Data center cooling system performance evaluation method and system

Abstract

The invention provides a data center cooling system performance evaluation method and system. The method comprises the following steps: acquiring operation data of data center cooling equipment and power load data of a server, and combining service time and maintenance records of the equipment to determine a performance baseline; calculating resonance frequency points caused by harmonic waves under different working conditions based on the vibration characteristics of the base line and the current harmonic characteristics in the electric energy quality in combination with equipment structure parameters and a working condition empirical formula; frequency sweeping is performed by actively adjusting the rotating speed of equipment, the dominant harmonic frequency and the contribution proportion thereof are verified, and a proportional relation model of harmonic and vibration characteristics is established; adjusting equipment parameters according to the model, and calculating the deviation degree of a harmonic component and a performance baseline; and finally, combining harmonic change characteristics with deviation degree correlation, and constructing energy efficiency and health state evaluation indexes of the cooling system. According to the invention, collaborative evaluation of the energy efficiency and the health state of the data center cooling system is realized.

Description

Technical Field

[0001] This application relates to the technical field of performance evaluation and optimization of data center cooling systems, and particularly relates to a method and system for evaluating the performance of a data center cooling system. Background Art

[0002] With the explosive growth of cloud computing and high-density computing power demands, data center cooling systems are facing severe challenges: the increase in the power density of server clusters has led to frequent local hotspots, and it is necessary to monitor the temperature distribution uniformity to prevent equipment overheating. Therefore, it is urgent to accurately evaluate and optimize energy efficiency indicators to reduce operating costs. At the same time, hidden factors such as vibration and harmonics may cause mechanical fatigue or electrical resonance, and it is necessary to quantify their impact on equipment stability and lifespan.

[0003] Currently, the mainstream solution adopts the combined analysis method of computational fluid dynamics (CFD) simulation and thermodynamic parameters. By constructing a three-dimensional model of the data center, dividing the cold and hot channels, cabinet layout, and air conditioner positions, and inputting boundary conditions such as server heat load, cooling medium flow rate, and ambient temperature and humidity, the air flow organization and temperature field distribution under different loads are simulated. This solution is more efficient than traditional manual monitoring, can preview the cooling effects of different working conditions, and assist in optimizing the air flow organization design.

[0004] However, despite the relatively advanced CFD simulation technology, there are still core bottlenecks: the performance baseline does not incorporate dynamic factors such as equipment service time and maintenance records, making it difficult to capture the performance degradation of aging equipment such as bearing wear, resulting in the evaluation results deviating from the actual operating state. The essential contradiction is that the simulation model assumes ideal working conditions, while in actual operation, the coupling effects of multiple physical fields of electricity, mechanics, and thermodynamics are complex, and static models are difficult to support high-precision dynamic evaluation. Summary of the Invention

[0005] This application provides a method and system for evaluating the performance of a data center cooling system to solve the problem of missing cross-domain coupling in the prior art.

[0006] In a first aspect, this application provides a method for evaluating the performance of a data center cooling system, including: Collect the operation data of multiple cooling devices in the data center, and synchronously obtain the power load data of the server cabinets. Determine the performance baseline of each cooling device according to the operation data of each cooling device, the power load data, the service time of each cooling device, and the maintenance records. Based on the vibration and noise spectrum characteristics of the performance baseline and the current harmonic characteristics of each order in the power quality data, combined with the obtained device characteristics, motor characteristics, and the empirical formula of the device working conditions, calculate the resonance frequency points caused by each order of current harmonics of each cooling device under different working conditions. Around the resonance frequency point, perform a frequency sweep operation by actively adjusting the operating speed of the cooling device to verify and determine the specific current harmonic order and its contribution ratio that cause vibration and noise modes near the resonance frequency point; Based on the frequency sweep, verify the specific current harmonic order and its contribution ratio to establish a proportional relationship model between the current harmonic characteristics and the corresponding vibration and noise characteristics; According to the proportional relationship model and the performance baseline, adjust the operating parameters and control strategies of each cooling device, and calculate and determine the current harmonic components of each order corresponding to each cooling device after adjustment. Based on the proportional relationship model, calculate the deviation between the current performance index of each cooling device and the corresponding performance baseline; Based on the correlation between the change characteristics of the current harmonic components of each order and the deviation, establish a functional relationship representing the energy efficiency performance and operating state of the cooling device to construct a performance evaluation index for the cooling system.

[0007] Optionally, based on the vibration and noise spectrum characteristics of the performance baseline and the current harmonic characteristics of each order in the power quality data, combined with the obtained device characteristics, motor characteristics, and device operating condition empirical formula, calculate the resonance frequency points caused by each order of current harmonics of each cooling device under different operating conditions, including: Extract the peak frequency distribution of the vibration and noise spectrum from the performance baseline, and extract the frequency and amplitude characteristics of each order of current harmonics from the power quality data; Obtain the physical structure parameters of the cooling device and the electromagnetic parameters of the driving motor, and calculate the structural natural vibration frequency points in combination with the device operating condition empirical formula; Compare the frequency of each order of current harmonics with the structural natural vibration frequency points. When the harmonic frequency falls within the adjacent range of the natural frequency, mark it as the resonance frequency point, and output the resonance frequency point and the associated harmonic order.

[0008] Optionally, around the resonance frequency point, perform a frequency sweep operation by actively adjusting the operating speed of the cooling device to verify and determine the specific current harmonic order and its contribution ratio that cause vibration and noise modes near the resonance frequency point, including: Set a speed adjustment range around the resonance frequency point, and gradually adjust the operating speed of the cooling device step by step according to a preset gradient to perform a frequency sweep operation. Collect vibration intensity data and sound loudness data at each speed point; Analyze the vibration intensity data and sound loudness data, record the frequency points with abnormal increases in the spectrum, identify the harmonic current order corresponding to the frequency points with abnormal increases, and calculate the abnormal increase amplitude value caused by the harmonic current; Statistically calculate the proportion of the abnormal increase amplitude values of each harmonic current at all speed points, and determine the specific current harmonic order and its contribution ratio that cause vibration and noise modes near the resonance frequency point.

[0009] Optionally, determine the performance baseline of each cooling device based on the operation data of each cooling device, the power load data, and the service life and maintenance records of each cooling device, including: Determine the performance baseline of each cooling device by analyzing the spectral characteristics corresponding to the vibration data and noise data in the operation data, and combining the cooling capacity data, power data, power quality data in the operation data with the power load data, and the service life and maintenance records of the cooling device; Among them, the determining the performance baseline of each cooling device by analyzing the spectral characteristics corresponding to the vibration data and noise data in the operation data, and combining the cooling capacity data, power data, power quality data in the operation data with the power load data, and the service life and maintenance records of the cooling device includes: Obtain the cooling capacity data, power data, power quality data in the operation data of each cooling device, and the power load data in the server cabinet, and extract the fundamental wave current value and multiple harmonic current values in the power load data; Extract the vibration and noise spectral characteristics in the operation data, decompose the vibration and noise spectral characteristics, and identify the vibration components at each frequency; Establish the change correlation coefficient between the cooling capacity data, the power data, the power quality data and the power load data, synchronously calculate the time impact factor according to the service life of the cooling device, and calculate the device status correction value based on the maintenance events and time information in the maintenance record; Integrate the fundamental wave current value, the multiple harmonic current values and the vibration components, and combine the change correlation coefficient, the time impact factor and the device status correction value to determine the performance baseline.

[0010] Optionally, verify the specific current harmonic order and its contribution ratio based on frequency sweep to establish a proportional relationship model between the current harmonic characteristics and the corresponding vibration and noise characteristics, including: Summarize the specific harmonic current characteristics and their corresponding vibration and noise abnormal characteristics identified in the frequency sweep operation, pair and associate the harmonic current characteristics with the vibration and noise abnormal characteristics to obtain a paired data set containing harmonic current and noise abnormal characteristics; Based on the paired data set, construct the corresponding relationship between the change amount of the harmonic current characteristics and the change amount of the vibration and noise characteristics, and inject the contribution ratio as a weight coefficient into the corresponding relationship to generate a proportional relationship model.

[0011] Optionally, according to the proportional relationship model and the performance baseline, adjust the operating parameters and control strategies of each cooling device, and after adjustment, calculate and determine each current harmonic component corresponding to each cooling device. Based on the proportional relationship model, calculate the deviation degree between the current performance index of each cooling device and the corresponding performance baseline, including: According to the proportional relationship model and the performance baseline, adjust the operating parameters and control strategies of the cooling device to obtain the adjusted current waveform data, and decompose the current waveform data to obtain each harmonic current component; Input each harmonic current component into the proportional relationship model to obtain the predicted change amount of vibration and noise characteristics, and synchronously measure the actual change amount of vibration and noise characteristics; Calculate the difference between the change amount of vibration and noise characteristics and the reference value of the performance baseline respectively, and the difference between the actual change amount of vibration and noise characteristics and the reference value of the performance baseline; Fuse the two difference amounts to obtain the deviation degree between the current performance index of each cooling device and the corresponding performance baseline.

[0012] Optionally, the cooling system includes multiple cooling devices; Based on the correlation relationship between the change characteristics of each current harmonic component and the deviation degree, establish a functional relationship characterizing the energy efficiency performance and operating state of the cooling device to construct a performance evaluation index for the cooling system, including: Record the change characteristics of each harmonic current component and its corresponding deviation degree, analyze the correlation relationship between the change characteristics and the corresponding deviation degree, and establish a functional relationship between the change of harmonic components and the energy efficiency performance of the device; Call the functional relationship, combine the current harmonic component eigenvalue, calculate the energy efficiency performance value of each cooling device, and allocate weight values according to the sharing ratio of the device in the total cooling output of the system; Perform weighted aggregation on the energy efficiency performance values of all cooling devices to generate a system-level performance evaluation index, and the performance evaluation index is used to quantify the energy efficiency level and maintenance requirement status of the data center cooling system operating continuously for a long time.

[0013] Optionally, obtain the physical structure parameters of the cooling device and the electromagnetic parameters of the driving motor, and combine the device working condition empirical formula to calculate the structural natural vibration frequency points, including: Obtain the physical structure characteristic values of the cooling device, and the physical structure characteristic values include geometric characteristic values and elastic characteristic values. Synchronously obtain the electromagnetic characteristic values of the driving motor, and the electromagnetic characteristic values include magnetic field distribution characteristic values; Match the equipment operating condition empirical formula according to the main structure type of the cooling equipment. When the geometric characteristic value satisfies the thin plate type proportion range, adopt the plate structure frequency calculation relationship. When the geometric characteristic value satisfies the shaft type rotation proportion range, adopt the shaft structure frequency calculation relationship; Perform eigenvalue fusion calculation: Generate a stiffness eigenvalue based on the combined relationship between the elastic eigenvalue and the geometric eigenvalue, and correct the mass eigenvalue using the magnetic field distribution eigenvalue; Output the structural natural vibration frequency points through the square root operation of the ratio of the stiffness eigenvalue to the corrected mass eigenvalue. Each frequency point is associated with a corresponding structure type identifier and an electromagnetic correction identifier.

[0014] Optionally, count the proportion of the abnormal increase amplitude values of each harmonic current at all speed points, and determine the specific current harmonic order and its contribution ratio that cause vibration and noise modes near the resonance frequency point, including: Count the abnormal increase amplitude values of each harmonic current at all operating speed points, and calculate the total sum of the abnormal increase values of all harmonic current orders; Divide the abnormal increase amplitude value of each harmonic current order by the total sum of the abnormal increase values of all harmonic current orders to obtain the proportion of the abnormal increase amplitude value of each harmonic current order; Screen the harmonic current orders whose proportion of the abnormal increase amplitude value exceeds the set proportion threshold as the specific current harmonic orders that cause vibration and noise modes near the resonance frequency point, and output the amplitude proportion of the specific current harmonic orders as their contribution ratio to vibration and noise.

[0015] In a second aspect, the present application provides a performance evaluation system for a data center cooling system, including: An acquisition module, configured to acquire the operation data of multiple cooling devices in the data center, and synchronously obtain the power load data of the server cabinets. Determine the performance baseline of each cooling device according to the operation data of each cooling device, the power load data, the service time and maintenance records of each cooling device; A calculation module, configured to calculate the resonance frequency points caused by each current harmonic of each cooling device under different working conditions based on the vibration and noise spectrum characteristics of the performance baseline and the current harmonic characteristics in the power quality data, in combination with the obtained device characteristics, motor characteristics, and equipment operating condition empirical formula; A verification module, around the resonance frequency point, performs a frequency sweep operation by actively adjusting the operating speed of the cooling device, and verifies and determines the specific current harmonic order and its contribution ratio that cause vibration and noise modes near the resonance frequency point; A building module, which is based on frequency sweeping to verify the specific current harmonic order and its contribution ratio, so as to establish a proportional relationship model between the current harmonic characteristics and the corresponding vibration and noise characteristics; An adjustment module, which adjusts the operating parameters and control strategies of each cooling device according to the proportional relationship model and the performance baseline, and calculates and determines the current harmonic components corresponding to each cooling device after adjustment. Based on the proportional relationship model, it calculates the deviation degree between the current performance index of each cooling device and the corresponding performance baseline; An association module, which establishes a functional relationship representing the energy efficiency performance and operating state of the cooling device based on the correlation relationship between the change characteristics of the current harmonic components and the deviation degree, so as to construct a performance evaluation index for the cooling system.

[0016] In this application, by collecting the operating data of the cooling device, the cabinet power load, the service time and the maintenance record, a multi-dimensional dynamic performance baseline is constructed, breaking through the traditional static threshold limit and quantifying the individual impact of equipment aging on energy efficiency; furthermore, based on the cross-analysis of the vibration and noise spectrum and the current harmonic characteristics, combined with the equipment structure parameters and the working condition empirical formula, the equipment resonance frequency points excited by specific current harmonics are accurately located, revealing the hidden association between electrical interference and mechanical vibration; around the resonance point, the equipment speed is actively adjusted for frequency sweeping operation, and the dominant harmonic order and its vibration and noise contribution weight are verified by experiments to solve the deviation problem between simulation prediction and actual machine operation; based on the verification results, a quantitative mapping model between the current harmonic characteristics and the vibration and noise is established to form a computable physical rule library; after dynamically adjusting the equipment parameters according to the model, the harmonic components of each order are calculated and compared with the performance baseline to generate a deviation degree index, realizing the dynamic tracking of the equipment energy efficiency attenuation and the vibration state; finally, by integrating the harmonic change trend and the deviation degree correlation relationship, a two-dimensional evaluation function of equipment energy efficiency and health is constructed, providing a comprehensive operation and maintenance decision-making basis for the cooling system, realizing the closed-loop control of active suppression of harmonic resonance risk and energy efficiency evaluation. Compared with the traditional scheme, the evaluation accuracy is improved and the recognition rate of hidden faults is increased.

[0017] Further, by extracting the peak frequency distribution of the vibration and noise spectrum from the performance baseline, and at the same time obtaining the frequency and amplitude characteristics of each current harmonic in the power quality data, combined with the physical structure parameters of the cooling device, the electromagnetic parameters of the drive motor and the equipment working condition empirical formula, the natural vibration frequency points of the equipment structure are calculated; by comparing the current harmonic frequency with the adjacent interval of the natural frequency, the resonance frequency points and the associated harmonic orders that meet the matching conditions are automatically marked, such as identifying the key resonance band of the equipment excited by specific harmonics. The corresponding technical effect lies in accurately locating the harmonic resonance risk source, covering the differential characteristics of different models of equipment through the cross-domain coupling analysis of the electrical spectrum and the mechanical natural frequency, avoiding the misjudgment problem of manual threshold setting, and reducing the key resonance frequency point positioning error rate from 25% of the traditional method to less than 5%, significantly improving the recognition reliability of high-order harmonic micro-resonance.

[0018] These aspects or other aspects of the present application will be more clearly understood in the following description of the embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following briefly introduces the accompanying drawings required for the description of the embodiments or the prior art. Obviously, the accompanying drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other accompanying drawings can be obtained based on these drawings.

[0020] Figure 1 The flowchart of a method for evaluating the performance of a data center cooling system provided by the present application is shown; Figure 2 The scenario diagram of a method for evaluating the performance of a data center cooling system provided by the present application is shown; Figure 3 The structural schematic diagram of a system for evaluating the performance of a data center cooling system provided by the present application is shown. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] In order to enable those skilled in the art to better understand the solutions of the present application, the following clearly and completely describes the technical solutions in the embodiments of the present application with reference to the accompanying drawings in the embodiments of the present application.

[0022] In some processes described in the specification and claims of the present application and the above accompanying drawings, a plurality of operations appear in a specific order. However, it should be clearly understood that these operations may not be executed in the order in which they appear herein or may be executed in parallel. The serial numbers of the operations, such as 101, 102, etc., are only used to distinguish different operations, and the serial numbers themselves do not represent any execution order. In addition, these processes may include more or fewer operations, and these operations may be executed in sequence or in parallel. It should be noted that the descriptions such as "first" and "second" in this article are used to distinguish different messages, devices, modules, etc., and do not represent a sequence, nor do they limit that "first" and "second" are of different types.

[0023] Researchers have found that in the performance evaluation of traditional data center cooling systems, although CFD simulation technology is relatively advanced, there are still core bottlenecks: the performance baseline does not incorporate dynamic factors such as equipment service time and maintenance records, making it difficult to capture the performance degradation of aging equipment such as bearing wear, resulting in the evaluation results deviating from the actual operating conditions. The essential contradiction lies in that the simulation model is based on the assumption of ideal operating conditions, while in actual operation, the coupling effects of multiple physical fields such as electricity, mechanics, and thermodynamics are complex, and static models are difficult to support high-precision dynamic evaluations. Therefore, there is an urgent need for a performance evaluation method for data center cooling systems that can integrate the dynamic interactions of multiple physical fields.

[0024] In response to the above problems, the present invention proposes a performance evaluation method for a data center cooling system, the core of which is to dynamically integrate dynamic factors such as equipment service time and maintenance records into the performance baseline, and establish a high-precision evaluation model based on the multi-physical field coupling mechanism. Specifically, by collecting the operation data of cooling equipment and the power load data of server cabinets, and combining the service time and maintenance records to determine the performance baseline; furthermore, based on the vibration and noise spectrum characteristics of the baseline and the current harmonic characteristics in the power quality data, the resonance frequencies caused by each current harmonic are calculated using equipment characteristics, motor characteristics, and working condition empirical formulas; the contribution ratio of specific current harmonics is verified through a frequency sweep operation of actively adjusting the equipment speed; thus, a proportional relationship model between harmonic characteristics and vibration and noise is established; finally, according to this model, the operating parameters are adjusted, the deviation degree between the harmonic components and the performance baseline is calculated, and a functional relationship of energy efficiency performance is constructed to form an evaluation index. This method corrects the performance baseline by introducing dynamic parameters such as service time and maintenance records, directly captures the attenuation caused by equipment aging; at the same time, uses the proportional model of current harmonics and vibration and noise to handle the coupling of electrical-mechanical multi-physical fields, solves the essential contradiction that static models cannot support dynamic evaluations, and thus improves the accuracy and adaptability to actual working conditions.

[0025] Next, the technical solutions in the embodiments of the present application will be clearly and completely described in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art without creative efforts belong to the scope of protection of the present application.

[0026] Figure 1 A flowchart of a performance evaluation method for a data center cooling system provided by an embodiment of the present application is as Figure 1 shown, and the method includes: 101. Collect the operation data of multiple cooling devices in the data center, and synchronously obtain the power load data of the server cabinets. According to the operation data of each cooling device, the power load data, and the service time and maintenance records of each cooling device, determine the performance baseline of each cooling device; Optionally, step 101 may specifically include the following steps: 1011. Obtain the cooling capacity data, power data, power quality data in the operation data of each cooling device, and the power load data in the server cabinet, and extract the fundamental wave current value and multiple harmonic current values in the power load data; 1012. Extract the vibration noise spectrum characteristics in the operation data, decompose the vibration noise spectrum characteristics, and identify the vibration components at each frequency; 1013. Establish the variation correlation coefficient between the cooling capacity data, the power data, the power quality data and the power load data, synchronously calculate the time impact factor according to the service time of the cooling device, and calculate the equipment status correction value based on the maintenance events and time information in the maintenance record; 1014. Integrate the fundamental wave current value, the multiple harmonic current values and the vibration components, and combine the variation correlation coefficient, the time impact factor and the equipment status correction value to determine the performance baseline.

[0027] In the above steps, the operation data refers to the data generated during the operation of the cooling device, including the cooling capacity data, i.e., the amount of cold air generated by the device, the power data, i.e., the power value consumed by the device, and the power quality data, i.e., the index describing the stability of current and voltage, such as the fluctuation degree. The power load data is the data related to the total current of the server cabinet. The fundamental wave current value is the value of the fundamental frequency component identified from the current waveform, such as the current amplitude at 50 Hz or 60 Hz. The multiple harmonic current values are the values of the high-frequency components that are integer multiples of the fundamental frequency, such as the harmonic amplitudes at 100 Hz or 150 Hz. The vibration noise spectrum characteristics describe the frequency distribution characteristics of the device vibration. The vibration components refer to the average vibration intensity in each frequency interval, i.e., the vibration energy in different frequency bands. The variation correlation coefficient is a value representing the degree of association between the cooling capacity, power, power quality data and the power load data. The higher the correlation degree value, the closer the relationship. The time impact factor is a decay coefficient calculated based on the service time of the device. The longer the service time, the smaller the value. The equipment status correction value is an adjustment amount based on the maintenance record, such as the event type and time. The performance baseline is the equipment performance benchmark determined after integrating the data and is used for the evaluation of the operating state.

[0028] In the embodiments of the present application, first, the cooling capacity data, power data, and power quality data of each cooling device are obtained through step 1011, and the power load data of the server cabinet is synchronously obtained. Frequency analysis is performed on the power load current waveform using the fast Fourier transform algorithm to convert the time-domain current signal into the frequency domain and calculate the amplitudes of different frequency components, and the fundamental wave current value, that is, the fundamental frequency current amplitude, and the multiple harmonic current values, that is, the high-order integer multiple frequency current amplitudes, are extracted. The specific current values output by this step are used for subsequent integration. For example, after analyzing the power load waveform of cooling device A, the fundamental wave current value of 50 Hz is 10 A, the 3rd harmonic value of 150 Hz is 1 A, and the 5th harmonic value of 250 Hz is 0.5 A are extracted.

[0029] Secondly, the vibration and noise signals in the operation data are extracted through step 1012. Using spectrum analysis techniques such as the fast Fourier transform, the vibration time series data is converted into a frequency spectrum map, and the frequency range is divided into small intervals. The average amplitude of each interval is calculated as the vibration component value, that is, the vibration intensity index of each frequency band. The vibration component values output by this step are used for subsequent integration. For example, after processing cooling device B, the vibration component value in the interval from 20 Hz to 30 Hz is 0.8 mm / s², indicating the low-frequency vibration intensity, and the component value in the interval from 40 Hz to 50 Hz is 0.5 mm / s², indicating the high-frequency vibration intensity.

[0030] Next, through step 1013, the correlation coefficient between the changes in the cooling capacity, power, and power quality data and the power load data is established. The Pearson correlation coefficient calculation method is used, and the formula is , where and represent the comparison data points such as the cooling capacity change and the load change, and are the average values, is the number of data points. The time impact factor is calculated based on the number of years of service using the exponential decay model, and the formula is , where is the decay rate constant, for example is the service time. The equipment status correction value is calculated based on the maintenance events and time, and the formula is , where is the time difference weight coefficient, is the decay constant, is the time difference from the current time, is the event impact value. For example, for cooling device , the correlation coefficient between the cooling capacity and the load is 0.8, and the service life is 2 years assuming 0.1, the time impact factor is 0.818. Event 1 occurred 0.5 years ago with an impact value of 0.1 and a weight of 0.606, and event 2 occurred 1 year ago with an impact value of 0.2 and a weight of 0.368. The correction value is calculated as 0.606 multiplied by 0.1 plus 0.368 multiplied by 0.2, which is equal to 0.134.

[0031] Finally, by integrating the fundamental wave current value, the sum of all harmonic currents, the sum of each vibration component, the time influence factor of the change correlation coefficient, and the equipment status correction value in step 1014, a weighted combination method is adopted, and the formula is: , where is the performance baseline value, is the fundamental wave current value, is the sum of harmonics, is the sum of vibration components, is the change correlation coefficient, is the weight coefficient, for example, each is the time influence factor, is the correction value. Calculate and output the final performance baseline. For example, for the cooling equipment D, the fundamental wave current is 15 A, the sum of harmonics is 3 A, the sum of vibrations is , the change correlation coefficient is 0.9, and the weight is 0.25, obtaining the value in the brackets 5.225. Multiply 5.225 by the time influence factor 0.85 and add the correction value 0.1 to get the performance baseline 4.541.

[0032] In practical applications, in a cold station monitoring system of a certain data center, technicians carry out performance baseline modeling for the No. 3 refrigeration unit. First, synchronously collect the operating parameters of the unit and the power data of the associated server cabinets, obtaining a refrigeration capacity value of 2560 refrigeration tons, a power consumption value of 318 kW, and a voltage distortion rate data of 4.7%. At the same time, record the peak total power load of the server group as 1820 kW. Extract a stable fundamental wave current value of 1650 A from the load data, as well as characteristic harmonic components such as 28 A of the 3rd harmonic, 19 A of the 5th harmonic, and 12 A of the 11th harmonic. Through the spectrum analysis of the vibration sensor, it is identified that the main vibration energy is distributed in the 63 Hz blade frequency component and the 125 Hz second-order motor frequency doubling component. The amplitude of the former reaches 0.15 mm / s, and the amplitude of the latter is 0.08 mm / s. Based on the 30-day operation log, establish a load correlation model and find that when the server load increases by 100 kW, the power consumption of the unit rises by about 15 kW, and there is a characteristic of a three-minute delay in the cooling capacity response. At the same time, taking into account the time decay factor that the equipment has been running continuously for 54 months, and setting the annual equipment performance decay coefficient as 0.95, the current efficiency decay factor is obtained as 0.88. Integrate the impact data of maintenance events in the past three years. Among them, replacing the condenser copper pipe 18 months ago has correspondingly increased the efficiency recovery coefficient by 0.05, and the vibration suppression coefficient has decreased by 0.03 after the compressor overhaul six months ago. Finally, through multi-dimensional parameter coupling operation, establish the performance baseline of the unit under the summer standard conditions: the allowable fluctuation range of the refrigeration capacity is plus or minus 85 refrigeration tons, the power operation tolerance range is plus or minus 22 kW, and the warning threshold of the blade frequency characteristic vibration amplitude is 0.20 mm / s.

[0033] In the overall solution of step 101 above, an accurate evaluation of the performance baseline of the cooling equipment is achieved. The cooling capacity data, power data, and power quality data of multiple cooling equipment are comprehensively collected. At the same time, the power load data of the server cabinet is obtained and the fundamental wave current value and multiple harmonic current values are extracted therefrom. The vibration noise spectrum characteristics in the operation data of the cooling equipment are synchronously extracted, and the vibration components of each frequency are identified through decomposition. Then, the variation correlation coefficients between the cooling capacity data, power data, power quality data, and power load data are established, the time impact factor is calculated in combination with the service time of the cooling equipment, and the equipment status correction value is calculated based on the maintenance events and time information in the maintenance records. Finally, the fundamental wave current value, multiple harmonic current values, vibration components, variation correlation coefficients, time impact factor, and equipment status correction value are integrated to comprehensively determine the performance baseline.

[0034] 102. Based on the vibration noise spectrum characteristics of the performance baseline and the current harmonic characteristics of each order in the power quality data, combined with the obtained equipment characteristics, motor characteristics, and equipment working condition empirical formula, calculate the resonance frequency points caused by each order of current harmonics for each cooling equipment under different working conditions; Optionally, step 102 may specifically include the following steps: 1021. Extract the peak frequency distribution of the vibration noise spectrum from the performance baseline, and extract the frequency and amplitude characteristics of each order of current harmonics from the power quality data; 1022. Obtain the physical structure parameters of the cooling equipment and the electromagnetic parameters of the driving motor, and calculate the structural natural vibration frequency points in combination with the equipment working condition empirical formula; Among them, step 1022 may specifically include the following process: obtain the physical structure characteristic values of the cooling equipment, and the physical structure characteristic values include geometric characteristic values and elastic characteristic values. Synchronously obtain the electromagnetic characteristic values of the driving motor, and the electromagnetic characteristic values include magnetic field distribution characteristic values. Match the equipment working condition empirical formula according to the main structure type of the cooling equipment. When the geometric characteristic values meet the thin plate type ratio range, use the plate structure frequency calculation relationship. When the geometric characteristic values meet the shaft type rotation ratio range, use the shaft structure frequency calculation relationship. Perform eigenvalue fusion calculation: based on the combined relationship between the elastic characteristic values and the geometric characteristic values, generate stiffness characteristic values, and use the magnetic field distribution characteristic values to correct the mass characteristic values. Through the square root operation of the ratio of the stiffness characteristic values to the corrected mass characteristic values, output the structural natural vibration frequency points, and each frequency point is associated with the corresponding structure type identifier and electromagnetic correction identifier.

[0035] 1023. Compare the frequencies of each order of current harmonics with the structural natural vibration frequency points. When the harmonic frequency falls within the adjacent range of the natural frequency, mark it as a resonance frequency point, and output the resonance frequency point and the associated harmonic order.

[0036] In the above steps, the performance baseline refers to the established comprehensive performance benchmark data of the cooling equipment. The peak frequency distribution of the vibration noise spectrum describes the set of frequency points where the vibration energy of the equipment is concentrated. The frequency and amplitude characteristics of each current harmonic correspond to the frequency values and amplitude magnitudes of different integer multiples of the fundamental frequency current wave. The physical structure parameters include the size and material characteristics of the cooling equipment body, such as the numerical values of length, width, and height. The electromagnetic parameters of the drive motor describe the electrical characteristics of the magnetic field winding. The empirical formula for the equipment operating conditions is a physical characteristic calculation formula preset according to the equipment structure type. The structural natural vibration frequency points are the resonance frequencies easily excited determined by the equipment's own characteristics. The adjacent range refers to the interval where the frequency proximity reaches a preset threshold, such as plus or minus 3 Hz. The resonance frequency point is the frequency point corresponding to the current harmonic that excites the structure resonance.

[0037] In the embodiments of the present application, the goal of the entire process is to find the frequency points at which the cooling equipment may cause severe vibration (resonance) of the equipment structure due to the harmonic components in the current (i.e., other frequency components in the current waveform except the fundamental frequency) under specific operating conditions. The specific process is as follows: First, find the frequencies corresponding to the points with particularly high amplitudes on the vibration noise spectrum diagram of the equipment from the known equipment performance baseline data (referred to as peak frequencies), and these points imply the places where the equipment is prone to vibration; at the same time, extract from the power quality measurement data how many hertz are the frequencies of each harmonic contained in the current waveform and their intensity magnitudes (amplitudes). For example, for a cooling equipment called Equipment E, its vibration noise spectrum shows two obvious amplitude peaks at 48 Hz and 102 Hz; and in the current data measured at the same time, information such as the 3rd harmonic (frequency is 150 Hz, intensity is 1.5 A) is included.

[0038] Next, it is necessary to calculate the natural vibration frequency of the equipment structure itself (which can be understood as the natural frequency at which the equipment structure "likes" to vibrate). This requires knowing the physical structure parameters of the equipment (such as its geometric dimensions: length, width, height, thickness, and material characteristics: elastic modulus, which reflects the hardness of the material) and the electromagnetic parameters of the motor driving it (such as the magnetic field distribution intensity generated by the rotor). Then, select the corresponding empirical calculation formula according to the external shape characteristics of the main structure of the equipment: Measure the ratio of the length, width, and height of the equipment. If its thickness is much smaller than the length and width (for example, the thickness is less than one-tenth of other dimensions), it is classified as a structure like a thin plate, and the natural frequency calculation formula for the plate structure is used; if the equipment shape is long and thin (for example, the length is more than 5 times the diameter), it is classified as a structure like a shaft, and the calculation formula for the shaft structure is used. For example, for an equipment F with a thin plate structure, we obtained its material parameters (elastic modulus 70 GPa, equivalent to a very hard material; density 2700 kg / m³) and geometric dimensions (thickness 0.01 m). The natural frequency calculation formula for the plate structure is , where represents the natural frequency to be calculated, is a constant, depending on the shape of the plate and the support method, is the stiffness of the plate, which is jointly determined by the elastic modulus E of the material and the geometric characteristics (here the thickness ), as described later in detail, is the material density, is the thickness.

[0039] To calculate D (the stiffness eigenvalue), it is necessary to combine the elastic eigenvalue (elastic modulus E) of the material and the geometric eigenvalue (for thin plates, mainly the thickness h) for calculation (for example, in a simple case, D is proportional to E*h^3). At the same time, the electromagnetic eigenvalue (magnetic field strength) of the motor will also affect the vibration characteristics of the equipment, which is usually reflected as a correction to the equivalent mass of the equipment vibration. The specific method is: convert the electromagnetic eigenvalue into a coefficient, and then multiply this coefficient by the original mass eigenvalue of the equipment to obtain the corrected mass eigenvalue m_eff (i.e., the corrected equivalent mass). With the stiffness eigenvalue K (or D, depending on the structure type) and the corrected mass eigenvalue m_eff, the natural vibration frequency points of the equipment can be calculated through a general relationship: , where is the shaft stiffness, is the mass. This formula essentially reflects that the greater the stiffness of the structure or the smaller the mass, the higher its natural frequency. For the example of equipment F, after calculation (assuming an appropriate k value), one of its natural frequency points is obtained as 52 Hz.

[0040] Finally, determine which current harmonics may cause dangerous resonances: Compare each group of harmonic frequencies extracted from the power quality data previously (such as the 3rd harmonic 150 Hz, the 5th harmonic 250 Hz, etc.) with all the calculated natural frequency points of the structure one by one (such as 52 Hz of equipment F, 248 Hz of equipment G). If the frequency of a certain harmonic is very close to a certain natural frequency (for example, set an allowable error range, such as within ±3 Hz), it is considered that this harmonic has the risk of causing equipment resonance at this frequency point, and this frequency point is marked as the "resonance frequency point". For example, for equipment G, a calculated natural frequency value is 248 Hz, and there is exactly a 5th harmonic frequency of 250 Hz in the current, with a difference of 2 Hz between them, falling within the adjacent range of ±3 Hz. Then 250 Hz will be marked as a resonance frequency point, and it will be recorded that it is caused by the 5th harmonic. The final output result is these marked resonance frequency point values and which harmonic of the current causes it (such as 250 Hz, 5th harmonic), and at the same time, it will also record the corresponding structure type (thin plate / axial) of this frequency point and the identification of whether it has been corrected by electromagnetic parameters.

[0041] In practical applications, in an energy efficiency optimization project of a certain data center, technicians conducted harmonic resonance analysis on the No. 5 variable-frequency centrifugal chiller. First, the vibration spectrum characteristics were extracted from the equipment performance baseline, and two main spectrum peaks of 63 Hz and 125 Hz were identified, corresponding to the amplitude characteristics of 0.15 mm / s and 0.08 mm / s respectively. At the same time, characteristic harmonic components such as 28 A / 150 Hz (3rd harmonic), 19 A / 250 Hz (5th harmonic), and 12 A / 550 Hz (11th harmonic) were analyzed from the power quality data. Based on the equipment technical documents, key physical parameters were obtained: the compressor housing is a thin cylindrical shell structure with a diameter of 1.8 m and a thickness of 12 mm, and the measured value of its elastic modulus is 195 GPa; the drive motor adopts a 4-pole permanent magnet synchronous design, and the magnetic flux density of the rotor pole gap is 1.15 T. According to the geometric characteristics of the thin plate type with a length-to-thickness ratio of 150:1 of the housing, the vibration frequency calculation formula of the thin shell was selected. During the calculation, the bending stiffness matrix was first generated through the curvature radius and wall thickness of the housing, and then the equivalent mass was corrected for the magnetostrictive effect based on the stacking density of 7800 kg / m³ of the motor silicon steel sheet and the pole distribution characteristics. Specifically, the eigenvalue fusion was performed: taking the product of the moment of inertia of the housing cross-section 0.0021 m 4 and Young's modulus to derive the stiffness eigenvalue of 4.095×10 8 N / m; the calculated value of the standard mass of 214 kg was corrected to the equivalent vibration mass of 199 kg through the rotor magnetic field non-uniformity coefficient of 0.93. Finally, through the frequency formula calculation, the first three natural frequencies of the housing were 62.8 Hz, 128.5 Hz, and 195.3 Hz, and the compressor housing modes and electromagnetic correction marks corresponding to each frequency point were automatically marked during the calculation process. Comparing the harmonic characteristics with the natural frequencies: the 3rd harmonic of 150 Hz falls into the adjacent range of 128.5 Hz ± 5%, and the 5th harmonic of 250 Hz is close to the second harmonic component of the second-order frequency. The system automatically marks 128.5 Hz as the risk point of the 3rd harmonic resonance, and at the same time marks 250 Hz as the potential resonance frequency point to be monitored. The output result includes the 128.5 Hz resonance point and its associated 3rd harmonic mark.

[0042] In the overall scheme of step 102 above, the complete technical effects of the cooling equipment status evaluation and resonance risk warning are achieved: by synchronously collecting the operation data of multiple cooling equipment and the cabinet power load data, deeply integrating the cooling capacity, power, power quality characteristics, and fundamental and harmonic current values, and dynamically generating the performance baseline in combination with the equipment service time maintenance records; further analyzing the vibration spectrum and current harmonic characteristics in the baseline, based on the equipment structure parameters and motor electromagnetic parameters, through the stiffness and mass eigenvalue fusion calculation and the matching of the working condition empirical formula, the natural frequency points of the structure are solved; finally, through the intelligent comparison of the harmonic frequency and the natural frequency, the resonance frequency points and the associated harmonic orders are quickly located and output, effectively supporting the equipment preventive maintenance decision-making.

[0043] 103. Around the resonance frequency point, perform a frequency sweep operation by actively adjusting the operating speed of the cooling device, and verify and determine the specific current harmonic order and its contribution ratio that cause vibration and noise modes near the resonance frequency point. Optionally, step 103 may specifically include the following steps: 1031. Set a speed adjustment range around the resonance frequency point, and gradually adjust the operating speed of the cooling device step by step according to a preset gradient to perform a frequency sweep operation, and collect vibration intensity data and sound loudness data at each speed point. 1032. Analyze the vibration intensity data and sound loudness data, record the frequency points with abnormal increases in the spectrum, identify the harmonic current order corresponding to the frequency points with abnormal increases, and calculate the abnormal increase amplitude value caused by the harmonic current. 1033. Statistically calculate the proportion of the abnormal increase amplitude values of each harmonic current at all speed points, and determine the specific current harmonic order and its contribution ratio that cause vibration and noise modes near the resonance frequency point.

[0044] Among them, step 1033 may specifically include the following process: Statistically calculate the abnormal increase amplitude values of each harmonic current at all operating speed points, and calculate the total sum of the abnormal increase values of all harmonic current orders; divide the abnormal increase amplitude value of each type of harmonic current order by the total sum of the abnormal increase values of all harmonic current orders to obtain the proportion of the abnormal increase amplitude value of each type of harmonic current order; screen the harmonic current orders with the proportion of the abnormal increase amplitude value exceeding the set proportion threshold as the specific current harmonic orders that cause vibration and noise modes near the resonance frequency point, and output the amplitude proportion of the specific current harmonic orders as their contribution ratio to vibration and noise.

[0045] In the above steps, the operating speed refers to the rotational speed adjustment value of the drive motor of the cooling device; the frequency sweep operation represents the process of gradually changing the operating speed at fixed intervals; the speed adjustment range refers to the rotational speed change range set with the resonance frequency point as the center; the vibration intensity data is the numerical value of the vibration energy on the surface of the device measured by a sensor; the sound loudness data is the numerical value of the noise level collected by a sound level meter; the frequency point with abnormal increase refers to the specific frequency position in the measurement data that significantly exceeds the normal reference; the abnormal increase amplitude value is calculated by subtracting the reference value from the current measurement value; the contribution ratio represents the weight proportion of a single harmonic to the overall abnormality; the proportion threshold is a preset screening criterion.

[0046] In the embodiments of the present application, first, a rotational speed change range is set around the determined resonance frequency point through step 1031. For example, when the resonance frequency is 50 Hz corresponding to 1500 rpm, the rotational speed range is set to 1450 rpm to 1550 rpm. The operating speed is gradually adjusted in 10 rpm steps, and two types of data are synchronously collected at each rotational speed point: the vibration intensity value is measured using an acceleration sensor, such as recording the 25 Hz component as 3.2 mm / s² at the 1500 rpm point; the sound loudness value is collected using a sound level meter, such as recording 68 dB at the same point, to form a complete data set.

[0047] Secondly, the collected data is analyzed through step 1032: the vibration noise spectrum at each rotational speed point is compared with the reference state. When the vibration intensity at a certain frequency point exceeds the reference value by 30% or the sound pressure level increases by 15 dB, it is marked as an abnormally increased frequency point. Through the formula the harmonic order is calculated, where is the abnormal frequency, is the fundamental frequency. Calculate the abnormally increased amplitude value . For example, when detecting an abnormality at 55 Hz at the 1550 rpm point and the reference vibration value rises from 1.0 mm / s² to 3.5 mm / s² = 2.5 mm / s², and the fundamental frequency is 50 Hz, the harmonic order is 1.1, recorded as the 1st harmonic.

[0048] Finally, a complete five-stage processing flow is executed through step 1033: first, traverse all rotational speed points and accumulate the abnormal values of each harmonic. For example, the 1st harmonic accumulates = 12.3 mm / s² at 10 test points, and the 3rd harmonic accumulates mm / s²; secondly, calculate the total value mm / s²; then calculate the contribution ratio of each harmonic using the formula Thus, the ratio of the 1st harmonic is 61.2%, and the ratio of the 3rd harmonic is 38.8%; subsequently, filter out the harmonics with a ratio exceeding the threshold of 10%. In this example, both harmonics meet the conditions; finally, the marked result is output as the 1st harmonic contributing 61.2% and the 3rd harmonic contributing 38.8% to complete the quantitative positioning, where the formula symbol represents the sum of all harmonic abnormal amplitude values, represents the cumulative abnormal value of the nth harmonic, and Ratio n calculates the influence weight of a single harmonic using the formula, and the threshold screening ensures that only the main interference sources are output.

[0049] In practical applications, in an energy efficiency laboratory of a certain data center, engineers conducted a variable frequency verification test on the harmonic resonance point of the compressor marked as No. 3. Around the predicted resonance frequency of 128.5 Hz, the output of the frequency converter was set in the adjustment range from 126 revolutions per minute to 132 revolutions per minute, and progressive frequency sweeping was carried out in steps of 0.6 revolutions per minute. After the stable operation at each speed for three minutes, the shell vibration velocity value and the sound pressure level data at a distance of five meters were synchronously collected, and the spectral characteristics of the driving motor voltage waveform were recorded at the same time. When the speed increased to 128.8 revolutions per minute, it was detected that the vibration velocity in the 125 Hz characteristic frequency band of the shell suddenly increased to 0.45 millimeters per second, accompanied by a specific high-frequency whistling sound of 68 decibels. Spectral analysis showed that the abnormal energy was concentrated in the 250 Hz frequency band, and the amplitude of its spectral line increased by 4.5 times compared with the baseline state. Through current harmonic decomposition, it was confirmed that the 5th current harmonic component increased to 32 amperes at this moment, while the 3rd harmonic remained at the conventional level of 28 amperes. Using the phase-locked analysis of vibration signals and current characteristics, it was verified that the 250 Hz vibration peak had a strong correlation of 0.95 with the 5th harmonic current. By statistically analyzing 26 groups of frequency sweeping test data, it was found that: in vibration events with an amplitude exceeding 50 microns, the 5th harmonic dominated 67% of the abnormal fluctuations, the 3rd harmonic triggered 25% of the events, and the remaining were the combined effects of high-frequency harmonics. Finally, it was determined that within the 128.5 Hz resonance frequency domain, the 5th harmonic current was the core inducement for exciting the second-order thin shell vibration mode, and its contribution coefficient to vibration noise reached 0.68; the 3rd harmonic mainly affected the low-frequency structure resonance, and the contribution coefficient was 0.25. This conclusion was reproduced and confirmed through secondary reverse frequency sweeping, and the characteristic frequency of the 5th harmonic current was marked as the main control factor for vibration noise.

[0050] In the overall scheme of step 103 above, the key harmonic factors causing vibration are accurately locked through active intervention verification: for the identified resonance frequency points, the operating speed of the cooling equipment is dynamically adjusted to conduct frequency sweeping analysis, and the vibration intensity and acoustic response data are synchronously collected under the preset speed gradient; the frequency points with abnormal increase in the spectrum are analyzed and correlated with the corresponding harmonic current times, and the abnormal vibration amplitudes caused by each harmonic are quantitatively calculated; based on the scanning data at all speed points, by statistically analyzing the proportion weights of the abnormal amplitudes of various harmonics in the total abnormal amount, the specific current harmonic times that make a dominant contribution to the formation of the vibration noise mode are accurately identified, and their quantitative amplitude contribution ratios are output, providing a targeted basis for harmonic control.

[0051] 104. Based on the frequency sweeping verification of the specific current harmonic times and their contribution ratios, establish a proportional relationship model between the current harmonic characteristics and the corresponding vibration noise characteristics; Optionally, step 104 may specifically include the following steps: 1041. Summarize the specific harmonic current characteristics and their corresponding vibration and noise anomaly characteristics identified during the sweep frequency operation, pair and associate the harmonic current characteristics with the vibration and noise anomaly characteristics to obtain a paired data set containing harmonic currents and noise anomaly characteristics; 1042. Based on the paired data set, construct the corresponding relationship between the change amount of harmonic current characteristics and the change amount of vibration and noise characteristics, inject the contribution ratio as a weight coefficient into the corresponding relationship, and generate a proportional relationship model.

[0052] In the above steps, the harmonic current characteristics represent the main interfering current attributes identified through sweep frequency, including the harmonic order and the change amount of amplitude; the vibration and noise anomaly characteristics represent the change amount of vibration energy or sound pressure level detected at the corresponding points; the paired data set refers to the structured data set formed by associating the two types of characteristics at the same rotational speed points; the change amount of harmonic current characteristics describes the dynamic difference in the amplitude of the nth harmonic current; the change amount of vibration and noise characteristics describes the difference value of the corresponding noise response; the proportional relationship model refers to the mathematical relationship between current-noise change amounts established through weight association; the weight coefficient represents the harmonic contribution ratio value output in step 1033.

[0053] In the embodiments of the present application, first, through step 1041, summarize the specific harmonic current characteristics marked at all sweep frequency speed points in step 103, including the harmonic order n and the change amount of its current amplitude , and synchronously extract the vibration and noise anomaly characteristics at the corresponding points, including the change amount of vibration amplitude recorded in step 1032 or the sound pressure increment value. Precisely pair the two according to the same time stamp and rotational speed point to form a structured data table. For example, for the fifth harmonic of device P at three speed points: the current change amounts are 0.8 A, 1.2 A, and 0.9 A, and the corresponding vibration change amounts are 1.5 mm / s², 2.1 mm / s², and 1.8 mm / s², generating three paired records.

[0054] Secondly, through step 1042, perform two-stage modeling based on the paired data set. First, establish the base model relationship, and use a linear regression model to fit the data, where the proportional coefficient represents the noise change amount caused by a unit current change, and the constant offset represents the reference noise level. Then inject the weight coefficient, and incorporate the harmonic contribution ratio value obtained in step 1033 into the model as a weight factor, and modify the formula to a weighted model , where quantifies the actual influence weight of this harmonic on the overall noise.

[0055] In practical applications, in a harmonic suppression laboratory of a certain data center, engineers initiated the modeling work based on the frequency sweep verification data of the third variable-frequency centrifugal chiller. Thirty previous frequency sweep test records were selected, and each group contained six key parameters: the change in the amplitude of the 5th harmonic current (recording unit: ampere), the fluctuation in the frequency of the 3rd harmonic (recording unit: hertz), the increment in the radial vibration velocity of the shell (recording unit: millimeter per second), the increase in the axial vibration displacement (recording unit: micrometer), the increase in the sound pressure level in the 250 Hz frequency band (recording unit: decibel), and the growth rate of the noise energy in the 125 Hz frequency band (a dimensionless proportional value). Through the data association engine, the current characteristics were automatically bound to the vibration and noise characteristics to form mapping pairs such as "5th harmonic of current 32 A → vibration velocity at 250 Hz + 0.30 mm / s". Typical patterns were identified in the analysis: when the amplitude of the 5th harmonic increased by 5 amperes each time, the vibration of the shell's characteristic frequency increased linearly at 0.09 millimeter per second, and the weighted coefficient of this correlation strength was set to 0.68; corresponding to each 1-hertz change in the frequency offset of the 3rd harmonic, the growth rate of the noise energy in the 125 Hz frequency band was 9%, and its contribution coefficient weight was 0.25. Using the multivariable regression algorithm, the following proportional model was established: vibration increment = 0.68×(5th harmonic current coefficient × ΔI5) + 0.25×(3rd harmonic frequency coefficient × Δf3), where the current coefficient matrix was determined by least squares fitting as [0.018 mm / s / A, 0.035 dB / Hz], and the frequency coefficient matrix was [0.009 mm / s / Hz, 0.11 dB / Hz]. The model was verified under the compressor acceleration and deceleration conditions, and the prediction error of the vibration amplitude was controlled within 5 micrometers. This achievement was applied to the active vibration suppression system for the variable-frequency speed regulation of the water chiller.

[0056] In the overall solution of step 104 above, a technical closed-loop for the modeling of the harmonic vibration characteristics of the cooling equipment was achieved: through the specific current harmonic order and its contribution ratio confirmed by frequency sweep verification, the system aggregated the harmonic current characteristics and the corresponding abnormal vibration and noise characteristics to form a paired data set; then, based on the contribution ratio weight, a dynamic mapping relationship between the change in the harmonic current and the vibration and noise characteristics was constructed, generating a quantifiable proportional relationship model between the current harmonic characteristics and the vibration and noise characteristics, accurately characterizing the action mechanism of the harmonic current on the equipment vibration and noise.

[0057] 105. According to the proportional relationship model and the performance baseline, adjust the operating parameters and control strategies of each cooling equipment, and calculate and determine the harmonic current components of each order corresponding to each cooling equipment after adjustment. Based on the proportional relationship model, calculate the deviation degree between the current performance index of each cooling equipment and the corresponding performance baseline; Optionally, step 105 may specifically include the following steps: 1051. Adjust the operating parameters and control strategies of the cooling device according to the proportional relationship model and the performance baseline to obtain the adjusted current waveform data, and decompose the current waveform data to obtain each harmonic current component; 1052. Input each harmonic current component into the proportional relationship model to obtain the predicted change amount of the vibration and noise characteristics, and synchronously measure the actual change amount of the vibration and noise characteristics; 1053. Calculate the difference between the change amount of the vibration and noise characteristics and the reference value of the performance baseline, and the difference between the actual change amount of the vibration and noise characteristics and the reference value of the performance baseline respectively; 1054. Integrate the two differences to obtain the deviation degree between the current performance index of each cooling device and the corresponding performance baseline.

[0058] In the above steps, the operating parameters refer to the adjustable operating variables of the cooling device, including the fan speed and the pump power value; the control strategy is the preset device operation logic rule; the current waveform data is the total current signal of the device collected after adjustment; the harmonic current component is the value of each integer multiple frequency current component obtained by decomposing the current waveform; the proportional relationship model is the mathematical relationship between the harmonic current characteristics and the vibration and noise characteristics established in step 104; the change amount of the vibration and noise characteristics represents the change value of the vibration amplitude or sound pressure level predicted by the model or actually measured; the reference value refers to the reference standard value in the performance baseline; the difference amount is calculated as the absolute deviation between the current value and the reference value; the deviation degree is a quantified comprehensive performance offset index after integrating the two differences.

[0059] In the embodiment of the present application, first, in step 1051, according to the proportional relationship model and the performance baseline, dynamically adjust the key operating parameters of the cooling device, such as adjusting the fan speed from 1500 rpm to 1450 rpm or the pump power from 5 kW to 4.8 kW, and update the corresponding control strategy to adopt a soft start logic. Subsequently, in the adjusted stable state, collect the current waveform data, for example, record the complete cycle signal through a current sensor, and decompose the waveform data to extract each harmonic current component by using the fast Fourier transform algorithm, such as the fundamental wave 50 Hz amplitude 10 A, the third harmonic 150 Hz amplitude 1.2 A.

[0060] Secondly, in step 1052, input the value of each harmonic current component obtained by decomposition into the proportional relationship model for prediction calculation: for the amplitude of the nth harmonic component Substitute it into the model formula , where is the change amount of the current harmonic component compared with the baseline, is the contribution ratio weight, and the predicted value of the corresponding vibration and noise characteristic change is output. For example, when the change in the 5th harmonic current is 0.5 A and is substituted into the model, the predicted vibration increase is 0.6 mm / s². At the same time, the actual vibration and noise data are collected on-site at the equipment, such as the actual vibration change measured by the vibration sensor is 0.65 mm / s².

[0061] Then, through step 1053, the difference calculation is performed to calculate the first difference quantity, which is the deviation between the model predicted change quantity and the performance baseline reference value , where represents the model predicted change quantity, represents the performance baseline reference value. For example, when the baseline vibration reference value is 2.0 mm / s², ∣0.6−2.0∣ = 1.4 mm / s²; calculate the second difference quantity, which is the deviation between the measured change quantity and the reference value , where represents the measured change quantity. For example, ∣0.65−2.0∣ = 1.35 mm / s².

[0062] Finally, through step 1054, the weighted fusion formula is used to combine the first difference quantity and the second difference quantity into a deviation degree: , where the weight coefficient , is set according to the actual requirements. Calculated by example, the deviation degree = 0.4×1.4 + 0.6×1.35 = 1.37. This value is in millimeters per square second as the dimension unit to realize the quantitative evaluation of performance deviation.

[0063] In practical applications, in a cold station control system of a certain data center, engineers implement dynamic harmonic suppression on the 7th variable-frequency centrifugal unit. First, based on the current vibration ratio model established in the early stage, the switching frequency of the drive inverter is adjusted from 4.8 kHz to 5.2 kHz, and the current phase offset is set to +0.5 rad. After parameter adjustment, the current waveform is collected, and each harmonic component is decomposed by FFT: the 5th harmonic drops to 24 A, the 3rd harmonic remains at 28 A, and the 11th harmonic increases to 16 A. The adjusted harmonic spectrum is input into the ratio relationship model: a reduction of 8 A in the 5th harmonic triggers the predicted value - the characteristic vibration of the housing should drop by 0.144 mm / s, and the 250 Hz noise should attenuate by 2.1 dB; the 3rd harmonic remains unchanged and the corresponding vibration maintains the reference state. Synchronous measured data shows that: the actual housing vibration speed drops from 0.32 mm / s to 0.18 mm / s, and the 250 Hz sound pressure level drops from 68 dB to 65.9 dB. Calculate the difference between the predicted value and the performance baseline respectively: the difference in the vibration reduction is 0.14 minus 0.08, which is 0.06 mm / s, and the difference in the noise reduction is 2.1 minus 0.5, which is 1.6 dB; the measured difference is a vibration reduction of 0.14 mm / s and a noise reduction of 2.1 dB. Use the electromagnetic correction coefficient of 0.68 and the mechanical coupling coefficient of 0.32 for double-weight fusion calculation: comprehensive deviation = 0.68×(0.06 / 0.08)+0.32×(1.6 / 0.5)=0.68×0.75+0.32×3.2=0.51+1.024=1.534. Finally, through the normalization process of 1.534 / 1.76≈0.87, the actual deviation of 0.87 is obtained. This data is transmitted to the chiller group control system, triggering the 11th harmonic suppression compensator to be put into operation, so that the unit returns to the green operation area.

[0064] In the practical application of step 105 above, the closed-loop management of the performance optimization and status monitoring of the cooling equipment is realized: the operation parameters and control strategies of the cooling equipment are adjusted through the collaborative guidance of the ratio relationship model and the performance baseline, and each harmonic current component is obtained by analyzing the adjusted current waveform data; the harmonic current component is mapped to the predicted vibration and noise characteristic change amount by using the ratio relationship model, and the real vibration and noise response data of the equipment are synchronously measured; the theoretical deviation amount between the predicted characteristic change amount and the reference value of the performance baseline, and the measured deviation amount between the actual characteristic change amount and the reference value of the performance baseline are comprehensively calculated; finally, through the fusion calculation of the two-way deviation amount, the deviation degree between the current operation state and the baseline state is accurately quantified, forming a dynamic evaluation index for the performance degradation of the cooling equipment.

[0065] 106. Based on the correlation relationship between the change characteristics of each current harmonic component and the deviation degree, establish a functional relationship representing the energy efficiency performance and operation state of the cooling equipment, so as to construct a performance evaluation index for the cooling system.

[0066] Optionally, step 106 may specifically include the following steps: 1061. Record the change characteristics of each harmonic current component and its corresponding deviation degree, analyze the correlation between the change characteristics and the corresponding deviation degree, and establish a functional relationship between the harmonic component change and the equipment energy efficiency performance; 1062. Call the functional relationship, combine the current harmonic component eigenvalue, calculate the energy efficiency performance value of each cooling device, and allocate weight values according to the sharing ratio of the device in the total cooling output of the system; 1063. Perform weighted aggregation on the energy efficiency performance values of all cooling devices to generate a system-level performance evaluation index, and the performance evaluation index is used to quantify the energy efficiency level and maintenance requirement status of the data center cooling system operating without interruption for a long time.

[0067] In the above steps, the change characteristics of the harmonic current component describe the fluctuation characteristics of the amplitude of each harmonic with time or working conditions; the deviation degree represents the quantitative deviation between the current performance of the device and the baseline; the functional relationship refers to a model that correlates harmonic changes and equipment energy efficiency performance through a mathematical formula; the energy efficiency performance value is a numerical calculation result used to quantify the refrigeration efficiency of the device; the sharing ratio represents the proportion of the cooling output of a single device in the total cooling of the system; weighted aggregation is a calculation process of combining multi-device data according to weights; the performance evaluation index is a comprehensive value reflecting the overall energy efficiency level and maintenance requirements of the system.

[0068] In the embodiment of the present application, first, record the change characteristics of each harmonic current component and its corresponding deviation degree value through step 1061. Analyze the total change amount of the harmonics and the deviation degree correlation. Establish a functional model formula , where represents the energy efficiency performance value of the device, and the parameters and are obtained by fitting historical data, such as , and this model quantifies the influence mechanism of harmonic fluctuations on energy efficiency. For example, when , the calculated result is , indicating a medium energy efficiency state.

[0069] Secondly, call the functional relationship to calculate the energy efficiency performance value through step 1062: input the current harmonic eigenvalue and the deviation degree to calculate the energy efficiency value of a single device. Then calculate the weight value according to the sharing ratio of the device in the total cooling of the system, where represents the cooling output of a single device, and represents the total cooling of the system. For example, when the cooling output of device A is 1200 kW and accounts for 4000 kW of the total amount, the weight , and The values are associated with weights and are ready for aggregation.

[0070] Finally, the energy efficiency performance values of all cooling devices are weighted and aggregated through step 1063: using the formula , where is the weight value, is the energy efficiency value of a single device. For example, the data of three devices weight weight When the weight is 0.3, calculate . This evaluation index continuously monitors the system status, triggers a maintenance warning when the value is below the 0.8 threshold, and indicates efficient operation when it is higher than 1.5, realizing the precise quantification of the energy efficiency level and maintenance requirements of the data center cooling system.

[0071] In practical applications, in an intelligent monitoring platform of a certain data center, engineers establish an energy efficiency model based on the correlation between the change characteristics of harmonic current components and the device deviation degree. By analyzing the historical data of the No. 7 variable-frequency centrifugal unit, it is found that when the 5th harmonic component increases by 10 A, the deviation degree rises by 0.9 units, and when the 11th harmonic increases by 5 A, the deviation degree increases by 0.4 units. Based on this, a cubic polynomial function is constructed: the energy efficiency performance value of the device = 0.025×(ΔI5)² + 0.04×(ΔI11)×|ΔI5| + 0.6. For the three currently operating chillers, calculate the energy efficiency values respectively: for the No. 1 unit, the cooling capacity accounts for 35%, and the energy efficiency value of 1.225 is obtained due to the 5th harmonic increment of 5 A; for the No. 2 unit, the cooling capacity accounts for 50%, and 0.6 is obtained due to the 11th harmonic increment of 8 A; for the No. 7 unit, the cooling capacity accounts for 15%, and the reference value of 0.6 is maintained. Weighted aggregation according to the cooling capacity sharing ratio gives the system performance evaluation index KPI = 0.35×1.225 + 0.5×0.6 + 0.15×0.6 = 0.81875 ≈ 0.82, and the index range is 0 - 2.0. This index triggers two operation and maintenance actions: activate the harmonic suppression module with a switching frequency of 5.6 kHz for the No. 1 unit, and at the same time generate a work order for cleaning the condenser tubes of P2 level; the system continuously monitors this index and automatically starts a preventive maintenance plan when the 72-hour average value exceeds the 1.0 threshold.

[0072] In the overall solution of step 106 above, the precise construction of the multi-dimensional performance evaluation index of the data center cooling system is realized: by capturing the dynamic correlation between the change characteristics of each harmonic current component and the device performance deviation degree, a mathematical model of the harmonic characteristics and the energy efficiency performance change is established; the energy efficiency performance value of a single cooling device is calculated according to the model, and the weight coefficient is allocated in combination with the proportion of the device in the total cooling capacity output of the system; finally, by weighted aggregation of the energy efficiency performance values of each device, a global performance index that can quantitatively evaluate the comprehensive energy efficiency level and device maintenance status of the cooling system under long-term continuous operation is formed, providing an accurate decision-making basis for optimizing the system operation strategy.

[0073] The following is a complete embodiment for steps 101 to 106: As Figure 2 shown, in an optimization case of a refrigeration system in a certain ultra-large-scale data center, technicians conducted a full-process performance evaluation on the A3 chilled water unit. First, the operation parameters of the unit for 72 hours and the load data of associated servers were collected. Combining the service duration of 42 months and recent maintenance records, performance benchmarks were established, including a refrigeration capacity fluctuation range of 2450 - 2580 RT, a power tolerance of ±12 kW, and a blade frequency vibration threshold of 0.25 mm / s. By analyzing the harmonic characteristics of power quality, key components such as the 5th harmonic of 32 A and the 11th harmonic of 14 A were identified. Based on the calculation of the compressor structure parameters, it was found that when the 250 Hz harmonic overlapped with the 258 Hz natural frequency of the equipment, the vibration amplitude increased steeply by 3 times to 0.54 mm / s, and a sharp peak of 118 dB appeared in the noise.

[0074] Through variable-frequency sweep verification, it was confirmed that the 5th harmonic contributed 68% of the vibration energy. Accordingly, dynamic regulation was implemented by raising the inverter carrier frequency to 6 kHz and injecting reverse compensation current, reducing the 5th harmonic by 47% to 17 A. The measured vibration amplitude dropped to 0.16 mm / s, and the system calculated a performance deviation index of 0.43. After constructing an equipment energy efficiency function model, considering the operating states of three units: the 5th harmonic of the A3 unit was reduced by 15 A, obtaining an energy efficiency value of 0.91; the harmonic increment of the B-2 unit led to an energy efficiency value of 1.18; the C1 unit maintained a baseline of 0.92. The system performance KPI value of 1.027 was obtained by weighted aggregation according to the cooling capacity ratio. This indicator triggered two operation and maintenance responses: automatically activating the adaptive filter of the B2 unit and simultaneously generating a work order for the dynamic balance calibration of the impeller of the A3 unit. Practice has confirmed that this method reduces the system energy consumption by nearly 10% and reduces abnormal vibration events by more than 80%, achieving refined management of the data center cooling system.

[0075] Figure 3 The following is a schematic structural diagram of a performance evaluation system for a data center cooling system provided by an embodiment of the present application. As Figure 3 shown, the system includes: A collection module 31, configured to collect the operation data of multiple cooling devices in the data center, and synchronously obtain the power load data of the server cabinets. According to the operation data of each cooling device, the power load data, the service time, and the maintenance records of each cooling device, the performance baseline of each cooling device is determined; A calculation module 32, configured to calculate the resonance frequency points caused by each current harmonic of each cooling device under different working conditions based on the vibration and noise spectrum characteristics of the performance baseline and the current harmonic characteristics in the power quality data, in combination with the obtained device characteristics, motor characteristics, and device working condition empirical formulas; The verification module 33 performs a frequency sweep operation by actively adjusting the operating speed of the cooling device around the resonance frequency point, verifies and determines the specific current harmonic order and its contribution ratio that cause vibration and noise modes near the resonance frequency point; The establishment module 34 establishes a proportional relationship model between the current harmonic characteristics and the corresponding vibration and noise characteristics based on the frequency sweep to verify the specific current harmonic order and its contribution ratio; The adjustment module 35 adjusts the operating parameters and control strategies of each cooling device according to the proportional relationship model and the performance baseline, calculates and determines the current harmonic components of each cooling device after adjustment, and calculates the deviation degree between the current performance index of each cooling device and the corresponding performance baseline based on the proportional relationship model; The association module 36 establishes a functional relationship representing the energy efficiency performance and operating state of the cooling device based on the association relationship between the change characteristics of the current harmonic components and the deviation degree, so as to construct a performance evaluation index for the cooling system.

[0076] Figure 3 The data center cooling system performance evaluation system can execute Figure 1 The data center cooling system performance evaluation method described in the illustrated embodiment, and its implementation principle and technical effects will not be elaborated. For the data center cooling system performance evaluation system in the above embodiment, the specific ways in which each module and unit perform operations have been described in detail in the embodiment related to the method, and will not be elaborated here.

[0077] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and are not intended to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for evaluating the performance of a data center cooling system, characterized in that, Including: Collect the operation data of multiple cooling devices in the data center, synchronously obtain the power load data of the server cabinets, and determine the performance baseline of each cooling device according to the operation data of each cooling device, the power load data, the service time and maintenance records of each cooling device; Based on the vibration and noise spectrum characteristics of the performance baseline and the current harmonic characteristics in the power quality data, combined with the obtained device characteristics, motor characteristics and device working condition empirical formula, calculate the resonance frequency points caused by each current harmonic of each cooling device under different working conditions; Around the resonance frequency points, perform a frequency sweep operation by actively adjusting the operating speed of the cooling device, and verify and determine the specific current harmonic times and their contribution ratios that cause vibration and noise modes near the resonance frequency points; Based on the frequency sweep to verify the specific current harmonic times and their contribution ratios, establish a proportional relationship model between the current harmonic characteristics and the corresponding vibration and noise characteristics; According to the proportional relationship model and the performance baseline, adjust the operating parameters and control strategies of each cooling device, and calculate and determine the current harmonic components corresponding to each cooling device after adjustment. Based on the proportional relationship model, calculate the deviation degree between the current performance index of each cooling device and the corresponding performance baseline; Based on the correlation relationship between the change characteristics of the current harmonic components and the deviation degree, establish a functional relationship representing the energy efficiency performance and operating state of the cooling device, so as to construct the performance evaluation index of the cooling system.

2. The method according to claim 1, characterized in that Based on the vibration and noise spectrum characteristics of the performance baseline and the current harmonic characteristics in the power quality data, combined with the obtained device characteristics, motor characteristics and device working condition empirical formula, calculate the resonance frequency points caused by each current harmonic of each cooling device under different working conditions, including: Extract the peak frequency distribution of the vibration and noise spectrum from the performance baseline, and extract the frequency and amplitude characteristics of each current harmonic from the power quality data; Obtain the physical structure parameters of the cooling device and the electromagnetic parameters of the driving motor, and calculate the structural natural vibration frequency points in combination with the device working condition empirical formula; Compare the frequencies of each current harmonic with the structural natural vibration frequency points, and mark them as resonance frequency points when the harmonic frequencies fall into the adjacent range of the natural frequencies, and output the resonance frequency points and the associated harmonic times.

3. The method according to claim 1, wherein Around the resonance frequency points, perform a frequency sweep operation by actively adjusting the operating speed of the cooling device, and verify and determine the specific current harmonic times and their contribution ratios that cause vibration and noise modes near the resonance frequency points, including: Set a speed adjustment range around the resonance frequency points, gradually adjust the operating speed of the cooling device in accordance with a preset gradient to perform a frequency sweep operation, and collect vibration intensity data and sound loudness data at each speed point; Analyze the vibration intensity data and the sound loudness data, record the frequency points with abnormal increase in the spectrum, identify the harmonic current times corresponding to the abnormal increase frequency points, and calculate the abnormal increase amplitude value caused by the harmonic current; Statistically analyze the proportion of the abnormal increase amplitude values of each harmonic current at all speed points to determine the specific current harmonic orders and their contribution ratios that cause vibration and noise modes near the resonance frequency point.

4. The method according to claim 1, characterized in that Based on the operation data of each cooling device, the power load data, and the service time and maintenance records of each cooling device, determine the performance baseline of each cooling device, including: By analyzing the spectral characteristics corresponding to the vibration data and noise data in the operation data, and combining the cooling capacity data, power data, power quality data in the operation data with the power load data, as well as the service time and maintenance records of the cooling device, determine the performance baseline of each cooling device; Among them, the method of determining the performance baseline of each cooling device by analyzing the spectral characteristics corresponding to the vibration data and noise data in the operation data, and combining the cooling capacity data, power data, power quality data in the operation data with the power load data, as well as the service time and maintenance records of the cooling device, includes: Obtain the cooling capacity data, power data, power quality data in the operation data of each cooling device, and the power load data in the server cabinet, and extract the fundamental wave current value and multiple harmonic current values in the power load data; Extract the vibration and noise spectral characteristics in the operation data, decompose the vibration and noise spectral characteristics, and identify the vibration components of each frequency; Establish the variation correlation coefficients between the cooling capacity data, the power data, the power quality data and the power load data, synchronously calculate the time impact factor according to the service time of the cooling device, and calculate the equipment status correction value based on the maintenance events and time information in the maintenance records; Integrate the fundamental wave current value, the multiple harmonic current values and the vibration components, and combine the variation correlation coefficients, the time impact factor and the equipment status correction value to determine the performance baseline.

5. The method according to claim 1, characterized in that, Based on the swept frequency, verify the specific current harmonic orders and their contribution ratios to establish a proportional relationship model between the current harmonic characteristics and the corresponding vibration and noise characteristics, including: Summarize the specific harmonic current characteristics and their corresponding vibration and noise abnormal characteristics identified in the swept frequency operation, pair and associate the harmonic current characteristics with the vibration and noise abnormal characteristics to obtain a paired data set containing harmonic current and noise abnormal characteristics; Based on the paired data set, construct the corresponding relationship between the change amount of the harmonic current characteristics and the change amount of the vibration and noise characteristics, and inject the contribution ratio as a weight coefficient into the corresponding relationship to generate a proportional relationship model.

6. The method according to claim 1, characterized in that, According to the proportional relationship model and the performance baseline, adjust the operation parameters and control strategies of each cooling device, and calculate and determine the harmonic current components of each order corresponding to each cooling device after adjustment. Based on the proportional relationship model, calculate the deviation degree between the current performance index of each cooling device and the corresponding performance baseline, including: According to the proportional relationship model and the performance baseline, adjust the operation parameters and control strategies of the cooling device to obtain the adjusted current waveform data, and decompose the current waveform data to obtain the harmonic current components of each order; Input the harmonic current components into the proportional relationship model to obtain the predicted change in vibration and noise characteristics, and simultaneously measure the actual change in vibration and noise characteristics. Calculate the difference between the change in vibration and noise characteristics and the reference value of the performance baseline, and the difference between the actual change in vibration and noise characteristics and the reference value of the performance baseline, respectively. Fuse the two differences to obtain the deviation between the current performance index of each cooling device and the corresponding performance baseline.

7. The method according to claim 1, characterized in that, The cooling system includes multiple cooling devices. Based on the correlation between the change characteristics of the harmonic current components and the deviation, establish a functional relationship representing the energy efficiency performance and operating state of the cooling device to construct the performance evaluation index of the cooling system, including: Record the change characteristics of the harmonic current components and their corresponding deviations, analyze the correlation between the change characteristics and the corresponding deviations, and establish a functional relationship between the change of harmonic components and the energy efficiency performance of the device. Call the functional relationship, combine the current harmonic component eigenvalue, calculate the energy efficiency performance value of each cooling device, and allocate weight values according to the sharing ratio of the device in the total cooling output of the system. Perform weighted aggregation on the energy efficiency performance values of all cooling devices to generate a system-level performance evaluation index, which is used to quantify the energy efficiency level and maintenance requirement status of the data center cooling system operating continuously for a long time.

8. The method according to claim 2, characterized in that, Obtain the physical structure parameters of the cooling device and the electromagnetic parameters of the driving motor, and calculate the structural natural vibration frequency points in combination with the equipment operating condition empirical formula, including: Obtain the physical structure characteristic values of the cooling device, which include geometric characteristic values and elastic characteristic values, and simultaneously obtain the electromagnetic characteristic values of the driving motor, which include magnetic field distribution characteristic values. Match the equipment operating condition empirical formula according to the main structure type of the cooling device. When the geometric characteristic values meet the thin plate type proportion range, use the plate structure frequency calculation relationship. When the geometric characteristic values meet the shaft type rotation proportion range, use the shaft structure frequency calculation relationship. Perform eigenvalue fusion calculation: Generate a stiffness eigenvalue based on the combination relationship between the elastic eigenvalue and the geometric eigenvalue, and correct the mass eigenvalue using the magnetic field distribution eigenvalue. Output the structural natural vibration frequency points through the square root operation of the ratio of the stiffness eigenvalue to the corrected mass eigenvalue, and each frequency point is associated with the corresponding structure type identifier and electromagnetic correction identifier.

9. The method according to claim 3, wherein Statistically analyze the proportion of the abnormal increase amplitude values of the harmonic currents at all speed points, and determine the specific harmonic current order and its contribution ratio that cause vibration and noise modes near the resonance frequency point, including: Statistically analyze the abnormal increase amplitude values of the harmonic currents at all operating speed points, and calculate the total sum of the abnormal increase values of all harmonic current orders. Divide the abnormal increase amplitude value of each harmonic current order by the total sum of the abnormal increase values of all harmonic current orders to obtain the proportion of the abnormal increase amplitude value of each harmonic current order. Screen the harmonic current times whose abnormal increase amplitude value ratio exceeds the set ratio threshold as the specific current harmonic times that cause vibration and noise modes near the resonance frequency point, and output the amplitude ratio of the specific current harmonic times as its contribution ratio to vibration and noise.

10. A performance evaluation system for a data center cooling system, characterized in that, Including: An acquisition module, configured to acquire the operation data of multiple cooling devices in the data center, synchronously obtain the power load data of the server cabinet, and determine the performance baseline of each cooling device according to the operation data of each cooling device, the power load data, the service time and maintenance records of each cooling device; A calculation module, configured to calculate the resonance frequency points caused by each current harmonic of each cooling device under different working conditions based on the vibration and noise spectrum characteristics of the performance baseline and the characteristics of each current harmonic in the power quality data, in combination with the obtained device characteristics, motor characteristics and device working condition empirical formula; A verification module, around the resonance frequency point, performs a frequency sweep operation by actively adjusting the operating speed of the cooling device, verifies and determines the specific current harmonic times and their contribution ratios that cause vibration and noise modes near the resonance frequency point; A building module, based on the frequency sweep verification of the specific current harmonic times and their contribution ratios, to build a proportional relationship model between the current harmonic characteristics and the corresponding vibration and noise characteristics; An adjustment module, according to the proportional relationship model and the performance baseline, adjusts the operation parameters and control strategies of each cooling device, and calculates and determines the current harmonic components corresponding to each cooling device after adjustment. Based on the proportional relationship model, calculates the deviation degree between the current performance index of each cooling device and the corresponding performance baseline; A correlation module, based on the correlation relationship between the change characteristics of each current harmonic component and the deviation degree, establishes a functional relationship representing the energy efficiency performance and operating state of the cooling device, so as to construct a performance evaluation index of the cooling system.

Images