Intelligent monitoring method for data center bus duct operating environment

Abstract

The present application relates to the technical field of data center bus duct temperature monitoring, in particular to a kind of data center bus duct operating environment intelligent monitoring method.The method first monitors temperature by laying distributed temperature measurement optical fiber on the surface of large bus, locates high temperature section;Call four-channel infrared temperature sensor data at the connector joint of high temperature section, identify neutral line abnormal high temperature;Retrieve main and branch harmonic data, evaluate whether harmonic is over standard by weighted calculation, and determine whether neutral line high temperature is caused by harmonic.If it is determined that harmonic causes, then use phase comparison method to lock the main harmonic source branch and its contribution ratio, and analyze the harmonic spectrum characteristics of the branch to identify the abnormal type, and generate automatic governance strategy, and then block the generation of heat source from the source.The present application realizes the depth perception of bus duct operating state, root cause diagnosis and active intervention, and improves the reliability and operation intelligent level of data center power supply system.

Description

Technical Field

[0001] This invention relates to the field of data center busbar temperature monitoring technology, and more specifically, to a method for intelligent monitoring of the operating environment of data center busbars. Background Technology

[0002] Data centers are the core infrastructure of the modern digital economy, and the reliability of their power supply systems directly affects the security and continuity of data services. Busbars, as the "arteries" of power distribution in data centers, are of paramount importance in their operational status. Traditional monitoring methods typically monitor busbar temperature or power quality (such as harmonics) in isolation, lacking in-depth analysis of the correlation between multiple parameters. For example, when the neutral line of a busbar overheats due to harmonics, existing systems often only issue temperature exceedance alarms, failing to automatically diagnose that the root cause of the overheating is excessive harmonics, let alone pinpoint the specific harmonic source branch (such as which server racks are located). Maintenance personnel must carry specialized equipment for on-site troubleshooting, which is inefficient and cannot prevent faults. Therefore, there is an urgent need for a monitoring method capable of intelligent correlation analysis, accurate root cause location, and automatic execution of remediation strategies to improve the predictive maintenance capabilities and overall resilience of data center power supply systems. Summary of the Invention

[0003] The purpose of this invention is to provide an intelligent monitoring method for the operating environment of data center busbar trunking, so as to solve the above-mentioned technical problems.

[0004] To achieve the above objectives, the embodiments of this application provide the following technical solutions:

[0005] This application provides an intelligent monitoring method for the operating environment of a data center busbar trunking. The method includes: real-time monitoring of the temperature of the main busbar in the data center busbar trunking using distributed temperature-measuring optical fibers installed on the surface of the main busbar; identifying multiple high-temperature busbar trunking segments when local high temperatures occur on the main busbar using the distributed temperature-measuring optical fibers; retrieving temperature data from a first temperature and humidity sensor installed on a first connector joint, the first connector joint being used to connect multiple adjacent high-temperature busbar trunking segments; the temperature monitoring module in the first temperature and humidity sensor is a four-channel infrared temperature sensor, with the four channels monitoring the temperatures of three phase lines and one neutral line, respectively; retrieving first harmonic data recorded by a first intelligent power measurement and control instrument installed on the transformer at the beginning of the main trunk of the busbar trunking when any neutral line temperature exceeds a preset first temperature threshold; calculating the current harmonic evaluation index based on the total harmonic current distortion rate and the third harmonic current content rate in the first harmonic data using a weighted algorithm; and determining that the main cause of the abnormal high temperature of the neutral line is harmonics when the harmonic evaluation index exceeds a preset index threshold; and generating a corresponding automatic mitigation strategy based on the first harmonic data to reduce the neutral line temperature.

[0006] Optionally, the step of generating a corresponding automatic governance strategy based on the first harmonic data includes:

[0007] Based on the first harmonic data and the second harmonic data fed back by the second intelligent power measurement and control instrument located in the branch plug-in box and at the front end of the important load, the main harmonic sources and the contribution ratio of each main harmonic source are determined by the phase comparison method, and the branches or important loads with the highest contribution ratios are recorded as the first branch.

[0008] The harmonic spectrum characteristics of each first branch are analyzed sequentially and matched with the built-in device harmonic fingerprint database to obtain an automatic governance strategy for each first branch.

[0009] Optionally, the harmonic spectrum characteristics of the first branch are analyzed, and a corresponding automatic mitigation strategy is formulated, including:

[0010] Retrieve the current waveform data of the first branch in the first 15-20 minutes, perform Fourier transform on it to generate a harmonic spectrum, and extract the core features of the harmonic spectrum, including the total harmonic current distortion rate, the content rate of each harmonic and the characteristic ratio.

[0011] If the core features are compared with the built-in device harmonic fingerprint database, and the 3rd harmonic content accounts for 70% to 85% of the total harmonic content, and the 5th and 7th harmonic content is low and shows a monotonically decreasing trend, then it is determined that a large number of downstream servers of the first branch start up simultaneously, resulting in pulse injection of the 3rd harmonic. The monotonically decreasing trend means that the content of each odd harmonic current decreases significantly with the increase of the harmonic number.

[0012] The system retrieves the operation and maintenance calendar of the downstream server cluster of the first branch to determine whether there are still batch power-on operations on the downstream servers of the first branch. If so, it adjusts the startup scheduling parameters of the downstream server cluster of the first branch so that the multiple servers to be started can perform power-on operations in batches according to the adjusted startup scheduling parameters, thereby reducing the concentration of batch power-on of servers.

[0013] The beneficial effects of this invention are as follows:

[0014] The method described in this invention deeply integrates distributed fiber optic temperature measurement, infrared temperature measurement from bus trunking connections, and full-network harmonic monitoring data to automatically determine whether abnormal high temperatures in the neutral line are caused by excessive harmonics. It then uses phase comparison and spectral feature recognition technology to accurately locate the specific branch causing the excessive harmonics and the abnormal cause type of downstream equipment. Based on the abnormal cause type, it formulates corresponding automatic governance strategies to eliminate the cause of abnormal high temperatures at the root, thereby improving the reliability and intelligent operation and maintenance level of the data center power supply system.

[0015] Secondly, the method described in this embodiment is not limited to monitoring and alarming; more importantly, it can automatically generate and execute governance strategies based on diagnostic results. In particular, for harmonic impacts caused by the batch startup of IT equipment (such as servers), this method can implement flexible power-on scheduling of unpowered devices in batches and with a time sequence by linking with intelligent PDUs or out-of-band management interfaces, smoothing out current impacts from the source, realizing intelligent coordination between infrastructure and IT load, and effectively avoiding the risks of excessive harmonics and overheating.

[0016] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description and the accompanying drawings. Attached Figure Description

[0017] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0018] Figure 1 This is a schematic diagram of a data center busbar connection system structure as described in an embodiment of the present invention;

[0019] Figure 2 This is a schematic diagram of a method for intelligent monitoring of the operating environment of a data center busbar trunking, as described in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0021] Example:

[0022] Before explaining the principle of the intelligent monitoring method for the operating environment of data center busbars described in this embodiment, it is necessary to briefly describe the structure of data center busbars, such as... Figure 1 As shown, the first end of the busbar trunking is connected to a transformer box or transformer, and multiple busbar trunking sections are connected by connectors. The connectors are equipped with infrared temperature and humidity sensors, with the temperature sensors being four-channel and compatible with three-phase four-wire power supply lines (N / C / B / A). Multiple wire branches are connected to the busbar trunking through branch plug-in boxes. A server cluster is connected downstream of the branch plug-in boxes. The busbar trunking contains a large busbar, and the surface of the large busbar is covered with distributed temperature measurement fiber optic (DTS). Furthermore, the first end of the busbar trunking and multiple branch plug-in boxes are equipped with intelligent power measurement and control instruments. These instruments can not only measure conventional voltage, current, and power, but also perform power quality analysis and calculate the content (THD) and amplitude of each harmonic (such as the 3rd, 5th, and 7th harmonics) in real time.

[0023] like Figure 2 As shown in the figure, this embodiment provides an intelligent monitoring method for the operating environment of a data center busbar trunking, the method including steps S100, S200 and S300.

[0024] Step S100: Monitor the temperature of the large busbar in the data center busbar trunking in real time through distributed temperature-measuring optical fibers installed on the surface of the large busbar. When the large busbar experiences local high temperature (greater than a preset first temperature threshold, which is low and mainly used to monitor slight abnormal high temperatures of the large busbar, but long-term slight abnormal high temperatures will affect the design service life of the large busbar), identify the corresponding multiple high-temperature busbar trunking segments through distributed temperature-measuring optical fibers.

[0025] Step S200: Retrieve the temperature data of the first temperature and humidity sensor installed on the first connector joint. The first connector joint is used to connect multiple adjacent high-temperature busbar sections. The temperature monitoring module in the first temperature and humidity sensor is a four-channel infrared temperature sensor. The four channels monitor the temperatures of three phase lines and one neutral line, respectively.

[0026] Step S300: When any neutral line temperature exceeds a preset first temperature threshold, retrieve the first harmonic data recorded by the first intelligent power measurement and control instrument on the transformer at the beginning of the main trunk of the busbar. Based on the total harmonic current distortion rate and the third harmonic current content rate in the first harmonic data, calculate the current harmonic evaluation index using a weighted algorithm. If the harmonic evaluation index exceeds a preset index threshold, determine that the main cause of the abnormal high temperature of the neutral line is harmonics. Based on the first harmonic data, generate a corresponding automatic mitigation strategy to reduce the neutral line temperature.

[0027] Many factors can induce high temperatures on the main busbar. Therefore, Distributed Temperature Sensing Fiber (DTS) is only used as a trigger for high temperatures. For some inconspicuous long-path abnormal high temperatures, its monitoring data is not very reliable. Only when obvious high temperatures occur can DTS quickly pinpoint the location of significant high temperatures. Therefore, when inconspicuous long-path abnormal high temperatures occur, infrared temperature and humidity sensors installed on the connectors of the high-temperature section can further determine which line is experiencing high-temperature overload. In real-world operation and maintenance environments, harmonics (especially the 3rd harmonic) are the main cause of abnormal heating of the neutral line. In a three-phase four-wire (A / B / C / N) system, the neutral line (N line) is the current return path. Under an ideal three-phase balanced linear load, the vector sum of the three-phase currents on the neutral line is zero, and there is almost no current on the N line. Modern data center loads (such as server and switch power supplies) are mainly switching power supplies, which generate a large amount of 3rd harmonic current. The key point is that the 3rd harmonic currents of the three phases are in phase in time. Therefore, they do not cancel each other out, but are directly algebraically added on the neutral line. This can cause the neutral current to be even greater than the phase current. According to the heating formula (heat ∝ current² × resistance), an excessively large neutral current will cause its temperature rise to far exceed expectations, making it one of the hottest conductors in a large busbar system, posing a risk of insulation aging and even fire. Therefore, monitoring the neutral temperature is a direct and effective indicator for early warning of harmonic hazards.

[0028] The automatic governance strategy generated based on the first harmonic data includes:

[0029] Step S310: Based on the first harmonic data and the second harmonic data fed back by the second intelligent power measurement and control instrument located in the branch plug-in box and at the front end of the important load, the main harmonic sources and the contribution ratio of each main harmonic source are determined by the phase comparison method, and the branches or important loads with the highest contribution ratios are recorded as the first branch.

[0030] The following is a brief explanation of the implementation principle of the phase comparison method for determining the main harmonic sources and the contribution ratio of each main harmonic source:

[0031] Acquire the current waveforms collected by the first intelligent power meter and controller and multiple second intelligent power meter and controllers, analyze the waveforms, extract the magnitude and phase angle of the specified harmonic (such as the third harmonic), compare the phase angle of the harmonic current measured at the branch point (corresponding to the second intelligent power meter and controller) with that at the main point (corresponding to the first intelligent power meter and controller), if the phase of the branch point leads that of the main point, it is determined that the harmonic is injected into the bus from the branch, and the branch is the harmonic source;

[0032] The total harmonic current measured at the main trunk point is considered as the vector sum of the currents injected by each harmonic source branch. Then, based on the magnitude and phase of the harmonic current in each source branch, its contribution weight to the total current is calculated (achieved by solving a non-negative least squares optimization problem). The contribution percentage of each harmonic source branch is output, and the branches with larger contributions are identified as the main sources. This method uses the synchronous harmonic current measurements (amplitude and phase) of all harmonic source branches and the main trunk point to find an optimal weight allocation scheme, ensuring that the current synthesized by each branch according to this weight is closest to the measured current of the main trunk. The final normalized weight is the contribution ratio. A specific implementation could be as follows:

[0033] The amplitude and phase angle of the target subharmonic current (target subharmonic current vector) are obtained from the main monitoring point, and the amplitude and phase angle of the corresponding subharmonic current (subharmonic current vector) are obtained from each harmonic source branch.

[0034] Calculate the target subharmonic current vector and the horizontal and vertical components of the subharmonic current vector, and construct the vector linear equations: Main trunk horizontal component ≈ (Branch 1 horizontal component × weight 1) + (Branch 2 horizontal component × weight 2) + ... + (Branch k horizontal component × weight k); Main trunk vertical component ≈ (Branch 1 vertical component × weight 1) + (Branch 2 vertical component × weight 2) + ... + (Branch k vertical component × weight k); where K is the number of harmonic source branches;

[0035] Due to measurement errors, the above equation is usually not strictly valid. The system uses a least squares optimization algorithm to solve the problem. Its goal is to find a set of non-negative contribution coefficients (weights) such that the overall error between the composite vector calculated from these weights and the components of each branch and the measured vector of the main trunk is minimized. Then, the set of non-negative contribution coefficients obtained are normalized to obtain the contribution percentage corresponding to each harmonic source branch.

[0036] Step S320: Analyze the harmonic spectrum characteristics of each first branch in sequence and match them with the built-in device harmonic fingerprint database to obtain an automatic governance strategy for each first branch.

[0037] The specific implementation method for analyzing the harmonic spectrum characteristics of the first branch and formulating the corresponding automatic mitigation strategy mentioned in step S320 can be as follows:

[0038] Step S321: Retrieve the current waveform data of the first branch in the first 15-20 minutes, perform Fourier transform on it to generate a harmonic spectrum diagram, and extract the core features of the harmonic spectrum diagram. The core features include the total harmonic current distortion rate, the content rate of each harmonic, and the characteristic ratio.

[0039] Step S322: The core feature is compared with the built-in device harmonic fingerprint database. If the 3rd harmonic content in the core feature accounts for 70% to 85% of the total harmonic content, and the 5th and 7th harmonic content is low and shows a monotonically decreasing trend, then it is determined that a large number of downstream servers of the first branch start up at the same time, resulting in pulse injection of the 3rd harmonic. The monotonically decreasing trend means that the content of each odd harmonic current decreases significantly with the increase of the harmonic number.

[0040] Step S323: Retrieve the operation and maintenance calendar of the downstream server cluster of the first branch, and determine whether there are still batch power-on operations of the downstream servers of the first branch. If so, adjust the startup scheduling parameters of the downstream server cluster of the first branch so that the multiple servers to be started can perform power-on operations in batches according to the adjusted startup scheduling parameters, thereby reducing the concentration of batch power-on of servers.

[0041] For example, if the maintenance calendar of the downstream server cluster in the first branch determines that there are 20 servers waiting to be powered on, then the above-mentioned method of adjusting the startup scheduling parameters to reduce the concentration of server batch startup can be implemented as follows: the startup operation of the 20 target servers, which may have been divided into 4 batches and completed within 15 minutes, can be adjusted to: within 18 minutes, the 20 target servers are divided into 5 batches of 4 servers each, with a batch interval of 240 seconds, and remote power-on is performed.

[0042] The overall optimization direction is to break down the total task into more batches within a reasonable range, reduce the number of servers started simultaneously in each batch, extend the batch interval, ensure sufficient interval time between adjacent batches, allow the harmonic current generated by the previous batch to be fully attenuated, and appropriately extend the total allowable time to complete all boot operations.

[0043] The method described in this embodiment deeply integrates distributed fiber optic temperature measurement, connection point infrared temperature measurement, and full network harmonic monitoring data. This method can automatically determine whether the abnormal high temperature of the neutral line is caused by excessive harmonics. By using phase comparison method and spectrum feature recognition technology, it can accurately locate the specific branch and equipment type that causes the excessive harmonics, upgrading operation and maintenance from "phenomenon alarm" to "root cause diagnosis", which greatly improves the efficiency of troubleshooting.

[0044] Secondly, the method described in this embodiment is not limited to monitoring and alarming; more importantly, it can automatically generate and execute governance strategies based on diagnostic results. In particular, for harmonic impacts caused by the batch startup of IT equipment (such as servers), this method can implement flexible power-on scheduling of unpowered devices in batches and with a time sequence by linking with intelligent PDUs or out-of-band management interfaces, smoothing out current impacts from the source, realizing intelligent coordination between infrastructure and IT load, and effectively avoiding the risks of excessive harmonics and overheating.

[0045] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A method for intelligent monitoring of the operating environment of a data center busbar trunking, characterized in that, The method includes: The temperature of the main bus in the data center busbar trunking is monitored in real time by distributed temperature-measuring optical fibers installed on the surface of the main busbar. When local high temperature occurs on the main busbar, the corresponding multiple high-temperature busbar trunking segments are identified by the distributed temperature-measuring optical fibers. The temperature data of the first temperature and humidity sensor located on the first connector is retrieved. The first connector is used to connect multiple adjacent high-temperature busbar sections. The temperature monitoring module in the first temperature and humidity sensor is a four-channel infrared temperature sensor. The four channels monitor the temperatures of three phase lines and one neutral line, respectively. If any neutral line temperature exceeds a preset first temperature threshold, the first harmonic data recorded by the first intelligent power measurement and control instrument on the transformer at the beginning of the main trunk of the busbar is retrieved. Based on the total harmonic current distortion rate and the third harmonic current content rate in the first harmonic data, the current harmonic evaluation index is calculated by a weighted algorithm. If the harmonic evaluation index exceeds a preset index threshold, the main cause of the abnormal high temperature of the neutral line is determined to be harmonics. Based on the first harmonic data, a corresponding automatic treatment strategy is generated to reduce the neutral line temperature. The automatic governance strategy generated based on the first harmonic data includes: Based on the first harmonic data and the second harmonic data fed back by the second intelligent power measurement and control instrument located in the branch plug-in box and at the front end of the important load, the main harmonic sources and the contribution ratio of each main harmonic source are determined by the phase comparison method, and the branches or important loads with the highest contribution ratios are recorded as the first branch. The harmonic spectrum characteristics of each first branch are analyzed sequentially and matched with the built-in device harmonic fingerprint database to obtain an automatic governance strategy for each first branch. The method for determining the main harmonic sources and the contribution ratio of each main harmonic source using the phase comparison method is as follows: Acquire the current waveforms collected by the first intelligent power meter and controller and multiple second intelligent power meter and controllers, analyze the waveforms, extract the magnitude and phase angle of the specified harmonic, compare the phase angle of the harmonic current measured at the branch point and the main point, if the phase of the branch point leads that of the main point, it is determined that the harmonic is injected into the bus from the branch, and the branch is the harmonic source. The total harmonic current measured at the main trunk point is considered as the vector sum of the currents injected by each harmonic source branch. Then, based on the magnitude and phase of the harmonic current in each source branch, a non-negative least squares optimization problem is used to calculate their contribution weights to the total current. The contribution percentage of each harmonic source branch is output, and the branches with larger contributions are identified as the main sources. Based on the amplitude and phase of the synchronous harmonic currents of all harmonic source branches and the main trunk point, an optimal weight allocation scheme is found, so that the current synthesized by each branch according to this weight is closest to the measured current of the main trunk. Finally, the normalized weight is the contribution ratio. The specific implementation is as follows: The amplitude and phase angle of the target subharmonic current are obtained from the main monitoring point, and the amplitude and phase angle of the corresponding subharmonic current are obtained from each harmonic source branch. Calculate the target subharmonic current vector and the horizontal and vertical components of the corresponding subharmonic current vector, and construct the vector linear equations: Main trunk horizontal component ≈ (Branch 1 horizontal component × weight 1) + (Branch 2 horizontal component × weight 2) + ... + (Branch k horizontal component × weight k); Main trunk vertical component ≈ (Branch 1 vertical component × weight 1) + (Branch 2 vertical component × weight 2) + ... + (Branch k vertical component × weight k); where k is the number of harmonic source branches; The system uses a least squares optimization algorithm to solve the problem. Its goal is to find a set of non-negative contribution coefficients such that the overall error between the composite vector calculated from these weights and the components of each branch and the measured vector of the main trunk is minimized. Then, the set of non-negative contribution coefficients is normalized to obtain the contribution percentage of each harmonic source branch.

2. The intelligent monitoring method for the operating environment of data center busbar trunking according to claim 1, characterized in that, Analyze the harmonic spectrum characteristics of the first branch and formulate corresponding automatic mitigation strategies, including: Retrieve the current waveform data of the first branch in the first 15-20 minutes, perform Fourier transform on it to generate a harmonic spectrum, and extract the core features of the harmonic spectrum, including the total harmonic current distortion rate, the content rate of each harmonic and the characteristic ratio. If the core features are compared with the built-in device harmonic fingerprint database, and the 3rd harmonic content accounts for 70% to 85% of the total harmonic content, and the 5th and 7th harmonic content is low and shows a monotonically decreasing trend, then it is determined that a large number of downstream servers of the first branch start up simultaneously, resulting in pulse injection of the 3rd harmonic. The monotonically decreasing trend means that the content of each odd harmonic current decreases significantly with the increase of the harmonic number. The system retrieves the operation and maintenance calendar of the downstream server cluster of the first branch to determine whether there are still batch power-on operations on the downstream servers of the first branch. If so, it adjusts the startup scheduling parameters of the downstream server cluster of the first branch so that the multiple servers to be started can perform power-on operations in batches according to the adjusted startup scheduling parameters, thereby reducing the concentration of batch power-on of servers.

Images