A UPS transformer voltage balance control method and system

Abstract

This application relates to the field of uninterruptible power supply (UPS) technology, and discloses a UPS transformer voltage balance control method and system. The method establishes a communication link between the UPS transformer and the power supply module of sensitive equipment in a data center; receives microscopic voltage quality data generated by the power supply module monitoring minute transient fluctuations in its input voltage; evaluates the impact of power supply quality on sensitive equipment based on the microscopic voltage quality data, obtaining an evaluation result; dynamically adjusts the target margin of the output voltage of each phase of the UPS transformer based on the evaluation result; and controls the output voltage of the UPS transformer based on the adjusted output voltage target margin, achieving refined and dynamic balance control of the UPS transformer output voltage. This effectively addresses minute transient voltage fluctuations, ensures the power supply quality of sensitive equipment, and solves the problem in existing technologies where minute voltage fluctuations are difficult to detect and handle.

Description

Technical Field

[0001] This application relates to the field of uninterruptible power supply (UPS) technology, and more specifically, to a UPS transformer voltage balance control method and system. Background Technology

[0002] In modern data center environments, uninterruptible power supply (UPS) systems are core infrastructure ensuring the continuous operation of critical information technology equipment. These UPS systems typically contain one or more transformers, and their voltage balancing control methods are crucial, designed to ensure the stability and symmetry of the output voltage, thereby providing high-quality power to downstream servers, storage arrays, and network equipment. A typical UPS transformer voltage balancing control system continuously acquires real-time output voltage data through voltage sampling sensors and, based on this feedback information, precisely adjusts the power output through internal control circuitry to maintain the voltage within a preset, narrow range.

[0003] Even in well-designed UPS systems, the delicate electronic components inside inevitably undergo physical changes during long-term, uninterrupted operation. For example, the precision resistors used in voltage sampling sensors experience minute Joule heating due to continuous current flow. Simultaneously, daily temperature fluctuations in data centers, even within non-extreme ranges, cause microscopic thermal expansion and contraction in these resistor materials. These seemingly insignificant physical effects, accumulated over years or even longer, can lead to extremely subtle changes in the material structure of the resistors, such as an increase in lattice defects or a redistribution of internal stress. This causes the resistance value to gradually deviate from its initial calibration value, resulting in what is known as "resistance drift." This drift is not a sudden failure but a slow, gradual, and imperceptible change in physical properties. This minute resistance drift, in turn, causes a continuous, imperceptible, subtle deviation between the voltage measurement value output by the voltage sampling sensor and the actual value under specific load conditions. When the duty cycle calculation logic of the pulse width modulation signal in the control circuit receives these slightly deviated sampling result data streams, its response accuracy to changes is affected. This can lead to potential slight lag or overshoot in the controller's adjustment actions when attempting to compensate for voltage changes. Meanwhile, to cope with continuous business growth, data centers typically introduce a large number of server devices employing advanced switching power supply technology. These servers' power supply modules (SMPS) no longer draw current smoothly from the grid like traditional linear power supplies; instead, they draw current in short, high-amplitude pulses within each AC cycle through high-speed switching. This operating method makes these devices significantly nonlinear loads, whose current and voltage waveforms no longer exhibit a simple linear relationship and possess rapidly changing transient characteristics.

[0004] When these nonlinear loads experience frequent power surges, control systems with computational biases cannot respond perfectly and quickly to these transient changes. This causes frequent, brief, and minute voltage overshoots or drops in the UPS transformer output voltage. While these frequent, minute transient voltage fluctuations may have little impact on an individual occasion, their cumulative effect creates continuous electrical stress on critical computing equipment within the data center that is sensitive to power quality. This repeated electrical shock accelerates the internal aging process of these components, leading to increased equivalent series resistance (ESR) of electrolytic capacitors, increased dielectric losses, or damage to the gate oxide layer of semiconductor devices. Over time, this can cause some equipment to experience performance degradation at unexpected times, such as increased calculation error rates, slower response times, or even intermittent failures, such as unexpected server restarts or failures of specific functional modules.

[0005] Because each voltage fluctuation is extremely brief and minor, far below the thresholds set by conventional UPS alarm systems for persistent imbalances or large single fluctuations, the alarm mechanism fails to trigger. When data center operations personnel face frequent performance degradation or intermittent failures of downstream IT equipment, they often attribute the equipment failures to hardware aging, software defects, firmware issues, or network configuration errors, since the UPS system reports everything as normal. This leads the operations team to spend a significant amount of time and resources troubleshooting, replacing hardware, and upgrading software. However, because the root cause of the problem—the implicit power quality issues at the UPS output—is not addressed, these failures will recur, severely impacting the operational efficiency and reliability of the data center. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this application provides a UPS transformer voltage balance control method and system.

[0007] In a first aspect, this application provides a UPS transformer voltage balance control method, which includes the following steps:

[0008] Establish communication links between UPS transformers and power modules of sensitive equipment within the data center;

[0009] Receives microscopic voltage quality data generated by the power module monitoring minute transient fluctuations in its input voltage and feeding it back;

[0010] Based on micro voltage quality data, the impact of power supply quality on sensitive equipment is assessed, and the assessment results are obtained.

[0011] Based on the evaluation results, dynamically adjust the target margin of the output voltage of each phase of the UPS transformer;

[0012] The output voltage of the UPS transformer is controlled based on the adjusted target output voltage margin.

[0013] By establishing a communication link, receiving micro voltage quality data, evaluating power supply quality, dynamically adjusting the output voltage target margin, and controlling the output voltage, the system achieves refined and dynamic balance control of the UPS transformer output voltage. This effectively addresses minute transient voltage fluctuations, ensures the power supply quality of sensitive equipment, and solves the problem of difficulty in detecting and handling minute voltage fluctuations in existing technologies.

[0014] Furthermore, this application also proposes that the method further includes the following steps:

[0015] After controlling the output voltage of the UPS transformer, the micro voltage quality data and the UPS transformer's own operating data are analyzed to identify power supply problems and generate diagnostic information.

[0016] Furthermore, this application also proposes a step for dynamically adjusting the target margin of the output voltage of each phase of the UPS transformer based on the evaluation results, including:

[0017] Obtain the output current data of each phase of the UPS transformer;

[0018] Based on the output current data, identify the transient load imbalance between each phase;

[0019] Based on the evaluation results and transient load imbalance, the target margin of the output voltage of each phase is adjusted in a coordinated manner.

[0020] Furthermore, this application also proposes that, based on the evaluation results, the step of dynamically adjusting the target margin of the output voltage of each phase of the UPS transformer further includes:

[0021] Monitor the balance of the output voltage of each phase, and adjust the target margin of the output voltage of each phase according to the balance compensation.

[0022] Furthermore, this application also proposes a method for assessing the impact of power supply quality on sensitive equipment based on microscopic voltage quality data, and obtaining the assessment results, including the following steps:

[0023] Receive the device type identifier and current operating load status sent by the power module;

[0024] Obtain the device type identifier and the voltage fluctuation tolerance parameters corresponding to the current operating load status;

[0025] Based on microscopic voltage quality data and voltage fluctuation tolerance parameters, the impact of power supply quality on sensitive equipment is assessed, and the assessment results are obtained.

[0026] Furthermore, this application also proposes steps for analyzing microscopic voltage quality data and UPS transformer operating data to identify power supply problems and generate diagnostic information, including:

[0027] Obtain the UPS transformer's own operating data and the current operating load of sensitive equipment;

[0028] Obtain the physical topology information of sensitive devices within the data center. This physical topology information includes the power supply path between the sensitive devices and the UPS transformer, the physical location of the sensitive devices, and the rack information to which the sensitive devices belong.

[0029] In micro voltage quality data, identify time-series data related to voltage fluctuation events occurring in sensitive devices under specific load conditions;

[0030] The identified time-series data is correlated with the UPS transformer's own operating data, the current operating load status of sensitive equipment, and physical topology information to infer the physical root cause of voltage fluctuation events. This physical root cause includes poor line contact, abnormal local impedance, or distributed electromagnetic interference sources, and the physical location of the physical root cause is determined.

[0031] Generate diagnostic information that includes the physical root cause and physical location.

[0032] Furthermore, this application proposes a method for correlating the identified time-series data with the UPS transformer's own operating data, the current operating load status of sensitive equipment, and physical topology information to infer the physical root cause of voltage fluctuation events. This method includes:

[0033] Features for distinguishing physical root causes are extracted from the identified time-series data, UPS transformer operating data, current operating load status of sensitive equipment, and acquired physical topology information, and combined into a fault signature.

[0034] According to the preset classification rules, the fault signatures are classified to obtain preliminary candidate physical sources;

[0035] Based on the physical topology information, preliminary candidate physical sources are logically eliminated in order to infer and determine the physical source.

[0036] Furthermore, this application proposes a step for extracting features to distinguish physical root causes from identified time-series data, UPS transformer operating data, current operating load status of sensitive equipment, and acquired physical topology information, including:

[0037] From the identified time-series data, extract the mean, variance, peak value, harmonic content, and transient duration of the voltage.

[0038] Extract voltage output deviation and current imbalance from the UPS transformer's own operating data;

[0039] Extract the load rate and power consumption change rate from the current operating load status of sensitive devices;

[0040] The power supply path length and the number of connected nodes are extracted from the acquired physical topology information.

[0041] Furthermore, this application proposes that the steps for assembling a fault signature include:

[0042] Weights are assigned based on the discriminative power of the extracted features against different physical sources.

[0043] Based on the assigned weights, the extracted features are weighted and combined to form a fault signature.

[0044] Secondly, this application also proposes a UPS transformer voltage balance control system, which includes:

[0045] The communication module is used to establish a communication link between the UPS transformer and the power modules of sensitive equipment in the data center;

[0046] The receiving module is used to receive microscopic voltage quality data generated by the power supply module monitoring minute transient fluctuations in its input voltage and feeding back the data.

[0047] The evaluation module is used to assess the impact of power supply quality on sensitive equipment based on micro voltage quality data and obtain evaluation results.

[0048] The adjustment module is used to dynamically adjust the target margin of the output voltage of each phase of the UPS transformer based on the evaluation results.

[0049] The control module is used to control the output voltage of the UPS transformer according to the adjusted output voltage target margin.

[0050] In summary, this application provides a UPS transformer voltage balance control method and system. By establishing a communication link between the UPS transformer and the power modules of sensitive equipment within the data center, it can receive micro-voltage quality data monitored and fed back by the power modules. Based on this micro-voltage quality data, it further assesses the impact of power supply quality on sensitive equipment and dynamically adjusts the target margin of the output voltage of each phase of the UPS transformer, ultimately controlling the output voltage of the UPS transformer. This application effectively solves the problem of minute transient voltage fluctuations caused by resistance drift of internal components and nonlinear load characteristics in existing UPS systems. Traditional UPS systems struggle to detect and respond to these minute fluctuations, leading to long-term accumulated electrical stress on sensitive equipment, resulting in performance degradation or intermittent failures. This application achieves refined perception of power supply quality by directly acquiring micro-voltage quality data from the power modules of sensitive equipment, overcoming the limitations of traditional UPS systems that rely solely on their own voltage sampling sensors for coarse control. By dynamically adjusting the target margin of the output voltage, this application can accurately and in real-time compensate and balance the output voltage of the UPS transformer based on the actual needs of the sensitive equipment and the power supply quality assessment results, effectively suppressing minute transient voltage fluctuations. This not only avoids the problem of traditional alarm systems failing to trigger due to fluctuations not reaching the threshold, but also fundamentally reduces the failure rate of sensitive equipment caused by power quality issues, significantly improving the operational efficiency and reliability of the data center. Therefore, this application can provide a more stable and reliable power supply, extend the lifespan of sensitive equipment, and reduce the time and resources spent by maintenance personnel on troubleshooting. Attached Figure Description

[0051] Figure 1 This is a flowchart illustrating a UPS transformer voltage balance control method provided in an embodiment of this application.

[0052] Figure 2 This is a schematic diagram of a UPS transformer voltage balance control system provided in an embodiment of this application.

[0053] Labeling explanation: 210, Communication module; 220, Receiving module; 230, Evaluation module; 240, Adjustment module; 250, Control module. Detailed Implementation

[0054] The technical solutions of this application will now be clearly and completely described with reference to the accompanying drawings. Obviously, the described embodiments are merely some embodiments of this application, and not all embodiments. The components of this application 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 this application provided in the accompanying drawings is not intended to limit the scope of the claimed application, but merely to illustrate selected embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are within the scope of protection of this application.

[0055] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this application, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0056] In traditional UPS systems, over long-term operation, the precision electronic components, such as voltage sampling sensors, experience resistance drift due to Joule heating and ambient temperature fluctuations. This minute resistance drift causes a slight deviation between the measured and actual voltage values, affecting the control circuit's accurate calculation of the pulse-width modulation signal duty cycle. When data centers introduce numerous non-linear loads (such as servers using switching power supply technology), the sudden power surges of these loads prevent the control system, already biased by calculation errors, from responding quickly enough. This results in frequent, brief, and minute transient fluctuations in the UPS transformer output voltage. These accumulated minute transient voltage fluctuations exert continuous electrical stress on sensitive equipment, accelerating component aging and leading to performance degradation or intermittent failures. Because these fluctuations do not reach conventional alarm thresholds, maintenance personnel struggle to identify the root cause, making troubleshooting difficult and severely impacting the operational efficiency and reliability of the data center.

[0057] Regarding this, firstly, see... Figure 1 This application proposes a UPS transformer voltage balance control method, including the following steps:

[0058] Establish communication links between UPS transformers and power modules of sensitive equipment within the data center;

[0059] Receives microscopic voltage quality data generated by the power module monitoring minute transient fluctuations in its input voltage and feeding it back;

[0060] Based on micro voltage quality data, the impact of power supply quality on sensitive equipment is assessed, and the assessment results are obtained.

[0061] Based on the evaluation results, dynamically adjust the target margin of the output voltage of each phase of the UPS transformer;

[0062] The output voltage of the UPS transformer is controlled based on the adjusted target output voltage margin.

[0063] The UPS transformer is the core component of a UPS system, responsible for voltage transformation and isolation of electrical energy to meet the power supply needs of sensitive equipment within the data center. Sensitive equipment in a data center typically refers to IT equipment with extremely high power quality requirements, such as servers, storage arrays, and network devices. These devices usually integrate power modules to convert the input AC power into the DC power required by the equipment. The communication link refers to the data transmission channel established between the UPS transformer and the power modules of the sensitive equipment. This can be a physical connection and logical protocol stack based on Ethernet, CAN bus, RS485, or other industrial communication protocols, used to achieve bidirectional data exchange. Microscopic voltage quality data refers to the minute transient fluctuations in the input voltage monitored by the power module. These fluctuations may include voltage dips, overshoots, and harmonic distortions, characterized by small amplitude and short duration, making them difficult to capture by traditional UPS monitoring systems. The target margin refers to the allowable deviation range of the output voltage of each phase of the UPS transformer from its rated value. By dynamically adjusting this margin, the output voltage can be controlled more precisely to meet the actual needs of the sensitive equipment.

[0064] Establishing a communication link between the UPS transformer and the power modules of sensitive equipment within the data center can be achieved in several ways. For example, a wired connection can be used, such as connecting the UPS transformer's control unit to the power modules of various sensitive devices in the data center via Ethernet cables, and configuring the appropriate TCP / IP protocol stack for data transmission. Another approach is to use a wireless connection, such as establishing a wireless communication network between the UPS transformer and the power modules via Wi-Fi or Bluetooth modules to achieve data exchange. A hybrid communication approach can also be used, where some devices are connected via wired connections and others via wireless connections, to adapt to different deployment environments and equipment types.

[0065] After the communication link is established, the power module can integrate a high-precision voltage sensor and data acquisition unit to monitor minute transient fluctuations in its input voltage in real time. For example, the power module can sample the input voltage at a sampling rate of thousands or even higher per second, and perform preliminary processing on the sampled data, such as filtering and noise reduction. Then, it encapsulates the identified micro-fluctuation information, such as voltage dips, overshoots, and harmonic content, into data packets and sends them to the UPS transformer's control system through the established communication link. Subsequently, based on the micro-voltage quality data, the impact of power supply quality on sensitive equipment is assessed, and an assessment result is obtained. The assessment process may include real-time analysis of the received micro-voltage quality data, such as comparing parameters like the amplitude, duration, and frequency of voltage fluctuations with preset voltage tolerance thresholds for sensitive equipment. For example, if a sensitive device has a tolerance of 5% voltage dip lasting 10 milliseconds, and the received data shows that a 4% voltage dip lasting 15 milliseconds actually occurred, it can be assessed that there has been a potential impact on the sensitive device. The assessment result can be a quantitative indicator, such as a power quality impact index, or a qualitative judgment, such as "minor impact," "moderate impact," or "serious impact."

[0066] Next, based on the evaluation results, the target margin of the output voltage of each phase of the UPS transformer is dynamically adjusted. For example, if the evaluation results show that the power quality has a moderate impact on sensitive equipment, the target margin of the UPS transformer output voltage can be tightened accordingly, such as adjusting the original allowable voltage fluctuation range of ±2% to ±1.5%. This adjustment can be based on a preset lookup table, directly mapping different evaluation results to the corresponding target margin; or it can be based on fuzzy logic or PID control algorithms, making continuous fine adjustments based on real-time changes in the evaluation results. Finally, the UPS transformer output voltage is controlled according to the adjusted output voltage target margin. The UPS transformer control system will precisely adjust the output voltage of each phase according to the new target margin through its internal power electronics (such as IGBTs, MOSFETs, etc.) and control algorithms (such as PWM control). For example, if the target margin is tightened, the control system will more actively perform voltage compensation to ensure that the output voltage is always maintained within a stricter range. This may involve switching the taps of the transformer windings or fine-tuning the duty cycle of the inverter output voltage to achieve precise control of the output voltage.

[0067] This application establishes a communication link between the UPS transformer and the power supply modules of sensitive equipment within the data center, enabling direct monitoring and feedback of minute transient fluctuations in the input voltage of sensitive equipment. Traditional UPS systems primarily rely on voltage sampling at their own output terminals, making it difficult to capture microscopic voltage quality issues at the sensitive equipment level caused by factors such as line impedance and load transient changes. This application, by receiving microscopic voltage quality data fed back from the power supply modules, can more accurately assess the impact of power supply quality on sensitive equipment, thus obtaining more refined evaluation results. Based on these evaluation results, the target margin of the output voltage of each phase of the UPS transformer is dynamically adjusted, and the output voltage of the UPS transformer is controlled accordingly.

[0068] Compared to existing technologies, the innovation of this application lies in extending the "tentacles" of voltage quality monitoring to the sensitive equipment end, realizing end-to-end voltage quality perception and control from the "source" to the "load". Traditional methods often only perform macroscopic voltage balance control at the UPS output end, and are powerless against local and microscopic voltage fluctuations generated in the complex power supply network within the data center. This application, by directly acquiring the microscopic voltage quality data of the power modules of sensitive equipment, can identify those tiny transient fluctuations that are undetectable by traditional UPS systems and have cumulative damage to sensitive equipment. For example, when a server power module reports frequent small drops in its input voltage, this application can respond immediately, dynamically adjusting the target margin of the output voltage of the corresponding phase of the UPS transformer for more precise compensation, thereby effectively suppressing these microscopic fluctuations. This refined and dynamic control strategy significantly improves the stability and reliability of data center power supply, reduces performance degradation and failures of sensitive equipment due to hidden power quality problems, thereby reducing operation and maintenance costs and improving the overall operational efficiency of the data center.

[0069] Furthermore, it also includes the following steps:

[0070] After controlling the output voltage of the UPS transformer, the micro voltage quality data and the UPS transformer's own operating data are analyzed to identify power supply problems and generate diagnostic information.

[0071] Specifically, after the UPS transformer completes output voltage control based on the adjusted output voltage target margin, this application will further analyze two key data points in depth. The first is micro-voltage quality data, monitored and fed back by the power modules of sensitive equipment within the data center. This data reflects minute transient fluctuations in the input voltage of the sensitive equipment, directly reflecting power quality at the terminal equipment side. The second is the UPS transformer's own operating data, encompassing various parameters of the UPS transformer during operation, such as phase output voltage, output current, frequency, internal temperature, load rate, and any internal alarms or fault indications. Comprehensive analysis of these two types of data aims to identify potential power supply problems. Identifying power supply problems means determining the specific cause of voltage fluctuations or imbalances, such as external power grid fluctuations, internal UPS transformer faults, distribution line problems, or changes in the load characteristics of the sensitive equipment itself. Once a power supply problem is identified, corresponding diagnostic information will be generated. This diagnostic information typically includes the type and severity of the problem, possible causes, and recommended solutions, such as indicating which line segments need to be inspected, which components need to be replaced, or which configuration parameters need to be adjusted.

[0072] This application effectively overcomes the limitations of relying solely on real-time voltage adjustment by introducing subsequent analysis of microscopic voltage quality data and the UPS transformer's own operational data. When a UPS transformer performs voltage balancing control, its primary focus is maintaining the output voltage within a preset range. However, if voltage fluctuations are caused by deeper physical causes (such as poor line contact or abnormal local impedance), simple voltage adjustment may only provide a temporary solution. By correlating the microscopic voltage quality data reported by sensitive devices with the UPS transformer's own operational status data, the power supply environment can be examined from multiple dimensions. For example, if a sensitive device reports a persistent voltage drop, while the UPS transformer's own output voltage and current data show normal readings, this may indicate a problem in the power distribution link between the UPS transformer and the sensitive device. Conversely, if the UPS transformer's own data also shows abnormalities, it may point to an internal fault within the UPS transformer. This comprehensive analysis goes beyond simple symptom responses, delving deeper to uncover and infer the physical root causes of power supply problems, thus providing a precise basis for subsequent maintenance and troubleshooting. Traditional voltage balancing control can only passively respond to voltage fluctuations. This application, however, by adding diagnostic steps, can proactively identify and locate the root cause of power supply problems, thereby preventing recurrence or escalation of issues. This enables a shift from passive response to proactive diagnosis, significantly shortening troubleshooting time, reducing maintenance costs, and effectively preventing equipment damage or service interruptions caused by undetected potential problems, ensuring the continuous and stable operation of sensitive data center equipment.

[0073] Furthermore, the steps for dynamically adjusting the target margin of the output voltage of each phase of the UPS transformer based on the evaluation results include:

[0074] Obtain the output current data of each phase of the UPS transformer;

[0075] Based on the output current data, identify the transient load imbalance between each phase;

[0076] Based on the evaluation results and transient load imbalance, the target margin of the output voltage of each phase is adjusted in a coordinated manner.

[0077] Specifically, obtaining the output current data of each phase of the UPS transformer refers to monitoring and collecting the instantaneous current values ​​of each phase in real time by configuring current sensors at the output end of the UPS transformer. This current data is a direct indicator reflecting the load status of each phase. Identifying the transient load imbalance between phases based on the output current data can be understood as comparing and analyzing the real-time collected output current data of each phase to calculate the degree of difference or imbalance coefficient between the currents of each phase. For example, the deviation of each phase current from the average current can be calculated, or the imbalance calculation method defined in the International Electrotechnical Commission (IEC) standard can be used to quantify the uniformity of the load distribution of each phase at the current moment, thereby determining whether there are transient load differences that may lead to voltage imbalance. In practical applications, coordinating and adjusting the target margin of the output voltage of each phase based on the evaluation results and the transient load imbalance specifically involves combining the overall power quality evaluation results obtained from microscopic voltage quality data with the real-time identified inter-phase transient load imbalance for comprehensive consideration. For example, when the evaluation results show that the overall power supply quality is good, but at the same time a sudden increase in the load of a certain phase is detected, which leads to an increase in the imbalance, the voltage target margin of that phase will be fine-tuned first to compensate for its voltage drop trend, while taking into account the overall voltage quality requirements. This coordinated adjustment aims to ensure that the balance of the output voltage of each phase is maintained to the maximum extent while meeting the overall power supply quality requirements.

[0078] This application introduces real-time monitoring of the output current data of each phase of the UPS transformer and identification of transient load imbalance. This allows for a more refined adjustment of the voltage target margin, moving beyond reliance solely on macroscopic power quality assessments to more precisely perceive and respond to dynamic changes in the load of each phase. By acquiring real-time information on inter-phase load imbalance, the target margin of the output voltage for each phase can be adjusted in a coordinated manner. For example, when the load on a phase suddenly increases, its output current will increase accordingly. This transient load imbalance can be immediately identified, and the voltage target margin for that phase can be appropriately increased to offset the voltage drop caused by the increased load, thereby effectively maintaining the balance of voltage across phases. This mechanism ensures that the UPS transformer can still provide a stable and balanced output voltage even under dynamic load changes or imbalances. By monitoring the output current of each phase in real time and identifying transient load imbalances, this application can more accurately and promptly adjust the target margin of the output voltage of each phase, effectively suppressing voltage imbalance caused by inter-phase load differences, and significantly improving the voltage balance control capability of the UPS transformer under dynamic and unbalanced load conditions. This not only helps extend the lifespan of sensitive equipment and reduce the risk of equipment failure caused by voltage fluctuations, but also ensures that all sensitive equipment in the data center receives a high-quality, high-stability power supply, thereby improving the operational reliability and efficiency of the entire data center.

[0079] Furthermore, based on the evaluation results, the steps for dynamically adjusting the target margin of the output voltage of each phase of the UPS transformer also include:

[0080] Monitor the balance of the output voltage of each phase, and adjust the target margin of the output voltage of each phase according to the balance compensation.

[0081] Specifically, monitoring the balance of each phase's output voltage refers to the real-time acquisition of instantaneous voltage values ​​for each phase using voltage sensors installed at the output of the UPS transformer. This voltage data is then sent to the control unit for analysis to calculate the degree of balance between the phase voltages. The balance can be understood as the degree to which the amplitude or phase angle of each phase voltage deviates from the ideal symmetry. For example, it can be quantified by calculating the voltage imbalance rate or the deviation of each phase voltage from the average phase voltage, accurately identifying any potential voltage asymmetry. Furthermore, adjusting the target margin of each phase's output voltage based on balance compensation means that once an imbalance is detected in the output voltage of each phase, the target margin of each phase's output voltage, previously determined based on evaluation results and transient load imbalance, is corrected according to the degree and direction of the imbalance. For example, if the voltage of a certain phase remains consistently low, the target margin for that phase can be appropriately increased; conversely, if the voltage of a certain phase remains consistently high, its target margin can be appropriately decreased. This compensation adjustment is a fine-tuning mechanism designed to guide the output voltage of each phase closer to an ideal balance.

[0082] This application introduces direct monitoring of the output voltage balance of each phase, enabling real-time acquisition of the symmetry information of the actual output voltage of the UPS transformer. Continuous monitoring of voltage balance identifies voltage deviations that are difficult to detect using current data alone, or that persist even after load imbalance adjustments. Based on this, the target margin of each phase's output voltage is compensated and adjusted according to the monitored voltage balance, forming a closed-loop voltage balance control mechanism. This ensures that even under complex load changes or fluctuations in internal system parameters, the UPS transformer maintains a highly balanced output voltage, effectively avoiding operational risks to sensitive equipment caused by voltage imbalance. Compared to adjustments based solely on current data, the introduction of voltage balance monitoring and compensation allows for a more comprehensive and accurate response to various unbalanced operating conditions, effectively reducing deviations in the output voltage of each phase. This significantly improves the stability and symmetry of power supply quality, providing a more reliable and high-quality power environment for sensitive equipment in data centers, further ensuring normal equipment operation and data security, and extending equipment lifespan.

[0083] Furthermore, based on microscopic voltage quality data, the steps for assessing the impact of power supply quality on sensitive equipment and obtaining the assessment results include:

[0084] Receive the device type identifier and current operating load status sent by the power module;

[0085] Obtain the device type identifier and the voltage fluctuation tolerance parameters corresponding to the current operating load status;

[0086] Based on microscopic voltage quality data and voltage fluctuation tolerance parameters, the impact of power supply quality on sensitive equipment is assessed, and the assessment results are obtained.

[0087] Specifically, upon receiving the device type identifier and current operating load status from the power module, the power module can periodically, or upon detecting a significant change in status, send the unique identifier of the sensitive device, its device type (e.g., CPU, GPU, storage device, network interface card, etc.), and its current operating load status (e.g., CPU utilization, memory utilization, I / O throughput, etc.) to the UPS transformer control system. The device type identifier distinguishes the inherent characteristics of different sensitive devices to voltage fluctuations, while the current operating load status reflects the device's power consumption and sensitivity to power at a specific moment. Obtaining the voltage fluctuation tolerance parameters corresponding to the device type identifier and current operating load status can be understood as retrieving the corresponding voltage fluctuation tolerance parameters from a pre-stored database or lookup table based on the received device type identifier and current operating load status. Voltage fluctuation tolerance parameters may include voltage sag thresholds, voltage overshoot thresholds, transient voltage change rate limits, harmonic content limits, etc. These parameters are calibrated and stored according to the electrical characteristics and reliability requirements of different device types under different load conditions. For example, high-performance computing devices may have a lower tolerance for voltage drops under heavy loads, but a relatively higher tolerance under light loads.

[0088] In practical applications, the impact of power supply quality on sensitive equipment is assessed based on micro-voltage quality data and voltage fluctuation tolerance parameters. This assessment involves comparing and analyzing the micro-voltage quality data monitored by the power module (e.g., instantaneous voltage value, voltage fluctuation amplitude, duration, etc.) with the acquired voltage fluctuation tolerance parameters. This comparison quantifies the degree of impact of the current power supply quality on specific sensitive equipment under specific operating conditions. For example, if the voltage drop in the micro-voltage quality data exceeds the voltage drop threshold for the equipment under its current load, the assessment result will indicate a potential power supply quality problem and may provide a specific risk level or degree of impact.

[0089] This application introduces the equipment type identifier, current operating load status, and corresponding voltage fluctuation tolerance parameters of sensitive devices, enabling power quality assessment to be targeted and refined, rather than a single, universal judgment. The limitations of traditional general assessment methods stem from the varying sensitivities of different sensitive devices to voltage fluctuations and the different tolerances exhibited by the same device under different loads. This application, by receiving the equipment type identifier and current operating load status, can accurately identify the specific operating condition of the sensitive device to be assessed. By acquiring voltage fluctuation tolerance parameters precisely matching this operating condition, a customized benchmark is provided for subsequent assessments. Comparing microscopic voltage quality data with these customized tolerance parameters allows for a more accurate determination of whether power quality issues have truly impacted the sensitive device, thus avoiding misjudgments or omissions and providing a more reliable basis for the dynamic adjustment of the UPS transformer output voltage. Compared to general assessment methods that do not distinguish between equipment type and load status, this application can more accurately identify the impact of power quality problems on specific sensitive devices, avoiding excessive or insufficient voltage adjustments due to inaccurate assessments. This significantly improves the accuracy and response speed of UPS transformer voltage balance control, ensuring that sensitive equipment can obtain a stable power supply that meets its specific needs under various operating conditions, thereby effectively improving the overall operational reliability and efficiency of the data center.

[0090] Furthermore, the steps of analyzing microscopic voltage quality data and UPS transformer operating data to identify power supply problems and generate diagnostic information include:

[0091] Obtain the UPS transformer's own operating data and the current operating load of sensitive equipment;

[0092] Obtain the physical topology information of sensitive devices within the data center. The physical topology information includes the power supply path between the sensitive device and the UPS transformer, the physical location of the sensitive device, and the rack information to which the sensitive device belongs.

[0093] In micro voltage quality data, identify time-series data related to voltage fluctuation events occurring in sensitive devices under specific load conditions;

[0094] The identified time-series data is correlated with the UPS transformer's own operating data, the current operating load status of sensitive equipment, and physical topology information to infer the physical root cause of voltage fluctuation events. The physical root cause includes poor line contact, abnormal local impedance, or distributed electromagnetic interference sources, and the physical location of the physical root cause is determined.

[0095] Generate diagnostic information that includes the physical root cause and physical location.

[0096] Specifically, when acquiring the UPS transformer's own operating data and the current operating load of sensitive equipment, the UPS transformer's own operating data can include real-time or historical operating parameters such as output voltage, output current, temperature, frequency, and power factor. The current operating load of sensitive equipment can refer to the equipment's current power consumption, CPU utilization, memory utilization, etc. These data reflect the equipment's power demand, and can be obtained through communication with the monitoring systems inside the UPS transformer and sensitive equipment. Acquiring the physical topology information of sensitive equipment within the data center is a crucial step. Physical topology information can be understood as a structured description of the power distribution network within the data center, clearly defining the electrical connections and spatial layout between sensitive equipment and the UPS transformer. The power supply path can refer to all electrical connections, such as distribution cabinets, PDUs, and cables, from the UPS transformer to the sensitive equipment. The physical location of sensitive equipment can refer to its specific coordinates within the data center server room or its rack number, U-position, etc. Rack information can be further refined to details such as power sockets and internal cabling within the rack. This information can be obtained and stored through the data center's asset management system, CAD drawings, or manual input.

[0097] Identifying time-series data related to voltage fluctuations in sensitive equipment under specific load conditions within micro-voltage quality data involves filtering data segments from continuously acquired micro-voltage quality data streams that are relevant to voltage anomalies (such as sudden drops, surges, momentary interruptions, harmonic distortion, etc.) occurring at specific points in time or within a specific time period for a particular sensitive device. Specific load conditions refer to the operating state of the sensitive device when voltage fluctuations occur, such as high-load operation, startup, or during specific task execution. The identification process can be automated using threshold detection, pattern matching, or machine learning algorithms. The identified time-series data is then correlated with the UPS transformer's own operating data, the current operating load state of the sensitive device, and physical topology information to uncover the underlying causes of voltage fluctuations from multi-dimensional data. Correlation analysis can employ statistical methods, data mining techniques, or expert system rules. Through cross-validation and comparison of multi-source data, the physical root cause of voltage fluctuation events can be inferred. Physical root causes may include poor line contact, such as a loose or oxidized connector; local impedance anomalies, such as increased impedance due to aging or damage to a cable segment; or distributed electromagnetic interference sources, such as electromagnetic interference generated by the startup or operation of nearby high-power equipment. By inferring the physical root cause and combining it with physical topology information, the specific physical location of the physical root cause can be determined, such as the output port of PDU 3 in rack A or a section of cable connecting racks B and C. Diagnostic information containing the physical root cause and physical location is generated, and the analysis results are output in a structured and easily understandable format. The diagnostic information includes the fault type, occurrence time, affected equipment, inferred physical root cause, precise physical location, and possible recommended remedial measures.

[0098] This application effectively addresses the limitation of relying solely on voltage quality data to accurately pinpoint the physical root cause of power supply problems by introducing multi-dimensional data sources and a refined correlation analysis mechanism. By acquiring UPS transformer operating data and the current operating load of sensitive equipment, more comprehensive background information can be provided for voltage fluctuation events, distinguishing between systemic problems and issues caused by local load characteristics. Simultaneously, acquiring the physical topology information of sensitive equipment within the data center provides the necessary spatial and connectivity context for subsequent root cause localization. Identifying time-series data related to specific events within micro-level voltage quality data ensures the targeted nature of the analysis. Most importantly, correlation analysis of these multi-source data sources allows for a deeper understanding of the underlying physical mechanisms beyond the surface-level voltage fluctuations. For example, by comparing voltage fluctuations at specific time points with UPS transformer internal parameters, load changes in sensitive equipment, and node states along the power supply path, it can be inferred whether the problem is caused by poor line contact, abnormal local impedance, or distributed electromagnetic interference sources. Combining this with physical topology information allows for precise determination of the physical location of the root cause, transforming the abstract "power supply problem" into a concrete and actionable "fault point." Compared to simply identifying power supply issues, this application can directly infer the physical root cause of voltage fluctuation events and determine their precise physical location. This allows maintenance personnel to quickly locate the fault point, avoiding blind troubleshooting and unnecessary downtime. It can significantly shorten fault recovery time, reduce operation and maintenance costs, and effectively ensure the power supply continuity and stability of sensitive equipment in the data center, thereby improving the reliability and availability of the entire data center.

[0099] Furthermore, the steps to correlate and analyze the identified time-series data with the UPS transformer's own operating data, the current operating load status of sensitive equipment, and the acquired physical topology information to infer the physical root cause of voltage fluctuation events include:

[0100] Features for distinguishing physical root causes are extracted from the identified time-series data, UPS transformer operating data, current operating load status of sensitive equipment, and acquired physical topology information, and combined into a fault signature.

[0101] According to the preset classification rules, the fault signatures are classified to obtain preliminary candidate physical sources;

[0102] Based on the physical topology information, preliminary candidate physical sources are logically eliminated in order to infer and determine the physical source.

[0103] Specifically, extracting features to distinguish physical root causes refers to identifying and quantifying data attributes from multi-source data that are significantly correlated with different physical root causes (such as poor line contact, local impedance anomalies, or distributed electromagnetic interference sources). These features can be numerical or descriptive, providing a basis for subsequent classification and inference. For example, voltage drop depth, duration, and recovery rate in time-series data; output current imbalance and harmonic content in UPS transformer operating data; instantaneous power changes in the load state of sensitive equipment; and power supply path length and number of connection points in physical topology information can all serve as potential features. Combining these features into a fault signature can be understood as integrating multiple extracted features according to certain logic or algorithms to form a data set that can uniquely or highly characterize a specific fault mode. This fault signature, as a comprehensive fingerprint, is used to describe the intrinsic characteristics of a specific voltage fluctuation event and its correlation with potential physical root causes. For example, it can be combined through weighted averaging, feature vector construction, or pattern matching.

[0104] In practical applications, classifying fault signatures according to preset classification rules to obtain preliminary candidate physical sources involves comparing the generated fault signatures with known fault modes using a pre-established knowledge base or model. These classification rules can be constructed based on expert experience, statistical models, or machine learning algorithms (such as decision trees, support vector machines, neural networks, etc.). Based on the characteristics of the fault signature, the most likely type of physical source causing the voltage fluctuation event is initially determined, such as poor line contact, abnormal local impedance, or a source of distributed electromagnetic interference. Logically eliminating preliminary candidate physical sources based on physical topology information to infer and determine the final physical source involves further screening and verification of candidate sources based on the preliminary classification results, combined with the physical topology information (including power supply paths, physical locations, rack information, etc.) of sensitive equipment within the data center. For example, if the preliminary classification result points to "poor line contact," but the physical topology information shows no connection points prone to poor contact near the specific sensitive equipment, then this candidate source may be eliminated. Through this logical elimination, candidate sources that do not conform to the actual physical layout or operating conditions can be eliminated, more accurately pinpointing the final physical source and its physical location.

[0105] This application effectively addresses the diagnostic ambiguity that may exist in traditional correlation analysis in complex power supply environments by introducing a structured root cause inference process. By extracting features from multi-dimensional data to distinguish physical root causes, it ensures that the diagnostic process can fully utilize various effective information, laying the foundation for subsequent accurate identification. The combination of these features forms a fault signature, giving each voltage fluctuation event a unique "fingerprint," thereby enhancing the distinguishability between different fault modes. Classifying the fault signatures according to preset classification rules can quickly and initially identify the most likely physical root cause type, greatly narrowing the scope of investigation. Logical elimination using physical topology information, comparing the preliminary classification results with the actual physical environment, effectively eliminates unreasonable or impossible candidate root causes, thereby avoiding misjudgments and ensuring the accuracy and reliability of the finally inferred physical root causes and their locations. By systematically extracting multi-source features and constructing fault signatures, this application enables different types of physical root causes to be more clearly identified, significantly improving the accuracy and reliability of power supply problem diagnosis in UPS transformer voltage balance control methods. By combining preliminary screening with preset classification rules and logical exclusion with physical topology information, this application effectively avoids misdiagnosis or missed diagnosis that may occur with traditional methods. Especially when multiple potential faults coexist or fault characteristics are not obvious, this application can more accurately infer the physical root cause and physical location of voltage fluctuation events.

[0106] Furthermore, the steps for extracting features to distinguish physical root causes from the identified time-series data, UPS transformer operating data, current operating load status of sensitive equipment, and acquired physical topology information include:

[0107] From the identified time-series data, extract the mean, variance, peak value, harmonic content, and transient duration of the voltage.

[0108] Extract voltage output deviation and current imbalance from the UPS transformer's own operating data;

[0109] Extract the load rate and power consumption change rate from the current operating load status of sensitive devices;

[0110] The power supply path length and the number of connected nodes are extracted from the acquired physical topology information.

[0111] Specifically, the mean, variance, peak value, harmonic content, and transient duration of voltage are extracted from the identified time-series data to quantify the basic characteristics of voltage fluctuations and the severity of transient events. The mean voltage reflects the overall voltage level, the variance characterizes voltage stability, the peak value captures extreme voltage fluctuations, the harmonic content reveals the degree of voltage waveform distortion, and the transient duration defines the time scale of fluctuation events. Extracting these features helps to comprehensively characterize the microscopic changes in voltage quality over time. Voltage output deviation and current imbalance are extracted from the UPS transformer's own operating data to assess its operational status. Voltage output deviation reflects the degree of deviation between the UPS transformer's output voltage and the set value, while current imbalance indicates the balance of the three-phase load. These data are directly related to the internal operating state of the UPS transformer and are crucial for determining whether power supply problems originate from the UPS transformer itself. Furthermore, load rate and power consumption change rate are extracted from the current operating load status of sensitive equipment to understand the response characteristics of sensitive equipment to power quality. The load rate reflects the actual operating load of the equipment, while the power consumption change rate reveals the fluctuations in the equipment's energy demand over a short period. These characteristics help analyze whether the operating modes of sensitive equipment are related to voltage fluctuation events. For example, a sudden high-power start-up or shutdown of certain equipment may cause a local voltage drop or rise. Simultaneously, the power supply path length and the number of connection nodes are extracted from the acquired physical topology information to quantify the physical characteristics of the power supply network. The power supply path length directly affects line impedance and voltage drop, while the number of connection nodes reflects the complexity of the power supply network and potential fault points. These topological features provide a spatial basis for analyzing the physical propagation path of voltage fluctuation events and locating potential physical sources.

[0112] This application extracts detailed and discriminative features from multiple dimensions and data sources, including the timing characteristics of voltage, the operating status of the UPS transformer, the load behavior of sensitive equipment, and the physical structure of the power supply network. This ensures that the unique "fingerprints" of voltage fluctuation events caused by different physical sources can be effectively reflected. For example, poor line contact may lead to a voltage drop with a specific transient duration, while distributed electromagnetic interference sources may exhibit an increase in specific harmonic content. This multi-dimensional feature extraction provides a rich and accurate data foundation for subsequent fault signature construction and classification. This application obtains a more comprehensive and refined feature set, which has higher sensitivity and specificity for distinguishing different types of physical sources. For example, by combining the timing characteristics of voltage such as mean, variance, and peak value with the voltage output deviation and current imbalance of the UPS transformer, it is possible to more accurately determine whether voltage fluctuations originate from external interference, line problems, or internal faults of the UPS transformer. Simultaneously, introducing the load rate and power consumption change rate of sensitive equipment helps identify voltage fluctuations caused by the equipment's own operation, while the power supply path length and the number of connection nodes provide crucial spatial information for locating the physical source.

[0113] Furthermore, the steps to assemble the fault signature include:

[0114] Weights are assigned based on the discriminative power of the extracted features against different physical sources.

[0115] Based on the assigned weights, the extracted features are weighted and combined to form a fault signature.

[0116] Specifically, the extracted features refer to parameters used to characterize the power supply system status and potential fault modes, obtained from time-series data identified in microscopic voltage quality data, UPS transformer operating data, the current operating load status of sensitive equipment, and acquired physical topology information. For example, these features may include time-series data features such as voltage mean, variance, peak value, harmonic content, and transient duration; UPS transformer operating data features such as voltage output deviation and current imbalance; current operating load status features of sensitive equipment such as load rate and power consumption change rate; and physical topology information features such as power supply path length and number of connection nodes. The discriminability of different physical sources can be understood as the effectiveness or importance of each extracted feature in distinguishing different types of physical sources (e.g., poor line contact, local impedance anomalies, or distributed electromagnetic interference sources). Specifically, the correlation or discriminative ability between each feature and a specific physical source can be evaluated through statistical analysis, machine learning methods, or expert experience. For example, some features may be highly sensitive in identifying poor line contact, while others may be more indicative of distributed electromagnetic interference sources.

[0117] In practical applications, weighting refers to assigning a numerical value to each feature based on its discriminative power against different physical sources, reflecting the relative importance of that feature in the final fault signature. Features with high discriminative power are assigned higher weights to enhance their influence in the fault signature; features with low discriminative power are assigned lower weights to avoid interfering with or diluting the fault signature. Weighting can be based on preset rules, machine learning model training results, or dynamic optimization algorithms. Weighted combination refers to integrating the weighted feature values ​​according to a specific mathematical model to generate a comprehensive numerical value or vector, i.e., the fault signature. For example, linear weighted sums, nonlinear function combinations, or other multi-dimensional feature fusion techniques can be used. Through weighted combination, it is ensured that those features more critical to distinguishing physical sources dominate the fault signature, thus enabling the fault signature to more accurately reflect the essence of the power supply problem.

[0118] This application addresses the issue of insufficient discriminative power in fault signatures caused by simple feature combinations by introducing a feature weighting allocation mechanism during fault signature generation. By evaluating the ability of each extracted feature to distinguish different physical sources, it identifies which features are more critical for diagnosing specific types of faults. For example, some voltage transient features may be more sensitive to identifying poor line contact, while other current imbalance features may be more effective at identifying local impedance anomalies. By assigning greater weights to these more discriminative features, their influence in the final fault signature is enhanced. Therefore, during weighted combination, features that better reflect specific physical sources will dominate the formation of the fault signature, enabling the generated fault signature to more accurately and effectively characterize and distinguish different physical sources. This avoids information redundancy or dilution of key information that might result from treating all features equally, significantly improving the accuracy of fault diagnosis. By fully considering the discriminative power of each feature against different physical sources during fault signature generation, the fault signature becomes more targeted and effective. This can significantly improve the accuracy and efficiency of identifying the physical root causes of power supply problems, reduce misjudgments and omissions, and thus provide a more reliable guarantee for the stable operation of data center power supply systems.

[0119] Secondly, see Figure 2 This application also discloses a UPS transformer voltage balance control system, which includes:

[0120] Communication module 210 is used to establish a communication link between the UPS transformer and the power modules of sensitive equipment in the data center;

[0121] The receiving module 220 is used to receive microscopic voltage quality data generated by the power supply module monitoring minute transient fluctuations in its input voltage and feeding back the data.

[0122] Evaluation module 230 is used to evaluate the impact of power supply quality on sensitive equipment based on micro voltage quality data and obtain evaluation results;

[0123] The adjustment module 240 is used to dynamically adjust the target margin of the output voltage of each phase of the UPS transformer based on the evaluation results.

[0124] The control module 250 is used to control the output voltage of the UPS transformer according to the adjusted output voltage target margin.

[0125] The core of this application lies in ensuring stable power supply to sensitive equipment within a data center through a refined voltage quality monitoring and dynamic adjustment mechanism. Compared to existing technologies, the innovation of this application lies in extending the "tentacles" of voltage quality monitoring to the sensitive equipment end, achieving end-to-end voltage quality perception and control from the "source" to the "load." Traditional methods often only perform macroscopic voltage balance control at the UPS output end, and are powerless against localized, microscopic voltage fluctuations generated in the complex power supply network within a data center. This application can identify minute transient fluctuations that traditional UPS systems cannot detect, which have cumulative damage to sensitive equipment. For example, when a server power module reports frequent minor drops in its input voltage, this application can respond immediately, dynamically adjusting the target margin of the output voltage of the corresponding phase of the UPS transformer and performing more precise compensation, thereby effectively suppressing these microscopic fluctuations. This refined and dynamic control strategy significantly improves the stability and reliability of data center power supply, reduces performance degradation and failures of sensitive equipment due to hidden power quality problems, thereby reducing maintenance costs and improving the overall operational efficiency of the data center.

[0126] The above description is merely an embodiment of this application and is not intended to limit the scope of protection of this application. Various modifications and variations can be made to this application by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method of UPS transformer voltage balancing control, characterized by, The method comprises the following steps: establishing a communication link between the UPS transformer and the power supply module of the sensitive equipment in the data center; receiving micro-transient fluctuations in the input voltage monitored by the power supply module and feeding back the generated micro-voltage quality data; evaluating the impact of power supply quality on the sensitive equipment according to the micro-voltage quality data, and obtaining an evaluation result; dynamically adjusting the target margin of the output voltage of each phase of the UPS transformer according to the evaluation result; controlling the output voltage of the UPS transformer according to the adjusted output voltage target margin; The method further comprises the following steps: After controlling the output voltage of the UPS transformer, analyzing the micro-voltage quality data and the UPS transformer's own operation data to identify power supply problems and generate diagnostic information; The step of analyzing the micro-voltage quality data and the UPS transformer's own operation data to identify power supply problems and generate diagnostic information comprises: obtaining the UPS transformer's own operation data and the current operating load of the sensitive equipment; obtaining the physical topology information of the sensitive equipment in the data center, which includes the power supply path between the sensitive equipment and the UPS transformer, the physical location of the sensitive equipment, and the rack information to which the sensitive equipment belongs; In the micro-voltage quality data, identify the time series data related to the voltage fluctuation event that occurs under a certain load condition of the sensitive equipment; correlate the identified time series data with the UPS transformer's own operation data, the current operating load state of the sensitive equipment, and the physical topology information to infer the physical root cause of the voltage fluctuation event, which includes poor line contact, local impedance abnormalities, or distributed electromagnetic interference sources, and determine the physical location of the physical root cause; generate diagnostic information containing the physical root cause and the physical location.

2. The method of claim 1, wherein, The step of dynamically adjusting the target margin of the output voltage of each phase of the UPS transformer according to the evaluation result comprises: obtaining the output current data of each phase of the UPS transformer; According to the output current data, identify the transient load imbalance between each phase; According to the evaluation result and the transient load imbalance, adjust the target margin of the output voltage of each phase.

3. The method of claim 2, wherein, The step of dynamically adjusting the target margin of the output voltage of each phase of the UPS transformer according to the evaluation result further comprises: monitoring the balance of the output voltage of each phase, and compensating for the adjustment of the target margin of the output voltage of each phase according to the balance.

4. The method of claim 1, wherein, The step of evaluating the impact of power supply quality on the sensitive equipment according to the micro-voltage quality data, and obtaining an evaluation result, comprises: receiving the device type identifier and the current operating load state sent by the power supply module; obtaining the voltage fluctuation tolerance parameters corresponding to the device type identifier and the current operating load state; According to the micro-voltage quality data and the voltage fluctuation tolerance parameters, evaluate the impact of power supply quality on the sensitive equipment, and obtain an evaluation result.

5. The method of claim 1, wherein, The step of correlating the identified time-series data with the UPS transformer's own operation data, the current operation load status of the sensitive equipment, and the physical topology information to infer the physical root cause of the voltage fluctuation event comprises: extracting features for distinguishing the physical root cause from the identified time-series data, the UPS transformer's own operation data, the current operation load status of the sensitive equipment, and the obtained physical topology information, and combining them into a fault signature; classifying the fault signature according to a preset classification rule to obtain a preliminary candidate physical root cause; logically excluding the preliminary candidate physical root cause according to the physical topology information to determine the physical root cause.

6. The method of claim 5, wherein, The step of extracting features for distinguishing the physical root cause from the identified time-series data, the UPS transformer's own operation data, the current operation load status of the sensitive equipment, and the obtained physical topology information comprises: extracting the mean, variance, peak value, harmonic content, and transient duration of the voltage from the identified time-series data; extracting the voltage output deviation and current imbalance from the UPS transformer's own operation data; extracting the load rate and power consumption change rate from the current operation load status of the sensitive equipment; extracting the power supply path length and the number of connected nodes from the obtained physical topology information.

7. The method of claim 5, wherein, The step of combining into a fault signature comprises: assigning weights according to the distinguishing degree of the extracted features for different physical root causes; weighting and combining the extracted features according to the assigned weights to form the fault signature.

8. A system for performing a method of voltage balancing control of a UPS transformer according to any one of claims 1 to 7, characterized by The system comprises: a communication module for establishing a communication link between the UPS transformer and the power supply module of the sensitive equipment in the data center; a receiving module for receiving micro voltage quality data generated by the power supply module monitoring the micro transient fluctuation of the input voltage and feeding back; an evaluation module for evaluating the impact of power supply quality on the sensitive equipment according to the micro voltage quality data to obtain an evaluation result; an adjustment module for dynamically adjusting the target margin of the output voltage of each phase of the UPS transformer according to the evaluation result; a control module for controlling the output voltage of the UPS transformer according to the adjusted target margin of the output voltage.

Images