A data center standby power energy storage optimization configuration energy-saving control method

Abstract

The application belongs to the technical field of energy storage optimization and energy-saving control, and specifically discloses a data center standby power energy storage optimization configuration energy-saving control method, which comprises the following steps: determining the grid disturbance intensity according to the grid-connected point voltage and frequency, and determining the minimum safety boundary of the discharge termination threshold in combination with the IT load standby power duration requirement; collecting photovoltaic power, identifying the change trend through linear regression, and determining the reference discharge termination threshold in combination with the minimum safety boundary; detecting the IT load power surge event, and if the event is detected, generating a pulse compensation amount according to the surge amplitude and duration, taking the sum of the reference threshold and the compensation amount as the final discharge termination threshold, or taking the reference threshold as the final threshold; when the energy storage SOC decreases to the final threshold, cutting off the non-IT load and retaining the IT load power supply. The application can improve the green power consumption capacity under the premise of guaranteeing the standby power reliability by dynamically adapting the grid disturbance, photovoltaic fluctuation and load pulse, and can avoid the power failure of the IT load.

Description

Technical Field

[0001] This invention belongs to the field of energy storage optimization and energy-saving control technology, and specifically relates to an energy-saving control method for optimizing the configuration of backup power energy storage in data centers. Background Technology

[0002] In photovoltaic-storage-charging microgrids, a fixed discharge termination threshold is typically set to balance daily economic dispatch with off-grid backup power reliability. However, in actual operation, factors such as sudden photovoltaic fluctuations and load pulses can cause the fixed lower limit to fail to respond in time, resulting in insufficient off-grid power supply duration or wasted dispatch capacity.

[0003] Currently, to address the discrepancy between the lower limit of the fixed discharge termination threshold and actual operating conditions, two main approaches are used: First, a large reserve power capacity is reserved to cover the uncertainties caused by sudden PV drops or load pulses. Second, after off-grid switching, non-critical loads are switched off tier by tier according to load priority, extending the power supply time for critical loads.

[0004] While reserving a margin can improve off-grid reliability, it significantly reduces the available capacity for daily dispatch and decreases the capacity for green energy absorption. For the tiered load shedding method, the shedding threshold is still based on a fixed discharge termination threshold, which cannot be dynamically adjusted according to real-time photovoltaic output, leading to early or late shedding. When the off-grid power supply duration is insufficient, it is difficult to determine whether the problem stems from insufficient configured capacity or from factors such as sudden photovoltaic drops, load pulses, or battery aging rendering the fixed lower limit inapplicable. Summary of the Invention

[0005] In view of this, in order to solve the above problems, an energy-saving control method for optimizing the configuration of backup power storage in data centers is proposed.

[0006] The objective of this invention can be achieved through the following technical solution: This invention provides a data center backup power storage energy-saving control method, which includes: determining the grid disturbance intensity based on the collected voltage and frequency values ​​of the grid connection point, and determining the minimum safe boundary of the discharge termination threshold by combining the grid disturbance intensity and the pre-installation power-on duration requirements of the IT load.

[0007] Photovoltaic power is collected in real time, and a photovoltaic power sequence is captured according to a preset capture time window. The photovoltaic power change trend is identified through linear regression, and the benchmark discharge termination threshold is determined based on the photovoltaic power change trend and the minimum safety boundary.

[0008] Based on real-time total power detection of IT load, if a sudden increase in IT load power is detected, a pulse compensation amount is generated according to the magnitude and duration of the sudden increase. The final discharge termination threshold is the sum of the reference discharge termination threshold and the pulse compensation amount. If no sudden increase is detected or the sudden increase subsides, the reference discharge termination threshold is used as the final discharge termination threshold.

[0009] The final discharge termination threshold is used as the cutoff condition for energy storage discharge. When the energy storage SOC drops to this threshold, non-IT loads are cut off while IT loads are kept powered.

[0010] Compared with the prior art, the beneficial effects of the present invention are as follows: (1) The present invention determines the minimum safe boundary of the discharge termination threshold based on the grid disturbance intensity and the backup power duration requirement, which solves the problem that the fixed threshold cannot quantify the incremental backup power demand when the grid is abnormal such as voltage sag, and realizes the adaptive adjustment of the grid quality to the energy storage discharge boundary, thereby avoiding the off-grid supply gap caused by the lack of disturbance perception.

[0011] (2) Based on the photovoltaic change trend identified by linear regression, this invention dynamically adjusts the benchmark discharge termination threshold according to the rising, falling or stable trend, which solves the contradiction of overestimating the available backup power energy when the photovoltaic power drops sharply, resulting in insufficient supply, and underestimating the dispatchable capacity when the photovoltaic power rises sharply, resulting in a decrease in the green electricity consumption capacity. It realizes the active release of the dispatch capacity occupied by redundancy under the premise of ensuring the reliability of backup power, thereby improving the green electricity consumption capacity.

[0012] (3) Based on the real-time total power detection of IT load, the present invention generates a pulse compensation amount according to the magnitude and duration of the surge and superimposes it on the reference discharge termination threshold to form the final discharge termination threshold. This avoids the occurrence of premature energy storage discharge cutoff and forced power outage of IT load caused by instantaneous load pulse impact. It can reserve corresponding energy at the moment of impact and automatically restore to the reference threshold after the impact subsides, avoiding frequent adjustment of daily scheduling capacity due to short pulse.

[0013] (4) By using the final discharge termination threshold as the cutoff condition for energy storage discharge, this invention can simultaneously respond to three types of dynamic factors: grid disturbance, photovoltaic fluctuation, and load pulse. When the off-grid power supply is insufficient, it can distinguish whether it is due to insufficient configuration capacity or the above disturbances. This provides a dynamic threshold benchmark for analyzing whether the insufficient off-grid power supply is due to insufficient configuration capacity or external disturbances, thereby assisting operation and maintenance personnel in making root cause judgments. Attached Figure Description

[0014] Figure 1 This is a schematic diagram of the overall implementation process of the present invention;

[0015] Figure 2 This is a schematic diagram of the process for determining the intensity of power grid disturbances according to the present invention;

[0016] Figure 3 This is a schematic diagram of the IT load power surge event detection process of the present invention. Detailed Implementation

[0017] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0018] Currently, a fixed discharge termination threshold is typically set for backup power storage in data centers. When photovoltaic power suddenly drops or rises due to weather changes, or when IT load experiences instantaneous high-power pulses, the actual available backup power energy of the energy storage terminal fluctuates dynamically. The fixed threshold cannot detect these physical changes and may misjudge normal energy fluctuations as insufficient backup power capacity or improper scheduling. This leads to insufficient off-grid power supply duration, wasted daily capacity, early or late power cut-offs, and difficulty in tracing the causes of insufficient power supply.

[0019] Based on this, this invention discloses an energy-saving control method for optimizing the configuration of backup power and energy storage in data centers, which can effectively and dynamically adapt to grid disturbances, photovoltaic fluctuations and load pulses, and improve the green electricity consumption capacity while ensuring the reliability of backup power.

[0020] Please refer to details. Figure 1 As shown, the present invention provides an energy-saving control method for optimizing backup power storage configuration in data centers. The method includes: S1, determining the grid disturbance intensity based on the collected voltage and frequency values ​​of the grid connection point, and determining the minimum safe boundary of the discharge termination threshold by combining the grid disturbance intensity and the pre-installation power-off duration requirements of the IT load.

[0021] Among them, the voltage value refers to the instantaneous value of the three-phase voltage at the grid connection point, and the frequency value refers to the grid frequency at the grid connection point. The voltage value can be obtained by the voltage transformer installed at the grid connection point, and the frequency value can be directly obtained by the frequency meter (such as a digital frequency meter) installed at the grid connection point.

[0022] It should be noted that when a short-circuit fault or heavy load switching occurs in the power grid, a voltage dip will occur. At this time, the ability of energy storage to draw charging power from the grid decreases, and it may even be forced to discharge, thereby accelerating the consumption of backup energy. When the active power of the power grid is unbalanced, the frequency will deviate from the rated value. A significant deviation (e.g., below 49.5Hz or above 50.5Hz) usually indicates that the power grid may enter islanded operation, requiring additional backup energy reserves. Therefore, this embodiment of the invention, after obtaining the voltage and frequency values, combines both to quantify the intensity of power grid disturbances.

[0023] Specifically, please refer to Figure 2 As shown, the specific quantification process of the power grid disturbance intensity is as follows: A1. Real-time acquisition of instantaneous voltage and frequency values ​​at the grid connection point, and construction of voltage data sequence and frequency data sequence respectively.

[0024] A2. The duration during which the instantaneous voltage value is lower than the rated voltage at the grid connection point is recorded as the voltage sag duration.

[0025] A3. If the duration of the voltage dip exceeds the first duration threshold, the grid disturbance intensity is determined to be the preset maximum value. The first duration threshold can be set according to the grid stability requirements. The higher the requirements, the smaller the first duration threshold. In this invention, it is preferably set to 0.5 seconds. The preset maximum value corresponds to a normalization result of 1.

[0026] A4. If the duration of the voltage dip does not exceed the first duration threshold, calculate the absolute value of the deviation between each instantaneous voltage value and the rated voltage at the grid connection point from the instantaneous voltage value sequence, and take the maximum value as the voltage deviation. At the same time, calculate the absolute value of the deviation between each instantaneous frequency value and the rated frequency at the grid connection point from the instantaneous frequency value sequence, and take the maximum value as the frequency deviation.

[0027] The rated voltage and rated frequency of the grid connection point refer to the standard voltage and standard frequency values ​​specified by the power grid design when the grid connection point is operating normally, which can be directly obtained from the equipment nameplate of the grid connection point.

[0028] Preferably, before calculating the voltage deviation and frequency deviation, the acquired instantaneous voltage value sequence and instantaneous frequency value sequence can be filtered, such as by moving average filtering or median filtering, to eliminate the influence of instantaneous noise spikes on the maximum value.

[0029] A5. Normalize the voltage deviation and frequency deviation separately. For example, divide the frequency deviation by the rated frequency and the voltage deviation by the rated voltage to obtain the normalized frequency deviation rate and voltage deviation rate. If the normalized frequency deviation rate and voltage deviation rate are greater than or equal to 1, assign a value of 1 to the normalized frequency deviation rate and voltage deviation rate. Then, perform a linear weighted summation of the normalized frequency deviation rate and voltage deviation rate, where the sum of the weights is 1. As an example, the weights can each be 0.5. Use the summation result as the power grid disturbance intensity.

[0030] It should be noted that voltage deviation and frequency deviation are both positively correlated with the intensity of power grid disturbance, that is, they exhibit an approximately linear relationship. Therefore, the embodiments of the present invention use a linear weighted summation method for fusion.

[0031] Understandably, grid disturbance intensity is used to characterize the degree of abnormality of the current grid operating state at the grid connection point relative to the rated operating condition. Its value ranges from [0, 1]. The greater the grid disturbance intensity, the more unstable the grid is, and the more backup power the energy storage terminal needs to reserve to cope with possible off-grid events. This causes the minimum safety boundary of the subsequent final discharge termination threshold to increase with the increase of disturbance intensity, realizing adaptive adjustment where the more disturbed the grid, the more sufficient the backup power.

[0032] Considering that the basic backup power energy can meet the ideal demand when there is no disturbance, but the grid disturbance in actual operation will lead to additional backup power demand, the embodiment of the present invention further determines the minimum safe boundary of the discharge termination threshold based on the grid disturbance intensity. The specific process is as follows: B1. Extract the pre-installation power duration threshold from the pre-installation power duration requirement, take the current time as the integration end point, and take the time obtained by subtracting the pre-installation power duration threshold from the current time as the integration start point. Based on the integration start point and integration end point, construct the integration time window.

[0033] Understandably, the pre-set power-on duration requirement includes, but is not limited to, the pre-set power-on duration threshold that the data center needs to maintain power supply to IT loads after a mains power failure. The pre-set power-on duration threshold is a design target parameter pre-set by the data center operator based on business needs, such as the requirement that IT loads must maintain power supply for 30 minutes after a mains power failure.

[0034] It should be noted that if the system has been running for less than the pre-set power-on time threshold, the integration start point is set to the system start time, and the integration window length is equal to the running time.

[0035] B2. Collect the total power value of IT load in real time, integrate it within the integration time window, and use the integration result as the basic backup power energy.

[0036] B3. Add the grid disturbance intensity to 1 to obtain the energy adjustment coefficient, and multiply the basic backup power energy by the energy adjustment coefficient to obtain the adjusted target backup power energy.

[0037] B4. Obtain the total energy capacity of the energy storage terminal, divide the adjusted target backup energy by the total energy capacity, and use the ratio as the minimum safe boundary for the discharge termination threshold. The total energy capacity of the energy storage terminal refers to the total electrical energy that the energy storage system can release in a fully charged state, which can be read by the battery management system (BMS) of the energy storage system.

[0038] Understandably, the minimum safety boundary refers to the minimum state of charge (SOC) that an energy storage terminal must reserve to meet the preset backup power duration requirements for IT loads, taking into account the current grid disturbance intensity. This boundary is expressed as a percentage and serves as the basic reference value for further adjustments to the discharge termination threshold based on photovoltaic power variation trends and load surge events. The base backup power reflects the ideal demand under undisturbed conditions, while the target backup power increases linearly with the disturbance intensity, ensuring sufficient backup power reserve even when the grid is highly unstable.

[0039] This step solves the problem that a fixed threshold cannot quantify the increase in backup power demand during grid anomalies such as voltage dips by determining the grid disturbance intensity and the minimum safe boundary of the discharge termination threshold based on the grid disturbance intensity and backup power duration requirements. It realizes the adaptive adjustment of the grid quality as the energy storage discharge boundary by dynamically quantifying the grid quality, thereby avoiding the off-grid power supply gap caused by the lack of disturbance perception.

[0040] S2. Real-time acquisition of photovoltaic power, capturing photovoltaic power sequence according to preset capture time window, identifying photovoltaic power change trend through linear regression, and determining the benchmark discharge termination threshold based on photovoltaic power change trend and minimum safety boundary.

[0041] The photovoltaic power can be acquired in real time by a photovoltaic power sensor (such as a Hall power sensor) installed on the DC output side of the photovoltaic array.

[0042] Considering the intermittent and fluctuating nature of photovoltaic (PV) power, when PV power is on an upward trend, it means that PV power generation will exceed the current level in the near future. Energy storage can obtain additional charging energy, thus allowing for a suitable reduction in the discharge termination threshold to release more daily dispatch capacity while ensuring backup power. When PV power is on a downward trend, future power generation will be lower than the current level, reducing the supplementary energy available to energy storage. Therefore, it is necessary to raise the discharge termination threshold in advance to reserve more backup power.

[0043] Based on this, after obtaining the minimum safety boundary, the present invention further determines the reference discharge termination threshold by identifying the photovoltaic power change trend.

[0044] Specifically, the process of identifying the photovoltaic power change trend includes: real-time acquisition of the output power of the photovoltaic array, extracting the photovoltaic power value within a recent period according to a preset interception time window (e.g., 30 seconds), and constructing a photovoltaic power sequence.

[0045] Perform linear regression on the sequence to obtain the slope of the regression line.

[0046] If the absolute value of the slope is greater than the preset dead zone threshold, the trend of photovoltaic power change is determined according to the sign of the slope, where a positive sign corresponds to an upward trend and a negative sign corresponds to a downward trend. The preset dead zone threshold is set to 0.01kW / s by default in this embodiment of the invention.

[0047] If the absolute value of the slope is less than or equal to the preset dead zone threshold, the photovoltaic power change trend is judged to be stable.

[0048] Understandably, the slope of the linear regression reflects the average rate of change of photovoltaic power within the time window, and the dead zone threshold is used to filter out small fluctuations caused by rapid cloud movement or measurement noise, so as to avoid misjudging a steady state as a trend change.

[0049] Furthermore, based on the identified photovoltaic power change trend, the determination process for the baseline discharge termination threshold is as follows: S21, estimate the net charging energy that the photovoltaic system can provide relative to the current time within a future preset time window, denoted as the additional charging energy. The length of the future preset time window is the same as the preset charging duration.

[0050] The estimation steps for additional charging energy are as follows: Starting with the current photovoltaic power, the slope obtained from linear regression is used as the rate of change to extrapolate the photovoltaic power change curve over the future pre-installed charging period. This curve is then integrated to obtain the estimated photovoltaic power generation. Next, the product of the current photovoltaic power and the pre-installed charging period is calculated. Subtracting this product from the estimated photovoltaic power generation yields the net charging energy that can be provided.

[0051] S22. The ratio of the additional charging energy to the total energy capacity of the energy storage terminal is used as the additional SOC offset. Then, the reference discharge termination threshold is determined according to the photovoltaic change trend: when the photovoltaic power change trend is upward, the minimum safety boundary is subtracted from the additional SOC offset, and the result is used as the reference discharge termination threshold.

[0052] When the photovoltaic power change trend is downward, the absolute value of the additional SOC offset is added to the minimum safety boundary, and the result is used as the reference discharge termination threshold.

[0053] When the photovoltaic power change trend is stable, the minimum safety boundary is output as the reference discharge termination threshold.

[0054] Understandably, the additional SOC offset is used to characterize the percentage impact of future photovoltaic power variation trends on energy storage backup capacity. This offset can be positive or negative, and its absolute value reflects the proportion of net charging energy (or net deficit energy) that photovoltaics can provide relative to the current moment within a future time window equal to the pre-installed charging duration, relative to the total energy storage capacity.

[0055] It should be noted that when photovoltaic (PV) power is on an upward trend, the additional charging energy available in the future allows for deeper discharge of energy storage (lowering the threshold), thereby releasing more daily dispatch capacity and improving the consumption of green electricity. When PV power is on a downward trend, future energy gaps need to be reserved in advance (raising the threshold) to ensure sufficient backup power duration when the grid is off-grid. When PV power is stable, no adjustments are needed.

[0056] This step is based on the photovoltaic change trend identified by linear regression. According to the rising, falling or stable trend, the benchmark discharge termination threshold is dynamically adjusted. This realizes the proactive release of the scheduling capacity occupied by redundancy while ensuring the reliability of backup power, thereby improving the green power consumption capacity.

[0057] S3. Based on the real-time total power detection of IT load, if a sudden increase in IT load power is detected, a pulse compensation amount is generated according to the magnitude and duration of the sudden increase. The sum of the reference discharge termination threshold and the pulse compensation amount is used as the final discharge termination threshold. If no sudden increase is detected or the sudden increase subsides, the reference discharge termination threshold is used as the final discharge termination threshold.

[0058] The real-time total power of the IT load can be obtained by reading the smart meter data of the IT load branch in the data center power distribution cabinet, or by collecting data in real time through a power sensor (such as a Hall power sensor) installed on the IT load power supply line.

[0059] When IT loads (such as servers and network devices) start up or execute sudden computing tasks, they generate instantaneous high-power surges. These surges can cause a rapid drop in the State of Charge (SOC) of the energy storage. If the discharge continues at the baseline threshold, the discharge may be prematurely triggered before the surge ends, causing the IT load to lose power. Therefore, this embodiment of the invention introduces power surge event detection to temporarily increase the discharge termination threshold and reserve energy for surges.

[0060] Specifically, please refer to Figure 3 As shown, the detection process of the IT load power surge event is as follows: S31, construct a power sampling sequence from the real-time total power of the IT load, perform first-order difference calculation on the sequence to obtain the power change sequence between adjacent time moments, and divide the current value in the power change sequence by the sampling period to obtain the power change rate. The sampling period is preferably 0.1 seconds.

[0061] S32. Compare the current value in the power change sequence with the preset pulse amplitude threshold, and compare the power change rate with the preset rate threshold.

[0062] S33. If the power change exceeds the pulse amplitude threshold and the power change rate exceeds the rate threshold, then an IT load power surge event is triggered, and the current moment is marked as the surge start point.

[0063] S34. After the initial point of the surge, continue to monitor the power change in real time. When the power change is lower than the pulse amplitude threshold for several consecutive sampling periods, the surge event is determined to have ended, and the number of consecutive periods is at least three.

[0064] Understandably, the pulse amplitude threshold is typically set to 20% to 30% of the total rated power of the IT load. In this invention, it is preferably set to 20% of the total rated power of the IT load. For example, if the total rated power of the data center IT load is 500kW, then the pulse amplitude threshold is 100kW.

[0065] The rate threshold is set according to the technical specifications of the IT equipment power supply. For example, the current change rate of a server power supply is typically 0.25. If the output voltage is 12V, then the corresponding power change rate is For the overall IT load of the data center, the rate threshold can be set to 5% to 10% of the total rated power of the IT load per millisecond. If specific parameters cannot be obtained, the default value can be set to... (Assuming a total rated power of 500kW).

[0066] It should be noted that employing both power change magnitude and power change rate as detection conditions can accurately distinguish between genuine load surges and slow power fluctuations or measurement noise. The power change magnitude reflects the amplitude of the surge, while the power change rate reflects the steepness of the surge. When both exceed a threshold simultaneously, an IT load power surge event is triggered, effectively avoiding false detections.

[0067] After identifying the surge event, the amplitude and duration of the surge event are further calculated: After the surge initiation point, the power sampling value continues to be monitored in real time, and the maximum value of the power sampling value is recorded. The difference between this maximum value and the power value at the surge initiation point is taken as the surge amplitude, and the time length between the surge initiation point and the surge event termination determination time is taken as the duration.

[0068] The surge magnitude reflects the instantaneous power increment of the load impact, and the duration reflects the duration of the impact. The product of the two is the total energy consumed by the impact, i.e., the surge energy of the surge event. Based on this, dividing the surge energy by the total energy capacity of the energy storage terminal yields the pulse compensation amount.

[0069] When multiple load surge events occur, the surge energy of each event needs to be calculated separately. Then, the sum of the surge energies is added together and divided by the total energy capacity of the energy storage terminal. The result is used as the final pulse compensation amount. This final pulse compensation amount is applied from the start of the first surge event and remains constant until the end of the last surge event. After the last surge event ends, the pulse compensation amount is directly set to zero.

[0070] It is important to note that the pulse compensation amount only takes effect during the duration of the surge event and is immediately revoked after the event ends, ensuring that energy storage does not maintain an excessively high discharge cutoff threshold for an extended period due to historical pulses, thereby not affecting daily dispatch capacity.

[0071] In addition, if the interval between two consecutive load surge events exceeds the preset interval, the pulse compensation amount will be set to zero after the previous event ends. After being set to zero, subsequent detected load surge events will be processed as a new event sequence, that is, the pulse compensation amount corresponding to the new event will be recalculated, and the corresponding application and cancellation will be performed according to the above rules.

[0072] Understandably, the preset interval duration is related to the preset power-on duration. As one implementation, the preset interval duration is set to half of the preset power-on duration. For example, if the preset power-on duration is 30 minutes, then the preset interval duration is 15 minutes.

[0073] It should be noted that the energy accumulation of multiple surge events can cope with continuous or overlapping load impacts, avoiding backup power gaps due to insufficient compensation from a single event. The constant maintenance of the pulse compensation amount and the direct zeroing method ensure that the threshold remains elevated during the impact period and quickly recovers to the reference threshold after the impact completely ends, avoiding long-term excessive reserve capacity. Furthermore, the design of resetting to zero after the interval exceeds the limit prevents erroneous energy accumulation due to two independent impact events being too far apart in time, further improving the accuracy of compensation.

[0074] S4. Use the final discharge termination threshold as the cutoff condition for energy storage discharge. When the energy storage SOC drops to this threshold, cut off non-IT loads and retain power supply to IT loads.

[0075] Specifically, the state of charge (SOC) of the energy storage system is monitored in real time (which can be read through the battery management system (BMS). When the SOC drops to the final discharge termination threshold, non-IT loads, such as air conditioners, lighting, and non-critical auxiliary equipment, are automatically cut off, while only IT loads, such as servers, network equipment, and storage devices, continue to be powered.

[0076] It should be noted that the purpose of cutting off non-IT loads is to prioritize the off-grid operation time of IT loads when energy storage capacity is limited.

[0077] Furthermore, since the final discharge termination threshold incorporates three dynamic factors—grid disturbances, photovoltaic trends, and load pulses—this cutoff condition can adaptively match real-time operating conditions. It can ensure the pre-set power-on time while avoiding the waste of scheduling capacity caused by prematurely cutting off non-IT loads, and also avoid the power outage of IT loads caused by cutting off too late.

[0078] The above content is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described, or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined by the present invention, and all such modifications and additions should fall within the protection scope of the present invention.

Claims

1. A method for optimizing and controlling energy-saving configuration of backup power storage in a data center, characterized in that, The method includes: The grid disturbance intensity is determined based on the collected voltage and frequency values ​​at the grid connection point. The minimum safe boundary for the discharge termination threshold is then determined by combining the grid disturbance intensity with the pre-equipment energization time requirement of the IT load. Real-time acquisition of photovoltaic power, capturing photovoltaic power sequence according to preset capture time window, identifying photovoltaic power change trend through linear regression, and determining benchmark discharge termination threshold based on photovoltaic power change trend and minimum safety boundary; Based on the real-time total power detection of IT load, if a sudden increase in IT load power is detected, a pulse compensation amount is generated according to the magnitude and duration of the sudden increase. The final discharge termination threshold is the sum of the reference discharge termination threshold and the pulse compensation amount. If no sudden increase is detected or the sudden increase has subsided, the reference discharge termination threshold is used as the final discharge termination threshold. The final discharge termination threshold is used as the cutoff condition for energy storage discharge. When the energy storage SOC drops to this threshold, non-IT loads are cut off while IT loads are kept powered.

2. The energy-saving control method for optimized configuration of backup power storage in a data center as described in claim 1, characterized in that: The specific process for determining the intensity of the power grid disturbance includes: Real-time acquisition of instantaneous voltage and frequency values ​​at grid connection points to construct voltage and frequency data sequences; The duration during which the instantaneous voltage value is lower than the rated voltage at the grid connection point is recorded as the voltage sag duration. When the duration of a voltage dip exceeds the first duration threshold, the grid disturbance intensity is set to the preset maximum value. Otherwise, the maximum absolute value of the deviation between each instantaneous voltage value and the rated voltage at the grid connection point is taken from the instantaneous voltage value sequence as the voltage deviation, and the maximum absolute value of the deviation between each instantaneous frequency value and the rated frequency at the grid connection point is taken from the instantaneous frequency value sequence as the frequency deviation. The frequency deviation and voltage deviation are normalized, and the normalized results are linearly weighted and calculated. The calculation results are used as the power grid disturbance intensity.

3. The energy-saving control method for optimized configuration of backup power storage in a data center as described in claim 1, characterized in that: The specific process for determining the minimum safe boundary of the discharge termination threshold includes: Extract the pre-equipment time threshold from the pre-equipment time requirement, take the current time as the integration end point, and take the time obtained by subtracting the pre-equipment time threshold from the current time as the integration start point. Based on the integration start point and integration end point, construct the integration time window. The total power value of IT load is collected in real time and integrated within the integration time window. The integration result is used as the basic backup power energy. The sum of the grid disturbance intensity and 1 is used as the energy adjustment coefficient, and the base backup power energy is multiplied by the energy adjustment coefficient to obtain the adjusted target backup power energy. Obtain the total energy capacity of the energy storage terminal, divide the adjusted target backup power energy by the total energy capacity, and use the ratio as the minimum safe boundary for the discharge termination threshold.

4. The energy-saving control method for optimized configuration of backup power storage in a data center as described in claim 1, characterized in that: The specific process for identifying the photovoltaic power change trend includes: Based on the extracted photovoltaic power values, a photovoltaic power sequence is constructed, and a linear regression is performed on the photovoltaic power sequence to obtain the slope of the regression line; If the absolute value of the slope is greater than the preset dead zone threshold, the trend of photovoltaic power change is determined by the sign of the slope, which indicates whether it is rising or falling. A positive sign corresponds to rising and a negative sign corresponds to falling. If the absolute value of the slope is less than or equal to the preset dead zone threshold, the photovoltaic power change trend is judged to be stable.

5. The energy-saving control method for optimized configuration of backup power storage in a data center as described in claim 4, characterized in that: The specific process for determining the reference discharge termination threshold includes: Based on the trend of photovoltaic power change, the net charging energy that photovoltaics can provide relative to the current moment within a future preset time window is estimated and denoted as additional charging energy. The length of the future preset time window is the same as the preset charging duration. The ratio of the additional charging energy to the total energy capacity of the energy storage terminal is used as the additional SOC offset. When the photovoltaic power change trend is upward, the minimum safety boundary is subtracted from the additional SOC offset, and the result is used as the benchmark discharge termination threshold. When the photovoltaic power change trend is downward, the absolute value of the additional SOC offset is added to the minimum safety boundary, and the result is used as the benchmark discharge termination threshold. When the photovoltaic power change trend is stable, the minimum safety boundary is output as the reference discharge termination threshold.

6. The energy-saving control method for optimized configuration of backup power storage in a data center as described in claim 5, characterized in that: The specific steps for estimating the net charging energy that photovoltaics can provide relative to the current time within a future preset time window are as follows: Starting with the current photovoltaic power, and using the slope of the linear regression line corresponding to the photovoltaic power sequence as the rate of change, we extrapolate to obtain the photovoltaic power change curve within the future pre-installed power generation period, and integrate the change curve to obtain the estimated photovoltaic power generation. Calculate the difference between the estimated photovoltaic power generation and the product of the current photovoltaic power and the estimated charging time, and use the difference as the net charging energy that can be provided.

7. The energy-saving control method for optimized configuration of backup power storage in a data center as described in claim 1, characterized in that: The specific detection process for the IT load power surge event includes: The real-time total power of the IT load is used to construct a power sampling sequence; The power sampling sequence is subjected to first-order difference calculation to obtain the power change sequence between adjacent time points, and the current value in the power change sequence is divided by the sampling period to obtain the power change rate. The current value in the power change sequence is compared with the preset pulse amplitude threshold, and the power change rate is compared with the preset rate threshold. If the power change exceeds the pulse amplitude threshold and the power change rate exceeds the rate threshold, then an IT load power surge event is triggered, and the current moment is marked as the surge start point. The power change is monitored in real time after the start of the surge. When the power change is lower than the pulse amplitude threshold for several consecutive sampling periods, the surge event is determined to have ended.

8. The energy-saving control method for optimized configuration of backup power storage in a data center as described in claim 7, characterized in that: The magnitude and duration of the sudden increase are determined as follows: After the surge initiation point, continue to monitor the power sampling value in real time, record the maximum value of the power sampling value, and use the difference between the maximum value and the power value at the surge initiation point as the surge magnitude; The duration is defined as the time between the start point of the surge and the time when the surge event is judged to end.

9. The energy-saving control method for optimized configuration of backup power storage in a data center as described in claim 1, characterized in that: The pulse compensation amount is generated in the following way: The product of the burst magnitude and duration is taken as the burst energy of the burst event; Divide the sudden energy increase by the total energy capacity of the energy storage terminal to obtain the pulse compensation amount.

10. The energy-saving control method for optimized configuration of backup power storage in a data center as described in claim 9, characterized in that: When there are multiple load surge events, calculate the surge energy of each surge event separately, sum up the surge energies and divide by the total energy capacity of the energy storage terminal, and use the result as the final pulse compensation amount; The final pulse compensation amount is applied from the start of the first burst event and remains constant until the end of the last burst event, and is set to zero directly after the last burst event ends. If the interval between two consecutive surge events exceeds the preset interval, the pulse compensation amount will be set to zero after the previous event ends.

Images