A communication machine room power environment self-adaptive regulation and control method and system

Abstract

This invention relates to the field of communication equipment room environmental control technology, specifically to an adaptive control method and system for the power environment of a communication equipment room. The method includes: collecting the current temperature of the equipment room and the current total power of the communication equipment; subtracting the static power consumption baseline to obtain the dynamic power consumption; dividing the difference in dynamic power consumption between adjacent samples by the sampling period to obtain the rate of change; multiplying the dynamic power consumption by a feedforward mapping coefficient to obtain the baseline cooling capacity; multiplying the rate of change by a trend compensation coefficient to obtain the trend compensation amount, with a positive rate of change using a temperature rise compensation coefficient and a negative rate of change using a smaller temperature fall compensation coefficient; inputting the temperature deviation into an incremental proportional-integral-derivative controller to calculate the feedback correction amount; and combining the baseline cooling capacity, the trend compensation amount, and the feedback correction amount to generate a total cooling command and send it to the air conditioner. This invention enables cooling to actively track changes in heat dissipation, solving the problems of temperature fluctuations and high energy consumption caused by dynamic mismatch.

Description

Technical Field

[0001] This invention relates to the field of communication equipment room environmental control technology. More specifically, this invention relates to an adaptive control method and system for the power environment of a communication equipment room. Background Technology

[0002] Communication equipment rooms are the infrastructure that houses base station main equipment, switches, servers, and other communication devices. During continuous operation, these devices convert electrical energy into heat, which is then dissipated into the room. To ensure the reliability and lifespan of the communication equipment, an air conditioning system is needed to remove this heat promptly, maintaining the room temperature within a safe range. Currently, air conditioning systems in communication equipment rooms generally operate using a fixed temperature threshold control method. This method pre-sets upper and lower temperature thresholds. When the room temperature exceeds the upper threshold, cooling output increases; when it falls below the lower threshold, cooling output decreases. The air conditioning system passively responds to temperature changes between these thresholds. This method is simple to implement, has low engineering deployment costs, and can meet basic temperature maintenance requirements in scenarios with relatively stable equipment heat loads.

[0003] However, with the continuous growth of communication network traffic and the increasing power density of equipment, the dynamic fluctuation range of heat dissipation load in computer rooms is becoming increasingly larger, and the heat dissipation power consumption of communication equipment exhibits significant time-varying characteristics with traffic flow. Heat dissipation increases dramatically during peak daytime traffic periods and decreases significantly during off-peak nighttime traffic periods. Fixed threshold control methods cannot detect real-time changes in equipment heat dissipation load. When traffic surges and heat dissipation rises rapidly, the air conditioning only responds passively after the computer room temperature has exceeded the safety limit, resulting in a control lag. When traffic decreases and heat dissipation reduces, the air conditioning continues to cool at the original setpoint, leading to wasted energy.

[0004] In existing technologies, fixed temperature threshold control methods may lead to a mismatch between cooling output and real-time heat dissipation load of equipment under rapidly changing load conditions, which may result in temperature fluctuations and increased energy consumption in the computer room. Summary of the Invention

[0005] To address the technical problem that existing fixed temperature threshold control methods cannot detect real-time changes in equipment heat dissipation load, resulting in a dynamic mismatch between cooling output and real-time heat dissipation load, leading to temperature fluctuation risks and high energy consumption, this invention provides solutions in the following aspects.

[0006] In a first aspect, the present invention provides an adaptive control method for the power environment of a communication equipment room, comprising: S1, collecting the current temperature of the equipment room and the current total power of the communication equipment; acquiring a pre-established and stored static power consumption baseline, and subtracting the static power consumption baseline from the current total power to obtain dynamic power consumption; S2, dividing the difference between two adjacent samples of the dynamic power consumption by the sampling period to obtain the dynamic power consumption change rate; multiplying the dynamic power consumption by a feedforward mapping coefficient to obtain a reference cooling capacity; S3, multiplying the dynamic power consumption change rate by a trend compensation coefficient to obtain a trend compensation amount; wherein, when the dynamic power consumption change rate is positive, the trend compensation coefficient is taken as the temperature rise compensation. Compensation coefficient; when the dynamic power consumption change rate is negative, the trend compensation coefficient is taken as the cooling compensation coefficient, and the cooling compensation coefficient is less than the heating compensation coefficient; S4, the difference between the current temperature and the target temperature is taken as the temperature deviation, the temperature deviation is input into the incremental proportional-integral-derivative controller, the feedback correction increment is calculated based on the proportional term, integral term and derivative term, and the feedback correction increment is superimposed on the feedback correction amount of the previous sampling period to obtain the feedback correction amount; S5, the reference cooling capacity is added to the trend compensation amount to obtain the feedforward cooling capacity, the feedforward cooling capacity is added to the feedback correction amount to obtain the total cooling command, and sent to the air conditioning system.

[0007] This invention separates dynamic power consumption by subtracting the static power consumption baseline from the current total power, thus removing the large and constant static power consumption component. This improves the sensitivity of the feedforward input to fluctuations in the actual heat dissipation load of the equipment. The baseline cooling capacity directly tracks the current heat dissipation load, while the trend compensation amount uses different compensation coefficients depending on the sign of the dynamic power consumption change rate. There is a natural difference in thermal inertia between the air in the computer room and the equipment body, resulting in rapid heating and slow cooling in the computer room. By setting the cooling compensation coefficient to be smaller than the heating compensation coefficient, the cooling withdrawal force during the cooling descent phase is less than the cooling addition force during the cooling rise phase. The feedforward cooling capacity calculated through feedforward can actively track changes in the equipment's heat dissipation load, while the feedback correction amount generated by the proportional-integral-derivative controller is used to compensate for disturbance errors not covered by the feedforward. The combination of these two methods eliminates the temperature over-limit problem caused by cooling output lag and the temperature rebound problem caused by cooling redundancy, simultaneously achieving temperature stability and energy consumption reduction.

[0008] Preferably, after the total cooling command is sent to the air conditioning system, the method further includes: in each sampling period, adding the integral term of the current sampling period to the integral accumulation of the previous sampling period to obtain the current integral accumulation; when the current integral accumulation has the same direction of increase or decrease relative to the integral accumulation of the previous sampling period, the number of sampling periods with the same direction of increase or decrease reaches the number of evaluation periods, and the absolute value of the current integral accumulation exceeds the correction threshold, dividing the current integral accumulation by the feedforward mapping coefficient to obtain the baseline correction amount; subtracting the baseline correction amount from the static power consumption baseline to update the static power consumption baseline; deducting the current integral accumulation from the feedback correction amount, and resetting the integral accumulation to zero.

[0009] By monitoring the unidirectional accumulation of integral terms, the system automatically converts them into baseline correction amounts to update the static power consumption baseline when set conditions are met. This mechanism enables the baseline to automatically adapt to long-term changes such as equipment aging or additions / removals, eliminating the need for manual recalibration by maintenance personnel and directly reusing existing data without incurring additional computational overhead.

[0010] Preferably, the determination of the number of evaluation cycles and the correction threshold includes: counting the longest continuous number of cycles in which a one-time disturbance in the computer room causes the integral term to accumulate continuously in the same direction; multiplying the longest continuous number of cycles by a preset safety factor to obtain the number of evaluation cycles; and taking a preset proportion of the static power consumption baseline as the correction threshold.

[0011] Multiplying the longest duration period by a safety factor as the trigger time condition can effectively distinguish between persistent baseline drift and one-off disturbances such as personnel entry and exit, avoiding the interference of short-term disturbances on feedforward accuracy due to miscorrection.

[0012] Preferably, determining the heating compensation coefficient and the cooling compensation coefficient includes: maintaining the communication equipment at a constant power consumption while the air conditioner is off, recording the temperature rise curve of the computer room, and fitting a first-order inertial element to obtain a heating time constant; maintaining the air conditioner at a constant cooling capacity after the power consumption of the communication equipment has dropped to its minimum, recording the temperature fall curve of the computer room, and fitting a first-order inertial element to obtain a cooling time constant; the heating compensation coefficient is a preset proportional value of the heating time constant; the cooling compensation coefficient is the heating compensation coefficient multiplied by the ratio of the heating time constant to the cooling time constant.

[0013] The compensation coefficient is set based on the actual measured heating and cooling time constants in the computer room, so that the strength ratio of asymmetric trend compensation is driven by real thermal response characteristic data, and the compensation accuracy is highly matched with the actual thermal inertia of the computer room environment.

[0014] Preferably, the determination of the feedforward mapping coefficient includes: during normal business hours and when the dynamic power consumption is greater than a preset lower limit, when the absolute value of the temperature deviation is less than a preset temperature threshold within a consecutive preset number of sampling periods, and the ratio of the absolute value of the feedback correction to the total cooling command is less than a preset ratio, it is determined to be in a steady temperature state; the actual cooling capacity and the dynamic power consumption of the air conditioning system under the steady temperature state are recorded, and the actual cooling capacity is divided by the dynamic power consumption to obtain a single mapping coefficient; the arithmetic mean of the single mapping coefficients of no less than a preset number of times is taken as the feedforward mapping coefficient.

[0015] When the computer room is running normally and has entered a steady temperature state, data is automatically recorded to calibrate the mapping coefficients. The arithmetic mean of multiple samples is used to suppress single random deviations, thereby improving the accuracy and operational stability of the feedforward mapping relationship.

[0016] Preferably, the establishment of the static power consumption baseline includes: continuously collecting the current total power during off-peak periods and removing abnormal sampling points to obtain valid sampling points; when the proportion of valid sampling points to the total number of sampling points is not less than a preset proportion, taking the arithmetic mean of the current total power corresponding to all valid sampling points as the static power consumption baseline; when the proportion of valid sampling points to the total number of sampling points is less than the preset proportion, using the previous valid static power consumption baseline.

[0017] By sampling multiple points and taking the arithmetic mean during off-peak hours, random noise in electrical measurements was suppressed; when there were insufficient effective sampling points, historical baselines were used to ensure the stability and reliability of baseline parameters.

[0018] Preferably, the determination of the business off-peak period includes: collecting hourly power consumption data of the data center for a consecutive preset number of days, calculating the power consumption mean and power consumption variance of each hourly period, selecting the candidate period for the consecutive preset number of hours with the lowest power consumption mean, and then selecting the period with the smallest power consumption variance from the candidate period as the business off-peak period.

[0019] The lowest average power consumption ensures the minimum dynamic power consumption during that period, and the minimum power consumption variance ensures the most stable power consumption fluctuations during that period. Combining these two dimensions accurately identifies the period that best matches the static power consumption characteristics, thus improving the basic accuracy of baseline calibration.

[0020] Preferably, the determination of the proportional gain, integral time, and derivative time of the incremental proportional-integral-derivative controller includes: applying a step increment to the cooling capacity of the air conditioning system under the condition that the feedforward cooling capacity is normally output, and recording the response curve of the room temperature; extracting the process gain, pure time delay, and response time constant from the response curve; and substituting the process gain, the pure time delay, and the response time constant into the open-loop step response tuning formula to obtain the proportional gain, the integral time, and the derivative time.

[0021] The characteristic parameters of the computer room system are extracted based on the standard open-loop step response test and substituted into the tuning formula to calculate the control parameters, which avoids the blindness of manual experience-based parameter tuning and ensures the adjustment quality of feedback correction.

[0022] Preferably, after collecting the current temperature of the equipment room and the current total power of the communication equipment, the method further includes: determining the validity of the current temperature and the current total power; when the current temperature exceeds the measurement range or the current total power is negative, determining that the data of the current sampling period is abnormal, and using the valid value of the previous sampling period for subsequent calculations; when the number of sampling periods with consecutive abnormal data exceeds a preset number of periods, setting the feedforward cooling capacity to zero, and sending the total cooling command obtained by adding the zeroed feedforward cooling capacity to the feedback correction amount to the air conditioning system.

[0023] By identifying anomalies in the collected data and switching to a backup mode, the system can still maintain basic temperature control capabilities by relying on the feedback channel when sensors fail or data is abnormal, thus improving the overall safety and fault tolerance of operation.

[0024] Secondly, the present invention provides an adaptive control system for the power environment of a communication equipment room, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the aforementioned adaptive control method for the power environment of a communication equipment room is implemented.

[0025] By adopting the above technical solution, a computer program is generated from the above-mentioned adaptive control method for the power environment of a communication equipment room, and stored in a memory so that it can be loaded and executed by a processor. Terminal equipment can then be manufactured based on the memory and the processor for convenient use.

[0026] The technical solution of the present invention has the following beneficial technical effects: This invention separates the dynamic load characteristics of power consumption of communication equipment that fluctuate with service traffic, and introduces a feedforward compensation model that conforms to the asymmetric thermal response law of the data center environment. It is supplemented by a feedback correction and adjustment architecture for residual deviation, which changes the working mode of passively waiting for temperature changes before performing regulation.

[0027] Furthermore, this enables the cooling output to predict and match the instantaneous heat dissipation rate of the communication load, mitigating the risks of temperature exceeding limits and rebound fluctuations, while avoiding energy waste caused by excessive cooling output, effectively improving the operational reliability and energy-saving performance of the power environment of communication facilities. Attached Figure Description

[0028] Figure 1 This is a flowchart of an adaptive control method for the power environment of a communication equipment room. Figure 2 This is the signal flow block diagram of the feedforward-feedback composite control architecture. Detailed Implementation

[0029] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings of the embodiments of the present invention. Obviously, the described embodiments are some embodiments of the present invention, but not all embodiments.

[0030] This embodiment discloses an adaptive control method for the power environment of a communication equipment room, using a base station equipment room managed by a telecommunications operator as the application scenario. The base station equipment room houses the main base station equipment, transmission equipment, and supporting power supply equipment. The equipment room area is 30 square meters and equipped with two precision air conditioners. During operation, the communication equipment generates heat, and the air conditioning cooling system removes this heat from the equipment room through cooling output to maintain the equipment room temperature within a safe range.

[0031] This embodiment adopts a feedforward-feedback composite control architecture, using the current total power of the communication equipment as the feedforward input and combining it with the current temperature of the computer room to form a feedback correction amount, so that the air conditioning cooling output can adapt to the changes in the dynamic heat dissipation load of the communication equipment.

[0032] Reference Figure 1 This includes steps S1 to S5: S1. Collect the current temperature of the computer room and the current total power of the communication equipment; establish a static power consumption baseline, and subtract the static power consumption baseline from the current total power of the communication equipment to obtain the dynamic power consumption.

[0033] In this embodiment, a temperature sensor deployed in the return air area of ​​the computer room is used to collect the ambient temperature of the computer room in real time to obtain the current temperature; a smart meter installed on the power distribution cabinet of the computer room is used to collect the input power of all communication equipment in real time to obtain the current total power; when there are multiple power distribution cabinets in the computer room, the sum of the meter readings of each power distribution cabinet is taken as the current total power. Both the temperature sensor and the smart meter are standard equipment of the power environment monitoring system for the communication computer room.

[0034] The current temperature of the computer room and the current total power of the communication equipment are collected using the same period. In this embodiment, the collection period is 30 seconds. When the collection period is shorter than 30 seconds, the impact of random noise in electrical measurements on the power change rate increases; when the collection period is longer than 30 seconds, power change information corresponding to rapid fluctuations in service traffic is missing. Specifically, computer room power data is collected at periods of 10 seconds, 30 seconds, and 60 seconds, and the power change rate sequence is calculated. Then, the signal-to-noise ratio (SNR) of each power change rate sequence is calculated, and the period corresponding to the maximum SNR is determined as the collection period. The SNR is obtained by dividing the absolute value of the mean of the power change rate sequence by the standard deviation of the power change rate sequence.

[0035] After acquiring the current temperature of the computer room and the current total power of the communication equipment in each sampling cycle, the control processing unit performs a validity judgment. If the current temperature of the computer room exceeds the range of the temperature sensor, or if the current total power of the communication equipment is negative, the control processing unit determines that the data of the current sampling cycle is abnormal and uses the valid current temperature of the computer room and the current total power of the communication equipment from the previous sampling cycle for subsequent calculations. When the number of sampling cycles with consecutive abnormal data exceeds the preset number of cycles, the control processing unit sets the feedforward cooling capacity to zero, adds the zero-value feedforward cooling capacity to the feedback correction amount to obtain the total cooling command, and sends the total cooling command to the air conditioning system. In this embodiment, the preset number of cycles is 10 sampling cycles, that is, when no valid data is obtained for 300 consecutive seconds, the system enters the backup mode. After the feedforward cooling capacity is set to zero, the total cooling command is composed of the feedback correction amount. The air conditioning system performs temperature control based on the temperature deviation between the most recent valid current temperature of the computer room and the target temperature. After the data is recovered, the feedforward channel is reactivated. The temperature sensor range is determined according to the factory calibration of the equipment. In this embodiment, the temperature sensor range is -40℃ to 85℃.

[0036] After data acquisition and validity assessment, the control processing unit establishes a static power consumption baseline. The total power consumption of the communication equipment includes static power consumption and dynamic power consumption. Static power consumption originates from leakage current consumption when the equipment is powered on, fixed losses in the power conversion circuit, and the constant-speed operation of the cooling fan; static power consumption does not change with communication traffic. Dynamic power consumption originates from the switching activity power consumption of the processor varying with service load and the power consumption of the RF power amplifier varying with transmit power; dynamic power consumption changes with communication traffic. Static power consumption accounts for 60% to 80% of the total power consumption of the communication equipment. When the current total power of the communication equipment is directly used as the feedforward input, static power consumption reduces the magnitude of dynamic power consumption changes, and decreases the sensitivity of the feedforward channel to identifying dynamic heat loads. Therefore, it is necessary to establish a static power consumption baseline to separate dynamic power consumption from the total power.

[0037] The establishment of a static power consumption baseline can be divided into two scenarios: initial initialization and routine calibration.

[0038] When the system is first powered on and before the actual measurement calibration during off-peak hours is completed, the control processing unit multiplies the sum of the rated power consumption of all communication devices in the equipment room by a preset proportional coefficient to obtain the static power consumption baseline. The preset proportional coefficient is determined based on the statistical median of the static power consumption of communication devices as a percentage of their rated power consumption; in this embodiment, the preset proportional coefficient is 0.7. After the initial actual measurement calibration during off-peak hours is completed, the control processing unit replaces the initial static power consumption baseline with the measured static power consumption baseline.

[0039] During routine calibration, the control processing unit continuously collects the current total power of the communication equipment during off-peak hours and eliminates abnormal sampling points based on the current temperature range of the equipment room and the non-negative condition of the current total power of the communication equipment. During off-peak hours, the dynamic power consumption of the communication equipment is at its lowest level, and the current total power of the communication equipment approximates its static power consumption. When the proportion of valid sampling points to the total number of sampling points is not less than a preset proportion, in order to suppress random noise in electrical measurements and improve baseline stability, the control processing unit determines the arithmetic mean of the current total power of the communication equipment corresponding to all valid sampling points as the static power consumption baseline. When the proportion of valid sampling points to the total number of sampling points is less than a preset proportion, the control processing unit uses the previous valid static power consumption baseline; in this embodiment, the preset proportion is 80%.

[0040] To enable automatic execution of routine calibration, the control processing unit pre-determines off-peak periods for routine calibration. Specifically, the control processing unit collects hourly power consumption data from the equipment room for a preset number of consecutive days, calculates the mean power consumption and variance for each consecutive preset hour window, identifies the consecutive hour window with the minimum mean power consumption as the candidate period, and determines the period with the minimum variance among the candidate periods as the off-peak period. The mean power consumption determines the dynamic power consumption level of the communication equipment, and the variance determines the degree of power consumption fluctuation. In this embodiment, the preset number of days is 7 days, and the preset number of consecutive hours is 2 hours. The 7-day collection window covers the changes in business traffic on weekdays and weekends; the 2-hour consecutive window ensures a sufficient number of sampling points during the off-peak period to obtain a stable mean. In this embodiment, the hourly power consumption statistics of the base station equipment room for 7 consecutive days show that the average power from 3:00 AM to 5:00 AM is 4.8 kW and the power variance is 0.12 kW squared, which is the lowest and most stable period of the day. The control processing unit determines the period from 3:00 AM to 5:00 AM as the low-end period and sets the static power consumption baseline at 4.8 kW.

[0041] After the static power consumption baseline is established, the dynamic power consumption is obtained by subtracting the static power consumption baseline from the current total power of the communication device in each sampling period. When the calculated result of subtracting the static power consumption baseline from the current total power is negative, the control processing unit sets the dynamic power consumption to zero. At this time, the feedforward cooling capacity is set to zero, and temperature regulation is performed by the feedback correction.

[0042] Thus, by introducing data validity judgment and backup mode switching in the acquisition stage, the basic temperature control capability can still be maintained even if the sensor or meter is abnormal; by establishing a static power consumption baseline and performing subtraction separation, the large and constant static power consumption is separated from the total power, and the dynamic power consumption signal that truly fluctuates with the traffic flow is retained, so that the subsequent feedforward calculation can sensitively perceive the real-time changes in heat dissipation load.

[0043] S2. Divide the difference between two adjacent samples of dynamic power consumption by the sampling period to obtain the dynamic power consumption change rate; multiply the dynamic power consumption by the feedforward mapping coefficient to obtain the reference cooling capacity.

[0044] In this embodiment, the control processing unit performs an adjacent sampling period difference operation on the dynamic power consumption: the dynamic power consumption of the current sampling period is subtracted from the dynamic power consumption of the previous sampling period, and then the difference is divided by the sampling period to obtain the dynamic power consumption change rate. The unit of the dynamic power consumption change rate is kilowatts per second. A positive dynamic power consumption change rate indicates an increase in heat dissipation load, while a negative rate indicates a decrease. The absolute value of the dynamic power consumption change rate represents the rate of change of heat dissipation load; a larger absolute value indicates a faster rate of change of heat dissipation load. It should be noted that the first sampling period after system startup lacks the dynamic power consumption of the previous sampling period, and the control processing unit sets the dynamic power consumption change rate to zero.

[0045] Due to the thermal inertia of the air and equipment in the computer room, it takes a period of time for the increased heat dissipation power of the equipment to be fully reflected in the room temperature. Simply tracking the current value of dynamic power consumption will cause the cooling capacity adjustment to lag behind the actual heat dissipation demand during the rapid rise of power consumption due to thermal inertia. The dynamic power consumption change rate can provide a quantitative basis for the direction and speed of load change for subsequent trend compensation.

[0046] After obtaining the dynamic power consumption, it is multiplied by the feedforward mapping coefficient to obtain the baseline cooling capacity. The feedforward mapping coefficient is used to characterize the calibration relationship between unit dynamic power consumption and the actual cooling capacity of the air conditioning system. The input electrical energy of communication equipment is mainly dissipated in the form of heat, and dynamic power consumption can be used as an input to characterize the dynamic heat dissipation load. According to the first law of thermodynamics, during steady-state operation, the control processing unit calibrates the feedforward mapping coefficient based on the measured ratio of the actual cooling capacity of the air conditioning system to the dynamic power consumption. The specific value of the feedforward mapping coefficient is related to the thermal environment of the computer room, the heat dissipation load corresponding to the dynamic power consumption, and the measurement conditions of the actual cooling capacity.

[0047] In this embodiment, the initial value of the feedforward mapping coefficient is 1. The calibration of the feedforward mapping coefficient is performed when the dynamic power consumption of the communication equipment is at a normal service level and greater than a preset lower limit, to ensure the stability of the division result. Specifically, the preset lower limit is determined according to the resolution of the smart meter and the power measurement error of the communication equipment, to avoid the denominator of the dynamic power consumption division operation being 0. During normal service periods, when the absolute value of the temperature deviation is less than the preset temperature threshold within a consecutive preset sampling period, and the ratio of the absolute value of the feedback correction to the total cooling command is less than a preset proportion, the control processing unit determines that the room temperature is in a steady state. In this embodiment, the preset sampling period is 10, the preset temperature threshold is 0.5℃, and the preset proportion is 5%. The above three conditions together constitute the criteria for determining the steady state of temperature: the absolute value of the temperature deviation is less than 0.5℃, which means that the current temperature of the computer room meets the target temperature control requirements; the temperature deviation condition is met for 10 consecutive sampling cycles, excluding instantaneous fluctuations; the proportion of feedback correction is less than 5%, which means that the feedforward cooling capacity undertakes most of the task of tracking the heat dissipation load of communication equipment, and the feedback channel is in a light-load state. At this time, the actual cooling capacity of the air conditioner is mainly determined by the feedforward, which is suitable for back-mapping the mapping relationship.

[0048] During the steady-state temperature period in the computer room, the control processing unit records the actual cooling capacity and dynamic power consumption of the air conditioning system. It divides the actual cooling capacity by the dynamic power consumption to obtain the single-time mapping coefficient. When the air conditioning controller provides actual cooling capacity operating parameters, the control processing unit reads the actual cooling capacity from the controller. When the air conditioning controller only supports temperature setpoint control, the control processing unit converts the temperature setpoint into actual cooling capacity according to the air conditioning system cooling capacity-setpoint correspondence table. After obtaining at least a preset number of single-time mapping coefficients, the control processing unit calculates the arithmetic mean of the single-time mapping coefficients to obtain the feedforward mapping coefficient. In this embodiment, the preset number of times is 3. (Feedforward mapping coefficient) The calibration formula is:

[0049] In the formula, This indicates the number of valid records, calculated from the number of records that meet the steady-state temperature conditions of the computer room, and is not less than the preset number. Indicates the first The actual cooling capacity of the air conditioning system under sub-steady-state temperature conditions is obtained from the operating parameters of the air conditioning controller, and the unit is kilowatts. For the first The dynamic power consumption under sub-temperature steady state is obtained by subtracting the static power consumption baseline from the current total power of the communication equipment, and the unit is kilowatt.

[0050] For example, taking the base station equipment room of this embodiment as an example, after deployment and continuous operation for 5 days, a total of 5 periods meeting the steady-state temperature conditions were recorded. The actual cooling capacity for each period was 2.21 kW, 2.27 kW, 2.16 kW, 2.23 kW, and 2.25 kW, respectively, and the corresponding dynamic power consumption was 2.1 kW, 2.12 kW, 2.08 kW, 2.11 kW, and 2.1 kW, respectively. The 5 single-time mapping coefficients were 1.05, 1.07, 1.04, 1.06, and 1.07, respectively. The arithmetic mean of the 5 single-time mapping coefficients was 1.06. The control processing unit determined 1.06 as the feedforward mapping coefficient, indicating that under the equipment room and the calibration conditions, the measured mapping relationship between the actual cooling capacity and dynamic power consumption of the air conditioning system was 1.06.

[0051] After obtaining the feedforward mapping coefficients, the reference cooling capacity The calculation formula is as follows:

[0052] In the formula, the reference cooling capacity The unit is kilowatt, used to represent the cooling demand of the feedforward channel corresponding to the current dynamic heat dissipation load; This represents the feedforward mapping coefficient, which is obtained from the arithmetic mean of the single mapping coefficients during the steady-state period of the computer room temperature; This represents the dynamic power consumption during the current sampling period, obtained by subtracting the static power consumption baseline from the current total power of the communication device, and is expressed in kilowatts.

[0053] Thus, by obtaining the dynamic power consumption change rate through differential operation, a quantitative input of the heat dissipation load change rate is provided for subsequent trend compensation; by calibrating the feedforward mapping coefficient in a data-driven manner under steady-state temperature conditions, the reference cooling capacity can accurately match the actual dynamic heat dissipation load level.

[0054] S3. The control processing unit multiplies the dynamic power consumption change rate by the trend compensation coefficient to obtain the trend compensation amount. When the dynamic power consumption change rate is positive, the trend compensation coefficient is the heating compensation coefficient; when the dynamic power consumption change rate is negative, the trend compensation coefficient is the cooling compensation coefficient; the cooling compensation coefficient is less than the heating compensation coefficient.

[0055] In this embodiment, the control processing unit multiplies the dynamic power consumption change rate by the trend compensation coefficient to obtain the trend compensation amount. The dynamic power consumption change rate is in kilowatts per second, the trend compensation coefficient is in seconds, and the trend compensation amount is in kilowatts. The trend compensation coefficient represents the lead compensation time of the feedforward channel. When the trend compensation coefficient increases, the trend compensation amount of the feedforward channel increases.

[0056] Because the heating and cooling processes of a computer room exhibit asymmetric thermal responses, when the dynamic power consumption of communication equipment increases, the equipment acts as an active heat source, with its surface temperature exceeding the surrounding air temperature. This temperature difference drives rapid convection heat transfer, resulting in a faster release of heat load into the computer room air. Conversely, when the dynamic power consumption of the communication equipment decreases, the heat stored in the equipment body and cabinet continues to be released, and the rate of decrease in the computer room's heat load lags behind the rate of decrease in the communication equipment's power consumption. This means that during the power consumption increase phase, a larger trend compensation is needed to quickly add cooling to prevent temperature exceeding limits. Conversely, if cooling is withdrawn with the same force during the power consumption decrease phase, the room temperature may rebound due to the continued release of residual heat.

[0057] Based on the aforementioned physical asymmetry, an asymmetric trend compensation strategy is adopted. Specifically, when the rate of change of dynamic power consumption is positive, indicating that the heat dissipation load is increasing, the control processing unit determines the trend compensation coefficient as the temperature rise compensation coefficient; when the rate of change of dynamic power consumption is negative, indicating that the heat dissipation load is decreasing, the control processing unit determines the trend compensation coefficient as the temperature fall compensation coefficient. The temperature fall compensation coefficient is less than the temperature rise compensation coefficient, ensuring that the cooling withdrawal force during the heat dissipation decrease phase is less than the cooling addition force during the heat dissipation increase phase. Trend compensation amount The calculation formula is as follows:

[0058] In the formula, the unit of trend compensation is kilowatt; Used to represent the rate of change of dynamic power consumption, calculated from the difference in dynamic power consumption between adjacent sampling periods. Divide by sampling period The unit is kilowatts per second; The trend compensation coefficient is determined by the sign of the dynamic power consumption change rate, and its unit is seconds. When the dynamic power consumption change rate is greater than 0, the trend compensation coefficient is... Take the temperature rise compensation coefficient; when the dynamic power consumption change rate is less than 0, take the trend compensation coefficient. Take the cooling compensation coefficient; when the dynamic power consumption change rate is equal to 0, the trend compensation amount is 0.

[0059] It is important to note that the values ​​of the heating compensation coefficient and the cooling compensation coefficient need to match the thermal response characteristics of the computer room itself. The determination of the two coefficients depends on the heating time constant and the cooling time constant of the computer room, and they are obtained through two sets of experimental calibrations.

[0060] Specifically, the calibration method for the temperature rise time constant is as follows: With the air conditioning off, the communication equipment operates at constant power consumption, and the temperature sensor continuously records the computer room temperature according to the sampling period. If the computer room temperature reaches the preset temperature protection upper limit, the air conditioning system resumes operation; if the computer room temperature does not reach the preset temperature protection upper limit, recording continues until the temperature curve enters a stable range. The control processing unit uses the computer room temperature data collected during the air conditioning shutdown period to fit the computer room temperature rise curve based on a first-order inertial element, obtaining the temperature rise time constant. The computer room temperature rise curve is expressed by the following formula:

[0061] In the formula, This indicates the elapsed time after the air conditioner was turned off. The temperature of the computer room at that time, in degrees Celsius; The initial temperature read by the temperature sensor at the moment the air conditioner is turned off, in degrees Celsius; This represents the steady-state temperature rise obtained from fitting a first-order inertial model, in degrees Celsius; when the preset temperature protection upper limit is not reached and the temperature curve enters the stable range... It can be determined by the difference between the stable temperature range of the computer room and the initial temperature; This represents the heating time constant, expressed in seconds. Under the first-order inertial model, the heating time constant is the time corresponding to when the temperature rise reaches 63.2% of the steady-state temperature rise obtained from the model fitting. The least squares method is used to fit the measured temperature data collected during the air conditioning shutdown period, and the steady-state temperature rise and heating time constant are jointly determined based on minimizing the sum of squared residuals. When the measured temperature curve enters the stable range, the heating time constant can also be directly determined by the time point corresponding to when the temperature rise reaches 63.2% of the steady-state temperature rise.

[0062] Specifically, the calibration method for the cooling time constant is as follows: After the power consumption of the communication equipment is reduced to its minimum, the air conditioner is kept running at a constant cooling capacity, the temperature drop curve of the computer room is recorded, and a first-order inertial element is used for fitting to obtain the cooling time constant. The fitting method for the cooling time constant is the same as the fitting method for the heating time constant. Since the residual heat of the communication equipment is continuously released during the cooling process, the temperature drop of the computer room is delayed. In this embodiment, the cooling time constant is greater than the heating time constant.

[0063] After obtaining the heating and cooling time constants, the control processing unit sets the candidate value of the heating compensation coefficient as the product of the heating time constant and a preset proportional value. The candidate range of the preset proportional value is 0.5 to 1 to ensure that the compensation time does not exceed the heating time constant of the controlled object and to avoid insufficient compensation due to an excessively small compensation time. Within the candidate heating compensation coefficient range corresponding to 50% to 100% of the heating time constant, the control processing unit performs a step disturbance test with a step size of 10 seconds, and determines the candidate heating compensation coefficient corresponding to the minimum temperature overshoot as the heating compensation coefficient. The control processing unit multiplies the heating compensation coefficient by the ratio of the heating time constant to the cooling time constant to obtain the cooling compensation coefficient. :

[0064] In the formula, the cooling compensation coefficient is... The unit is seconds; This represents the temperature rise compensation coefficient, obtained from the test result of the smallest temperature overshoot in the step disturbance test, and is expressed in seconds. This represents the heating time constant, in seconds. This represents the cooling time constant, obtained by fitting the temperature drop curve in the communication equipment load reduction experiment, and is expressed in seconds. When the cooling time constant is greater than the heating time constant, the ratio of the heating time constant to the cooling time constant is less than 1, and the cooling compensation coefficient is less than the heating compensation coefficient. This causes the compensation intensity during the cooling phase to decrease accordingly based on the difference in thermal inertia between the heating and cooling processes.

[0065] For example, taking the base station equipment room of this embodiment as an example, the air conditioning shutdown experiment obtained a heating time constant of 180 seconds, and the communication equipment load reduction experiment obtained a cooling time constant of 260 seconds. The control processing unit performs a step test in 10-second increments within the range of 90 to 180 seconds. The heating compensation coefficient corresponding to the minimum temperature overshoot is 120 seconds, and the cooling compensation coefficient is... Seconds. Assuming the dynamic power consumption change rate is 0.04 kW / s for a certain sampling period, and the control processing unit uses a temperature rise compensation coefficient of 120 seconds, the trend compensation amount is... At one moment, the dynamic power consumption change rate is -0.03 kW / s. The control processing unit uses a cooling compensation coefficient of 83 seconds, and the trend compensation amount is... kilowatts. If the asymmetric strategy is not used here and 120 seconds is still used, the calculation result is: The reduction in cooling capacity is approximately 1.1 kilowatts, thus preserving more cooling capacity to cope with the slow release of residual heat from communication equipment and reduce the risk of room temperature rebound.

[0066] Thus, by multiplying the dynamic power consumption change rate with the trend compensation coefficient to obtain the trend compensation amount, the feedforward channel can make advance compensation according to the rate of change of heat dissipation load; the asymmetric trend compensation strategy incorporates the thermodynamic asymmetry of rapid heating and slow cooling of the computer room into the feedforward model, reducing the risk of temperature exceeding the limit and temperature rebound.

[0067] S4. The control processing unit subtracts the target temperature from the current temperature of the computer room to obtain the temperature deviation. The control processing unit inputs the temperature deviation into the incremental proportional-integral-derivative controller, calculates the feedback correction increment based on the proportional term, integral term and derivative term, and adds the feedback correction increment to the feedback correction amount of the previous sampling period to obtain the feedback correction amount.

[0068] In this embodiment, the feedforward channel can actively pre-compensate for most dynamic heat dissipation loads, but it cannot cover all disturbance factors. Factors such as heat transfer from the building envelope, hot air intrusion caused by personnel entering and exiting doors, and local efficiency differences in communication equipment can all cause the actual temperature in the computer room to deviate from the target temperature. Therefore, a feedback channel is needed to correct the actual temperature deviation and compensate for the residual error of the feedforward channel.

[0069] The temperature deviation is obtained by subtracting the target temperature from the current temperature. The target temperature is set according to communication industry standards, with a recommended operating temperature range of 23°C to 27°C for communication equipment rooms. In this embodiment, the preset target temperature is 25°C. When the temperature deviation is positive, the feedback correction increases the cooling capacity; when the temperature deviation is negative, the feedback correction decreases the cooling capacity.

[0070] Using temperature deviation as input, the standard incremental proportional-integral-derivative (PID) control algorithm is employed to calculate the feedback correction. The standard incremental PID algorithm outputs a correction increment in each sampling period, which is accumulated to the feedback correction of the previous sampling period to obtain the feedback correction for the current sampling period. The correction increment consists of three terms: a proportional term, an integral term, and a derivative term. The proportional term generates an instantaneous adjustment increment based on the change in temperature deviation between two adjacent sampling periods. The integral term equals the current temperature deviation multiplied by the ratio of the sampling period to the integral time, accumulating persistent deviations over time to eliminate steady-state error. The derivative term generates a predictive adjustment increment based on the second-order difference of the temperature deviation over the most recent three sampling periods, suppressing overshoot.

[0071] The feedback correction update formula is:

[0072] In the formula, Indicates the first Feedback correction amount per sampling period, in kilowatts; This indicates the feedback correction amount from the previous sampling period; Indicates the first The feedback correction increment for each sampling period is calculated using the proportional, integral, and differential terms, and is expressed in kilowatts. The formula for calculating the feedback correction increment is:

[0073] Among them, the proportional correction increment The calculation formula is:

[0074] Integral Correction Increment The calculation formula is:

[0075] Differential correction increment The calculation formula is:

[0076] In the formula, Indicates proportional gain. Integral gain represents the integral gain. By proportional gain Multiply by the sampling period Divide by integration time The value is obtained in kilowatts per degree Celsius; differential gain. By proportional gain Multiply by the differential time Divide by sampling period The value is obtained, measured in kilowatts per degree Celsius. Indicates the first The temperature deviation for each sampling period is obtained by subtracting the target temperature from the current temperature of the computer room, and the unit is degrees Celsius. and The temperature deviations for the previous and two previous sampling periods are respectively obtained by the control processing unit and are expressed in degrees Celsius. It is important to note that for the first two sampling periods after system startup, missing historical deviation values ​​are replaced with 0. An incremental algorithm is used, where the feedback correction is recursively obtained from the feedback corrections of adjacent sampling periods and the feedback correction increment of the current period. This ensures that the calculated deviation for a single sampling period is primarily reflected in the feedback correction increment of the current period, facilitating the monitoring and limiting of the feedback correction increment, thereby improving the safety of control operation.

[0077] The determination of the three core parameters of the proportional-integral-derivative controller includes: applying a step increment to the cooling capacity of the air conditioning system under normal feedforward cooling capacity output conditions, recording the response curve of the computer room temperature, and extracting the process gain, pure time delay, and response time constant from the response curve; substituting the three characteristic parameters into the tuning formula to obtain the proportional gain, integral time, and derivative time.

[0078] In this embodiment, the step increment is 5% of the current total cooling capacity, thereby stimulating an observable temperature response without causing excessive disturbance to the computer room temperature. After the control processing unit applies the step increment, the temperature sensor records the computer room temperature according to the sampling period until the computer room temperature enters the stable range. The process gain is obtained by dividing the total change in computer room temperature by the change in cooling capacity, in degrees Celsius per kilowatt; the pure lag time is obtained by subtracting the step application time from the moment when the computer room temperature first deviates from the initial steady-state value by more than 5% of the total steady-state change, in seconds; the response time constant is obtained by subtracting the moment when the computer room temperature begins to respond from the moment when the computer room temperature reaches 63.2% of the steady-state change, in seconds.

[0079] The control processing unit substitutes the process gain, pure time delay, and response time constant into the Ziegler-Nichols tuning formula:

[0080]

[0081]

[0082] In the formula, For proportional gain, The response time constant is expressed in seconds. Process gain, expressed in degrees Celsius per kilowatt. The time delay is the pure time lag, in seconds. The integration time is in seconds. The time is the differential time, expressed in seconds. In this embodiment, the process gain measured by the step response experiment is 0.3 degrees Celsius per kilowatt, the pure time delay is 20 seconds, and the response time constant is 150 seconds. Substituting these values ​​into the tuning formula: , Second, Second.

[0083] In this way, the feedback channel uses a proportional-integral-derivative control algorithm to perform closed-loop correction on unknown disturbances that the feedforward cannot cover, ensuring that the room temperature eventually converges to the target value. The feedforward and feedback channels work together, and the temperature control accuracy is better than that of a solution that uses either channel alone.

[0084] S5. Add the baseline cooling capacity to the trend compensation amount to obtain the feedforward cooling capacity. Add the feedforward cooling capacity to the feedback correction amount to obtain the total cooling command. Send the total cooling command to the air conditioning system to perform cooling control.

[0085] like Figure 2As shown, the control processing unit adds the baseline cooling capacity to the trend compensation amount to obtain the feedforward cooling capacity; the control processing unit adds the feedforward cooling capacity to the feedback correction amount to obtain the total cooling command. When the total cooling command exceeds the rated cooling capacity of the air conditioning system, the control processing unit determines the rated cooling capacity as the actual output value and sends a high-temperature warning message; when the total cooling command is less than 0, the control processing unit sets the total cooling command to zero, indicating that there is currently no cooling demand.

[0086] After receiving the total cooling command, the air conditioning system adjusts the operating frequency of the inverter compressor or the fan speed to ensure that the actual cooling output tracks the total cooling command. In another embodiment, if the air conditioning system only supports temperature setpoint control, the control processing unit converts the total cooling command into a temperature setpoint based on the air conditioning system's cooling capacity-setpoint correspondence table and sends the temperature setpoint to the air conditioning system. After execution, the system enters the next sampling cycle to re-collect the current temperature of the computer room and the current total power of the communication equipment.

[0087] During continuous operation, the static power consumption baseline is calibrated and stored as a fixed value during off-peak periods. Over long-term operation in the data center, factors such as equipment aging leading to increased power consumption of communication equipment, the addition or removal of communication equipment causing abrupt changes in the static power consumption baseline, and seasonal changes in ambient temperature affecting power conversion efficiency, all cause the actual static power consumption to gradually deviate from the initially calibrated baseline. When the static power consumption baseline is lower than the actual static power consumption, the calculated dynamic power consumption increases, and the feedforward cooling capacity increases; conversely, when the static power consumption baseline is higher than the actual static power consumption, the calculated dynamic power consumption decreases, and the feedforward cooling capacity decreases. Therefore, the control processing unit utilizes the cumulative behavior of the integral term in the proportional-integral-derivative (PID) controller to detect and correct baseline drift.

[0088] Specifically, in each sampling period, the cycle count of the integral term in the proportional-integral-derivative controller accumulating continuously in the same direction is recorded. When the cycle count reaches the evaluation cycle number and the absolute value of the accumulated integral term exceeds the correction threshold, the accumulated integral term is divided by the feedforward mapping coefficient to obtain the baseline correction. The static power consumption baseline is updated by subtracting the baseline correction, the accumulated integral term is deducted from the feedback correction, and the integral term is reset to zero.

[0089] When the static power consumption baseline is accurate, the current temperature of the computer room fluctuates around the target temperature, and the integral term alternates around zero, without accumulating continuously in the same direction. When the static power consumption baseline drifts, the feedforward cooling capacity is continuously too high or too low, the temperature deviation maintains the same sign, and the integral term accumulates continuously in the same direction. The integral operation uses time averaging to cancel out high-frequency random fluctuations and accumulates low-frequency persistent deviations to gradually amplify them. Therefore, the integral term naturally filters out high-frequency interference and retains low-frequency baseline drift information.

[0090] Continuous accumulation in the same direction includes: the integral term continuously increasing or decreasing over multiple consecutive sampling periods without changing direction. In each sampling period, the magnitude of the current integral term is compared with that of the integral term in the previous period. If the direction is the same as the previous period, the period count is incremented by 1; if the direction changes, the period count is reset to zero.

[0091] When triggering correction, the formula for calculating the baseline correction is:

[0092] In the formula, This is the baseline correction, in kilowatts. This represents the cumulative amount of the integral term, expressed in kilowatts. The feedforward mapping coefficient is obtained from the arithmetic mean of the single mapping coefficients recorded in the steady-state temperature data of the computer room. It represents the cooling capacity corresponding to a unit of dynamic power consumption. Its reciprocal converts the cumulative integral term of the cooling capacity into a baseline deviation in the power dimension. The updated static power consumption baseline is the original baseline value minus the baseline correction. At the instant the correction is completed, the change in feedforward cooling capacity caused by the static power consumption baseline correction is equal to the cumulative integral term. At the same time, the cumulative integral term is subtracted from the feedback correction. The changes in feedforward cooling capacity and feedback correction are equal in magnitude and opposite in direction, and the total cooling command remains continuous.

[0093] The purpose of continuous unidirectional cumulative cycle counting is to distinguish between static power consumption baseline drift and one-off disturbances. One-off disturbances, such as the intrusion of hot air caused by the brief opening of the computer room door due to personnel entering or exiting, will cause the integral term to temporarily increase in one direction, but the integral term will fall back in the opposite direction after the disturbance is eliminated, and unidirectional accumulation will not last for too long; only persistent deviation sources such as static power consumption baseline drift can cause the integral term to continuously accumulate in the same direction for a sufficiently long period of time.

[0094] The determination of the number of evaluation periods includes: the control processing unit counts the longest duration period of continuous accumulation of the integral term in the same direction caused by a one-time disturbance, multiplies the longest duration period by a safety factor to obtain the number of evaluation periods. In this embodiment, the safety factor is 1.5 to ensure that the fluctuation of the integral term caused by a one-time disturbance has sufficient margin to subside before triggering correction. During the initial stage of system launch, 30 days of continuous operating data are recorded. The control processing unit records the direction of integral term accumulation in each sampling period, counts the number of consecutive accumulation periods in the same direction caused by personnel opening door disturbances, and takes the maximum value as the longest duration period. For example, the longest duration of personnel opening door disturbances is 10 minutes, corresponding to 20 sampling periods, and the number of evaluation periods is... .

[0095] The determination of the correction threshold includes: the control processing unit multiplies the static power consumption baseline by a preset ratio to obtain the correction threshold. In this embodiment, the preset ratio is 1%. The annual static power consumption drift of the communication equipment is determined according to 5% of the rated power consumption. Calculated over 12 months, the monthly static power consumption drift is 0.4167% of the rated power consumption. Taking 1% as the correction threshold allows for engineering margin for a single drift. When the static power consumption baseline is 4.8 kW, the correction threshold is 4.8 multiplied by 1%, resulting in 0.048 kW.

[0096] In another embodiment, when the data center network management system reports a large-scale deployment or removal of communication equipment, the control processing unit immediately performs a recalibration of the static power consumption baseline during off-peak hours. Before the off-peak hours arrive, the control processing unit performs dynamic power consumption calculations using the static power consumption baseline corrected by the most recent integral term, in order to shorten the response waiting time after a sudden change in the baseline.

[0097] In this way, by combining the feedforward cooling capacity and the feedback correction quantity into a total cooling command and applying physical upper and lower limit constraints, the air conditioning system is guaranteed to receive executable commands under any operating condition. The baseline self-calibration mechanism enables the static power consumption baseline to be automatically updated according to the slow changes in the room's operating status, without the need for maintenance personnel to manually recalibrate it periodically. Moreover, this mechanism directly reuses the integral term data already generated by the proportional-integral-derivative controller, without the need for additional sensors or calculation models.

[0098] This invention also discloses an adaptive control system for the power environment of a communication equipment room, including a processor and a memory. The memory stores computer program instructions, and when the computer program instructions are executed by the processor, the adaptive control method for the power environment of a communication equipment room according to this invention is implemented.

[0099] The system also includes other components well known to those skilled in the art, such as communication buses and communication interfaces, the settings and functions of which are known in the art and will not be described in detail here.

[0100] In the description of this specification, "multiple" or "several" means at least two, such as two, three or more, unless otherwise expressly and specifically defined.

Claims

1. A method for adaptive control of the power environment in a communication equipment room, characterized in that, include: S1. Collect the current temperature of the computer room and the current total power of the communication equipment; Obtain the pre-established and stored static power consumption baseline, and subtract the static power consumption baseline from the current total power to obtain the dynamic power consumption; S2. Divide the difference between two adjacent samples of the dynamic power consumption by the sampling period to obtain the dynamic power consumption change rate; multiply the dynamic power consumption by the feedforward mapping coefficient to obtain the reference cooling capacity; S3. Multiply the dynamic power consumption change rate by the trend compensation coefficient to obtain the trend compensation amount; wherein, when the dynamic power consumption change rate is positive, the trend compensation coefficient is the heating compensation coefficient; when the dynamic power consumption change rate is negative, the trend compensation coefficient is the cooling compensation coefficient, and the cooling compensation coefficient is less than the heating compensation coefficient. S4. The difference between the current temperature and the target temperature is taken as the temperature deviation. The temperature deviation is input into the incremental proportional-integral-derivative controller. The feedback correction increment is calculated based on the proportional term, integral term and derivative term. The feedback correction increment is added to the feedback correction amount of the previous sampling period to obtain the feedback correction amount. S5. Add the baseline cooling capacity to the trend compensation amount to obtain the feedforward cooling capacity, add the feedforward cooling capacity to the feedback correction amount to obtain the total cooling command, and send it to the air conditioning system.

2. The adaptive control method for the power environment of a communication equipment room as described in claim 1, characterized in that, After the total cooling command is sent to the air conditioning system, the system also includes: In each sampling period, the integral term of the current sampling period is added to the integral accumulation of the previous sampling period to obtain the current integral accumulation. When the current integral accumulation is in the same direction of increase or decrease as the integral accumulation of the previous sampling period, the number of sampling periods with the same direction of increase or decrease reaches the number of evaluation periods, and the absolute value of the current integral accumulation exceeds the correction threshold, the current integral accumulation is divided by the feedforward mapping coefficient to obtain the baseline correction amount. The static power consumption baseline is updated by subtracting the baseline correction amount from the static power consumption baseline; the current integral accumulation amount is deducted from the feedback correction amount, and the integral accumulation amount is reset to zero.

3. The adaptive control method for the power environment of a communication equipment room as described in claim 2, characterized in that, The determination of the number of evaluation periods and the correction threshold includes: The longest duration period in which a single disturbance in the computer room causes the integral term to accumulate continuously in the same direction is calculated. The longest duration period is multiplied by a preset safety factor to obtain the number of evaluation periods. A preset proportion of the static power consumption baseline is taken as the correction threshold.

4. The adaptive control method for the power environment of a communication equipment room as described in claim 1, characterized in that, The determination of the heating compensation coefficient and the cooling compensation coefficient includes: With the air conditioning off, the communication equipment is kept running at a constant power consumption. The temperature rise curve of the computer room is recorded, and the temperature rise time constant is obtained by fitting a first-order inertial element. After the power consumption of the communication equipment is reduced to the minimum, the air conditioner is kept running at a constant cooling capacity, the temperature drop curve of the computer room is recorded, and the cooling time constant is obtained by fitting a first-order inertial element; the heating compensation coefficient is taken as a preset ratio value of the heating time constant. The cooling compensation coefficient is the ratio of the heating compensation coefficient multiplied by the heating time constant to the cooling time constant.

5. The adaptive control method for the power environment of a communication equipment room as described in claim 1, characterized in that, The determination of the feedforward mapping coefficients includes: During normal business hours and when the dynamic power consumption is greater than a preset lower limit, when the absolute value of the temperature deviation is less than a preset temperature threshold within a preset number of consecutive sampling periods, and the ratio of the absolute value of the feedback correction to the total cooling command is less than a preset ratio, the temperature is determined to be in steady state. Record the actual cooling capacity and dynamic power consumption of the air conditioning system under steady-state temperature. Divide the actual cooling capacity by the dynamic power consumption to obtain the single-time mapping coefficient. Take the arithmetic mean of the single-time mapping coefficients of no less than a preset number of times as the feedforward mapping coefficient.

6. The adaptive control method for the power environment of a communication equipment room as described in claim 1, characterized in that, The establishment of the static power consumption baseline includes: During periods of low business activity, the current total power is continuously collected, and abnormal sampling points are removed to obtain valid sampling points; When the proportion of the number of valid sampling points to the total number of sampling points is not less than a preset proportion, the arithmetic mean of the current total power corresponding to all valid sampling points is taken as the static power consumption baseline. When the proportion of the number of valid sampling points to the total number of sampling points is lower than the preset proportion, the previous valid static power consumption baseline is used.

7. The adaptive control method for the power environment of a communication equipment room as described in claim 6, characterized in that, The determination of the business off-peak period includes: Collect hourly power consumption data for a consecutive preset number of days in the data center, calculate the average power consumption and power consumption variance for each hourly period, select the candidate period for the consecutive preset number of hours with the lowest average power consumption, and then select the period with the smallest power consumption variance from the candidate period as the business off-peak period.

8. The adaptive control method for the power environment of a communication equipment room as described in claim 1, characterized in that, The determination of the proportional gain, integral time, and derivative time of the incremental proportional-integral-derivative controller includes: Under the condition that the feedforward cooling capacity is normally output, a step increment is applied to the cooling capacity of the air conditioning system, and the response curve of the computer room temperature is recorded. Extract the process gain, pure time delay, and response time constant from the response curve; Substituting the process gain, the pure time delay, and the response time constant into the open-loop step response tuning formula yields the proportional gain, the integral time, and the derivative time.

9. The adaptive control method for the power environment of a communication equipment room as described in claim 1, characterized in that, After collecting the current temperature of the data acquisition room and the current total power of the communication equipment, the following are also included: The validity of the current temperature and the current total power is determined; When the current temperature exceeds the measurement range or the current total power is negative, the data of the current sampling period is determined to be abnormal, and the effective value of the previous sampling period is used for subsequent calculations. When the number of sampling periods with continuous data anomalies exceeds the preset number of periods, the feedforward cooling capacity is set to zero, and the total cooling command obtained by adding the zeroed feedforward cooling capacity to the feedback correction amount is sent to the air conditioning system.

10. A power environment adaptive control system for a communication equipment room, characterized in that, include: A processor and a memory, wherein the memory stores computer program instructions that, when executed by the processor, implement an adaptive control method for the power environment of a communication equipment room according to any one of claims 1 to 9.

Images