate an individual health score. This application obtains the individual health scores corresponding to each operating parameter based on any of the above methods. After obtaining the individual health scores corresponding to each operating parameter, the health assessment value of the cooling fan module can be determined by weighted summation.

[0092] Specifically, based on the health assessment value of each cooling fan module, the optimal heat dissipation power is allocated among multiple cooling fan modules. This can specifically include first dividing the multiple cooling fan modules into a main fan group and a backup fan group based on the health assessment value of each cooling fan module. For example, cooling fan modules with a health assessment value greater than a preset threshold can be assigned to the main fan group because of their high health, while cooling fan modules with a health assessment value less than or equal to the preset threshold can be assigned to the backup fan group because of their low health.

[0093] Then, the total rated power of each cooling fan module in the main fan group can be determined. For example, the rated power of each cooling fan module in the main fan group can be summed to obtain the total rated power. Then, it can be determined whether the total rated power is greater than or equal to the optimal cooling power.

[0094] If the total rated power is greater than or equal to the optimal heat dissipation power, it means that each cooling fan module in the main fan group is sufficient to support the optimal heat dissipation power. Therefore, the optimal heat dissipation power can be evenly distributed among the cooling fan modules in the main fan group, so that only each cooling fan module in the main fan group needs to be used.

[0095] Of course, if the total rated power is less than the optimal heat dissipation power, relying solely on the individual cooling fan modules in the main fan group is insufficient to achieve the total rated power. Therefore, on the one hand, the individual cooling fan modules in the main fan group can be controlled to operate at the rated power, and the power difference between the optimal heat dissipation power and the total rated power can be calculated. This difference is the power that the total rated power is insufficient relative to the optimal heat dissipation power.

[0096] After obtaining the power difference, M cooling fan modules can be selected from the standby fan group based on this power difference, and the power difference can be evenly distributed among the selected M cooling fan modules. Here, M = [power difference / (rated power × reduction factor)] + 1, where [] represents the integer part, and the reduction factor is a parameter greater than 0 and less than 1, such as 0.9, 0.8, etc. In this way, each cooling fan module in the main fan group can operate at its total rated power, while each cooling fan module in the standby fan group can operate at a relatively lower power, thereby improving the overall lifespan of the cooling fan modules.

[0097] Based on the intelligent monitoring method for building electrical systems provided in the embodiments of this application, the embodiments of this application can also provide an intelligent monitoring system for building electrical systems. For any unclear points regarding the content of this system embodiment, please refer to the relevant content in the above method embodiments. Figure 4 The diagram shows the specific structure of the intelligent monitoring system 20 (hereinafter referred to as system 20) for the building's electrical system. System 20 includes: an acquisition unit 201, a determination unit 202, a calculation unit 203, a prediction unit 204, and a second determination unit 205, wherein: The acquisition unit 201 is used to acquire the historical monitoring traffic of high-power electronic devices in a large computing building through the monitoring system. The determining unit 202 is used to determine the intraday flow change curve and the interday flow change trend based on the historical monitored flow. The calculation unit 203 is used to calculate the average estimated flow for a preset future time period based on the intraday flow change curve, the interday flow change trend and the actual monitored flow on the day. The estimation unit 204 is used to determine the estimated average operating power of high-power electronic devices in the large computing building during the preset future time period based on the average estimated flow rate. The second determining unit 205 is used to determine the optimal heat dissipation power of the forced air cooling system in the preset future time period based on the estimated average operating power, so as to control the operation of the forced air cooling system. The system 20 provided in the embodiments of this application adopts the same inventive concept as the method provided in the embodiments of this application. Since the method can solve the problems in the prior art, the system 20 can also solve the problems in the prior art. This will not be elaborated here.

[0098] The forced air cooling system includes multiple cooling fan modules; the system 20 also includes a power distribution unit, which is used to acquire the operating parameters of each cooling fan module through the monitoring system; determine the health assessment value of each cooling fan module based on the operating parameters; and distribute the optimal heat dissipation power among the multiple cooling fan modules based on the health assessment value of each cooling fan module.

[0099] The operating parameters include at least one of the following parameters of the cooling fan module: electrical parameters, performance parameters, reliability parameters, and physical state parameters; and, Based on the operating parameters, the health assessment value of each cooling fan module is determined. Specifically, the health assessment value of each cooling fan module is determined in the following manner: Based on each parameter in the operating parameters, calculate the individual health score corresponding to each parameter; The health assessment value of the cooling fan module is determined by weighted summation based on the individual health scores corresponding to each parameter in the operating parameters.

[0100] Specifically, based on the health assessment values ​​of each of the cooling fan modules, the optimal heat dissipation power is allocated among the multiple cooling fan modules, including: Based on the health assessment value of each cooling fan module, the multiple cooling fan modules are divided into a main fan group and a backup fan group. Determine the total rated power of each cooling fan module in the main fan group, and determine whether the total rated power is greater than or equal to the optimal cooling power; If the total rated power is greater than or equal to the optimal heat dissipation power, then the optimal heat dissipation power is evenly distributed among the heat dissipation fan modules of the main fan group; or, If the total rated power is less than the optimal heat dissipation power, then each heat dissipation fan module in the main fan group is controlled to operate at rated power, and the power difference between the optimal heat dissipation power and the total rated power is calculated. Based on the power difference, N heat dissipation fan modules are selected from the standby fan group, and the power difference is evenly distributed among the selected M heat dissipation fan modules, where M = [power difference / (rated power × reduction factor)] + 1.

[0101] Specifically, based on the historical monitored flow rates, the intraday flow rate variation curve and interday flow rate variation trend are determined, including: S121. Arrange the original sequence of the historical monitored flow in time sequence according to a preset time granularity; S122. Based on the fluctuation amplitude and periodic characteristics of historical monitoring flow after time-series arrangement, determine the optimal smoothing coefficient for extracting long-term trends. S123. Using the exponentially weighted moving average algorithm and applying the optimal smoothing coefficient, iterative calculations are performed on the historical monitoring flow after time-series arrangement to obtain an initial diurnal variation trend sequence characterizing the long-term change direction. S124. Remove the initial diurnal variation trend sequence from the original sequence of the historical monitored flow to obtain the first residual sequence; S125. Perform a stationarity test on the first residual sequence. If the test passes, use the initial diurnal variation trend sequence as the diurnal flow variation trend. If the test fails, adjust the optimal smoothing coefficient and re-execute steps S123 and S124 until the new first residual sequence passes the stationarity test, and then obtain the diurnal flow variation trend. S126. Remove the daytime flow change trend from the original sequence of the historical monitored flow to obtain the second residual sequence; S127. Based on the second residual sequence, calculate the historical flow statistics at the same time points within a single natural day, wherein the historical flow statistics include the arithmetic mean, median, or mode. S128. Connect the historical flow statistics at each time point to form the intraday flow change curve that characterizes the periodic fluctuation of intraday flow.

[0102] Specifically, based on the intraday flow rate change curve, the inter-day flow rate change trend, and the actual monitored flow rate for the day, the average estimated flow rate for a preset future time period is calculated, including: Based on the daytime flow variation trend, the theoretical total flow for the day is predicted using an extrapolation algorithm; Based on the intraday flow change curve, calculate the area under the first curve from midnight to the current time, the area under the second curve for a preset future time period, and the area under the third curve for the remaining time of the day. The theoretical flow rate for the day is determined based on the area under the first curve and the area under the third curve, as well as the theoretical total flow rate for the day. Based on the area under the second curve, the area under the third curve, and the area under the first curve, as well as the theoretical total flow for the day, estimate the theoretical total flow for the preset future time period; The ratio of the actual monitored flow rate to the theoretical flow rate of the day is used as a real-time correction factor. Multiply the theoretical total flow of the preset future time period by the real-time correction factor, and then divide by the duration of the preset future time period to obtain the average estimated flow of the preset future time period.

[0103] Specifically, determining the estimated average operating power of high-power electronic devices within the large computing building during the preset future time period based on the average estimated traffic flow includes: Multiple constant traffic loads Fi at different levels are applied to the high-power electronic device, and the operating power Pi of the high-power electronic device is measured simultaneously to obtain a series of data pairs (Fi, Pi); Regression analysis was performed on the series of data pairs (Fi, Pi) to obtain the corresponding relationship between flow rate and operating power. Substitute the average estimated flow rate into the corresponding relationship to obtain the estimated average operating power.

[0104] Specifically, determining the optimal heat dissipation power of the forced air cooling system within the preset future time period based on the estimated average operating power includes: Multiply the estimated average operating power by a heat dissipation efficiency coefficient to obtain the baseline heat dissipation power; Based on the real-time environmental parameters of the large computing building, a heat dissipation power correction value is calculated; the environmental parameters include at least one of temperature, humidity, air pressure and natural wind speed inside and outside the building. The optimal heat dissipation power is calculated using the heat dissipation power correction value and the baseline heat dissipation power.

[0105] The system 20 further includes a heat dissipation efficiency coefficient determination unit, used to deploy the high-power electronic equipment and the forced air cooling system within an experimental unit; and to control the high-power electronic equipment within the experimental unit to operate at at least two different operating power P. it Below, and at each operating power P it Next, adjust the heat dissipation power P of the forced air cooling system. cool This continues until the temperature at the key temperature measurement points within the experimental unit stabilizes within the safe target temperature range, in order to obtain multiple power pairs (P... it P cool ); through each power pair (P) it P cool P in ) cool With P it The ratio of the values ​​is used to calculate the heat dissipation efficiency coefficients corresponding to the multiple operating powers.

[0106] Figure 5An example is a schematic diagram of the physical structure of an electronic device, such as... Figure 5 As shown, the electronic device may include a processor 310, a communications interface 320, a memory 330, and a communication bus 340, wherein the processor 310, communications interface 320, and memory 330 communicate with each other via the communication bus 340. The processor 310 can call logical instructions in the memory 330 to execute the intelligent monitoring method for a building electrical system provided in this application embodiment. This method includes: acquiring historical monitoring traffic flow of high-power electronic devices within a large computing building through a monitoring system; determining a daily traffic flow change curve and a daytime traffic flow change trend based on the historical monitoring traffic flow; calculating the average estimated traffic flow for a preset future time period based on the daily traffic flow change curve, the daytime traffic flow change trend, and the actual monitoring traffic flow for the day; determining the estimated average operating power of the high-power electronic devices within the large computing building during the preset future time period based on the average estimated traffic flow; and determining the optimal heat dissipation power of the forced air cooling system during the preset future time period based on the estimated average operating power, for controlling the operation of the forced air cooling system. This method obtains the daily flow change curve and the daytime flow change trend by monitoring historical flow, and then combines the actual monitored flow on the day to estimate the average flow for a preset future time period. By using this estimated flow, the forced air cooling system can be fed forward controlled. This control method is more proactive, can overcome the delay defects caused by the current feedback control strategy, and is more conducive to the long-term effective operation of high-power electronic devices.

[0107] Obviously, since the processor 310 can call the logical instructions in the memory 330 to execute the method provided in the embodiments of this application, it can also solve the problems in the prior art.

[0108] Furthermore, the logical instructions in the aforementioned memory 330 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0109] On the other hand, the present invention also provides a computer program product, which includes a computer program that can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the intelligent monitoring method for building electrical systems provided in the embodiments of this application. This method includes: acquiring historical monitoring flow of high-power electronic devices in a large computing building through a monitoring system; determining a daily flow change curve and a daytime flow change trend based on the historical monitoring flow; calculating the average estimated flow for a preset future time period based on the daily flow change curve, the daytime flow change trend, and the actual monitoring flow for the day; determining the estimated average operating power of the high-power electronic devices in the large computing building during the preset future time period based on the average estimated flow; and determining the optimal heat dissipation power of the forced air cooling system during the preset future time period based on the estimated average operating power, for controlling the operation of the forced air cooling system. This method obtains the daily flow change curve and the daytime flow change trend by monitoring historical flow, and then combines the actual monitored flow on the day to estimate the average flow for a preset future time period. By using this estimated flow, the forced air cooling system can be fed forward controlled. This control method is more proactive, can overcome the delay defects caused by the current feedback control strategy, and is more conducive to the long-term effective operation of high-power electronic devices.

[0110] Obviously, since the computer can execute the method provided in the embodiments of this application when the computer program is executed by the processor, it can also solve the problems in the prior art.

[0111] In another aspect, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, provides the method provided in the embodiments of this application.

[0112] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0113] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0114] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for intelligent monitoring of a building electrical system, characterized in that, include: Historical monitoring traffic of high-power electronic devices within large computing power buildings is obtained through a monitoring system. Based on the historical monitored flow rates, determine the intraday flow rate change curve and the interday flow rate change trend; Based on the intraday flow change curve, the interday flow change trend, and the actual monitored flow on the day, the average estimated flow for a preset future time period is calculated. Based on the average estimated traffic, the estimated average operating power of high-power electronic devices in the large computing building is determined in the preset future time period. The optimal heat dissipation power of the forced air cooling system in the preset future time period is determined based on the estimated average operating power, so as to control the operation of the forced air cooling system.

2. The method according to claim 1, characterized in that, The forced air cooling system includes multiple cooling fan modules; the method further includes: The operating parameters of each of the cooling fan modules are obtained through the monitoring system. Based on the operating parameters, determine the health assessment value of each of the cooling fan modules; Based on the health assessment value of each of the cooling fan modules, the optimal heat dissipation power is allocated among the multiple cooling fan modules.

3. The method according to claim 2, characterized in that, The operating parameters include at least one of the following parameters of the cooling fan module: electrical parameters, performance parameters, reliability parameters, and physical state parameters; and, Based on the operating parameters, the health assessment value of each cooling fan module is determined. Specifically, the health assessment value of each cooling fan module is determined in the following manner: Based on each parameter in the operating parameters, calculate the individual health score corresponding to each parameter; The health assessment value of the cooling fan module is determined by weighted summation based on the individual health scores corresponding to each parameter in the operating parameters.

4. The method according to claim 2, characterized in that, Based on the health assessment values ​​of each of the aforementioned cooling fan modules, the optimal heat dissipation power is allocated among the multiple cooling fan modules, specifically including: Based on the health assessment value of each cooling fan module, the multiple cooling fan modules are divided into a main fan group and a backup fan group. Determine the total rated power of each cooling fan module in the main fan group, and determine whether the total rated power is greater than or equal to the optimal cooling power; If the total rated power is greater than or equal to the optimal heat dissipation power, then the optimal heat dissipation power is evenly distributed among the heat dissipation fan modules of the main fan group; or, If the total rated power is less than the optimal heat dissipation power, then each heat dissipation fan module in the main fan group is controlled to operate at rated power, and the power difference between the optimal heat dissipation power and the total rated power is calculated. Based on the power difference, N heat dissipation fan modules are selected from the standby fan group, and the power difference is evenly distributed among the selected M heat dissipation fan modules, where M = [power difference / (rated power × reduction factor)] + 1.

5. The method according to claim 1, characterized in that, Based on the historical monitored flow rates, the intraday flow rate variation curve and interday flow rate variation trend are determined, specifically including: S121. Arrange the original sequence of the historical monitored flow in time sequence according to a preset time granularity; S122. Based on the fluctuation amplitude and periodic characteristics of historical monitoring flow after time-series arrangement, determine the optimal smoothing coefficient for extracting long-term trends. S123. Using the exponentially weighted moving average algorithm and applying the optimal smoothing coefficient, iterative calculations are performed on the historical monitoring flow after time-series arrangement to obtain an initial diurnal variation trend sequence characterizing the long-term change direction. S124. Remove the initial diurnal variation trend sequence from the original sequence of the historical monitored flow to obtain the first residual sequence; S125. Perform a stationarity test on the first residual sequence. If the test passes, use the initial diurnal variation trend sequence as the diurnal flow variation trend. If the test fails, adjust the optimal smoothing coefficient and re-execute steps S123 and S124 until the new first residual sequence passes the stationarity test, and then obtain the diurnal flow variation trend. S126. Remove the daytime flow change trend from the original sequence of the historical monitored flow to obtain the second residual sequence; S127. Based on the second residual sequence, calculate the historical flow statistics at the same time points within a single natural day, wherein the historical flow statistics include the arithmetic mean, median, or mode. S128. Connect the historical flow statistics at each time point to form the intraday flow change curve that characterizes the periodic fluctuation of intraday flow.

6. The method according to claim 1, characterized in that, Based on the intraday flow rate change curve, the inter-day flow rate change trend, and the actual monitored flow rate for the day, the average estimated flow rate for a preset future time period is calculated, specifically including: Based on the daytime flow variation trend, the theoretical total flow for the day is predicted using an extrapolation algorithm; Based on the intraday flow change curve, calculate the area under the first curve from midnight to the current time, the area under the second curve for a preset future time period, and the area under the third curve for the remaining time of the day. The theoretical flow rate for the day is determined based on the area under the first curve and the area under the third curve, as well as the theoretical total flow rate for the day. Based on the area under the second curve, the area under the third curve, and the area under the first curve, as well as the theoretical total flow for the day, estimate the theoretical total flow for the preset future time period; The ratio of the actual monitored flow rate to the theoretical flow rate of the day is used as a real-time correction factor. Multiply the theoretical total flow of the preset future time period by the real-time correction factor, and then divide by the duration of the preset future time period to obtain the average estimated flow of the preset future time period.

7. The method according to claim 1, characterized in that, Based on the average estimated traffic, the estimated average operating power of high-power electronic devices within the large computing building is determined during the preset future time period, specifically including: Multiple constant service traffic loads Fi at different levels are applied to the high-power electronic device, and the operating power Pi of the high-power electronic device is measured simultaneously to obtain a series of data pairs (Fi, Pi); Regression analysis was performed on the series of data pairs (Fi, Pi) to obtain the corresponding relationship between flow rate and operating power. Substitute the average estimated flow rate into the corresponding relationship to obtain the estimated average operating power.

8. The method according to claim 1, characterized in that, Determining the optimal heat dissipation power of the forced air cooling system in the preset future time period based on the estimated average operating power specifically includes: Multiply the estimated average operating power by a heat dissipation efficiency coefficient to obtain the baseline heat dissipation power; Based on the real-time environmental parameters of the large computing building, a heat dissipation power correction value is calculated; the environmental parameters include at least one of temperature, humidity, air pressure and natural wind speed inside and outside the building. The optimal heat dissipation power is calculated using the heat dissipation power correction value and the baseline heat dissipation power.

9. The method according to claim 1, characterized in that, The method further includes: The high-power electronic device and the forced air cooling system are deployed in an experimental unit. Control the high-power electronic devices within the experimental unit to operate at at least two different operating power P it Below, and at each operating power P it Next, adjust the heat dissipation power P of the forced air cooling system. cool This continues until the temperature at the key temperature measurement points within the experimental unit stabilizes within the safe target temperature range, in order to obtain multiple power pairs (P... it P cool ); Through each power pair (P) it P cool P in ) cool With P it The ratio of the values ​​is used to calculate the heat dissipation efficiency coefficients corresponding to the multiple operating powers.

10. An intelligent monitoring system for a building electrical system, characterized in that, include: The acquisition unit is used to acquire historical monitoring traffic of high-power electronic devices in large computing buildings through the monitoring system. The determining unit is used to determine the intraday flow change curve and the interday flow change trend based on the historical monitored flow. The calculation unit is used to calculate the average estimated flow for a preset future time period based on the intraday flow change curve, the interday flow change trend and the actual monitored flow on the day. The estimation unit is used to determine the estimated average operating power of high-power electronic devices in the large computing building during the preset future time period based on the average estimated traffic. The second determining unit is used to determine the optimal heat dissipation power of the forced air cooling system in the preset future time period based on the estimated average operating power, so as to control the operation of the forced air cooling system.

Images