Multi-level safety response control method for industrial and commercial liquid-cooled energy storage system

Abstract

The invention discloses a multi-level safety response control method for an industrial and commercial liquid cooling energy storage system, and relates to the technical field of electric energy storage systems. A sensor network collects a time series matrix of operating status data, and performs abnormal data cleaning on the time series matrix; preprocesses the vibration signal of the cooling system, extracts characteristic data of abnormal vibration through Fourier transform and energy distribution analysis, and locates the abnormal area of ​​time-frequency distribution; generates a time series multivariate data set from the characteristic data of the abnormal vibration, determines the cause label with the fused time series multivariate data set, obtains the natural frequency and resonant frequency of the cooling system in combination with simulation analysis, and matches and determines the cause of the abnormal vibration; uses an LSTW neural network to predict the risk of abnormal resonance, and takes corresponding countermeasures for the cooling network according to the prediction result; avoids unnecessary vibration caused by misadjustment, balances the relationship between cooling efficiency and safety, and reduces unnecessary power loss.

Description

Technical Field

[0001] The present invention relates to the technical field of electric energy storage systems, and in particular to a multi-level safety response control method for industrial and commercial liquid cooling energy storage systems. Background Art

[0002] With the rapid development of energy storage technology, liquid-cooled energy storage systems are widely used in data centers, new energy power generation, industrial energy storage, and electric vehicles due to their advantages in efficient heat dissipation, improved energy density, and stable operation. Liquid cooling technology removes heat through the flow of coolant, which can significantly improve the thermal management efficiency of energy storage equipment, extend battery life, and reduce system operating costs. However, in liquid-cooled energy storage systems, complex pipeline design and high-speed coolant flow often cause fluid-induced vibrations, thermal expansion and contraction stress, and multivariable coupling anomalies in equipment operation, posing huge challenges to the safety and reliability of the system.

[0003] Specifically, during the operation of the liquid-cooled energy storage system, the high-speed flow of the coolant is prone to turbulence, cavitation, and local pressure fluctuations at key locations such as bends, high-pressure gradient areas, and pump outlets. These factors can cause significant vibration and mechanical fatigue effects. Especially in the context of the continuous increase in the power density of energy storage equipment, the fluid vibration frequency of the system may be coupled with the natural frequency of the equipment structure, inducing resonance, and causing serious damage to core components such as pipelines, pump bodies, and energy storage units. In addition, liquid-cooled energy storage systems are also faced with operational anomalies caused by the coupling of thermal-mechanical-fluid multi-fields. For example, when the system is operating under high load or rapid charging and discharging conditions, hot and cold cycles may cause thermal fatigue and deformation of the pipeline material, further exacerbating the risk of vibration and leakage hazards.

[0004] In the Chinese invention patent with application publication number CN118655869A, a method for preventing and responding to faults in a liquid-cooled energy storage system based on multi-level safety control is disclosed, which includes the following steps: constructing a hierarchical safety monitoring architecture; based on the hierarchical safety architecture, obtaining system operation data and performing data preprocessing, and predicting faults based on the preprocessed data; automatically triggering emergency response strategies based on the fault prediction results, and optimizing the operation of the system; constructing a data analysis model based on the system's operating data to iteratively optimize the system's performance; defining the failure modes of each component in the hierarchical safety architecture, and configuring the risk priority value of each failure mode, and performing fault processing based on the risk priority value. This solves the technical problem of how to accurately and quickly monitor the faults of liquid-cooled energy storage systems and optimize the fault response capability.

[0005] Combining the above applications and the prior art:

[0006] Existing liquid-cooled energy storage systems have significant deficiencies in dealing with the above-mentioned problems. First, traditional thermal management and vibration control methods are mostly static designs, such as reducing vibration risks by increasing pipeline support points, optimizing structural layout, or manually adjusting the coolant flow rate, but these methods lack real-time and dynamic adaptability, and are difficult to cope with complex and changeable operating conditions. Secondly, existing monitoring technologies usually only focus on single parameters such as coolant temperature, pressure or flow rate, lack comprehensive monitoring of multivariable coupling effects, and are difficult to effectively warn of vibration anomalies or system failures. In addition, when vibration or thermal anomalies cause local failures, the system lacks a multi-level response safety control mechanism, making it difficult to achieve active suppression of vibration problems and hierarchical processing of abnormal situations, which may cause local failures to expand into full system failures, increasing safety risks and maintenance costs of equipment.

[0007] To this end, the present invention provides a multi-level safety response control method for an industrial and commercial liquid-cooled energy storage system. Summary of the invention

[0008] 1. Technical issues to be resolved

[0009] In view of the deficiencies in the prior art, the present invention provides a multi-level safety response control method for industrial and commercial liquid-cooled energy storage systems. The vibration signal of the cooling system is preprocessed, and the characteristic data of the abnormal vibration is extracted through Fourier transform and energy distribution analysis, and the abnormal area of ​​the time-frequency distribution is located; a time series multivariate data set is generated from the characteristic data of the abnormal vibration, and the cause label is determined by the fused time series multivariate data set, and the natural frequency and resonant frequency of the cooling system are obtained in combination with simulation analysis, and the cause of the abnormal vibration is determined by matching; the LSTW neural network is used to predict the risk of abnormal resonance, and corresponding countermeasures are taken for the cooling network according to the prediction results; unnecessary vibration caused by misadjustment is avoided, the relationship between cooling efficiency and safety is balanced, and unnecessary power loss is reduced, thereby solving the technical problems in the background technology.

[0010] (II) Technical solution

[0011] To achieve the above objectives, the present invention is implemented through the following technical solutions: A multi-level safety response control method for industrial and commercial liquid cooling energy storage system, comprising: deploying a sensor network at key nodes of the power energy storage system, and collecting the operation status data time series matrix by the sensor network , for the time series matrix Clean up abnormal data;

[0012] Preprocess the vibration signal of the cooling system, extract the characteristic data of abnormal vibration through Fourier transform and energy distribution analysis, and analyze the abnormal area of ​​time-frequency distribution. To locate;

[0013] Generate a time series multivariate dataset from the characteristic data of abnormal vibration , with the fused time series multivariate dataset Determine the cause label and combine simulation analysis to obtain the natural frequency of the cooling system and resonant frequency , match and determine the cause of abnormal vibration;

[0014] The LSTW neural network is used to predict the risk of abnormal resonance, and corresponding countermeasures are taken for the cooling network based on the prediction results, including dynamic adjustment of the coolant flow rate, adjustment of the support point layout, automatic reduction of the pump speed, suspension of system operation and switching to the backup cooling system.

[0015] Furthermore, different types of sensors are deployed at the key nodes of the liquid cooling system to form a sensor network. The sampling frequencies of all sensors are kept consistent, and the sampling data of all sensors are synchronized according to the timestamp to form a complete time series matrix. , apply preset constraint rules to each data series, remove abnormal points that exceed the physical reasonable range; Perform real-time smoothing to eliminate random noise during data acquisition; Fluctuation of flow rate Conduct joint testing to determine whether there are any abnormal correlations between the two; label and classify all verified data, including normal data, single variable anomalies, and multivariate correlation anomalies.

[0016] Furthermore, the vibration signal is processed to retain the vibration signal within the natural frequency range of the system, and the low-frequency drift and high-frequency noise are removed, and the transient signal is eliminated by using a detection method based on autoregressive residuals;

[0017] Focusing on key vibration characteristics, eliminating transient signals and reducing the impact of abnormal sudden interference on spectrum analysis can improve the reliability of analysis results.

[0018] Furthermore, the normalized signal is subjected to Fourier transform to obtain the amplitude distribution of the vibration signal in the frequency domain; the signal is divided into multiple time windows, and each time window is subjected to spectrum analysis to form a time-frequency diagram, and the vibration signal is subjected to wavelet transform to capture abnormal frequencies in the signal. and abnormal amplitude ;

[0019] Extract abnormal features and identify abnormal patterns based on statistical analysis of the spectrum baseline and time-frequency distribution of normal operating conditions;

[0020] If the spectrum intensity exceeds the base spectrum value, the frequency point is marked as an abnormal frequency point By calculating the local energy density, the time-frequency distribution of abnormal areas to locate.

[0021] Furthermore, characteristic data of abnormal vibrations are imported, including abnormal frequencies. , abnormal amplitude , abnormal time-frequency distribution area , combined with the collected and verified coolant flow rate , pressure gradient and temperature changes , fused to generate a unified time series multivariate dataset : ； The fused time series multivariate dataset As input, the output is the cause label ,in,

[0022] ;

[0023] in is the posterior probability of each cause; is the optimal prediction value.

[0024] Furthermore, based on the fluid mechanics simulation software, the coolant flow characteristics of the pipeline system are simulated and the natural frequency is output. and the resonance frequency under different fluid conditions Distribution, the abnormal frequency The natural frequency obtained from simulation and resonant frequency Matching, the cause of abnormal vibration.

[0025] Furthermore, the multi-source data fusion analysis results are used as input, including the coolant flow rate , pressure gradient and temperature changes , vibration frequency , pressure gradient , predict the probability of obtaining resonance risk , Risk impact range ,like , send a first-level alarm command to the outside. If , send a secondary alarm command to the outside.

[0026] Furthermore, when a first-level alarm command is received, the coolant flow rate is dynamically adjusted. Combined with the relationship between the resonance frequency and the flow rate, the flow rate is adjusted to avoid the dangerous resonance frequency range. The output flow of the pump is dynamically adjusted to gradually adjust the flow rate to the target value. The changes in vibration amplitude and frequency are monitored through real-time feedback to verify the adjustment effect.

[0027] Furthermore, if the coolant flow rate adjustment effect is insufficient, adjust the arrangement of the support points, increase the number of support points in the high-risk area or adjust the support point spacing to , because the arrangement of pipeline support points will directly affect its natural frequency and vibration modes, and establish a support point optimization model by combining fluid mechanics simulation and vibration analysis to calculate the ideal support spacing and stiffness :

[0028] ;

[0029] in, is the support point stiffness, is the pipe mass per unit length, is the support point spacing;

[0030] Determine the optimal parameter combination through particle swarm optimization algorithm , so that the natural frequency and vibration frequency Avoid dangerous resonance ranges.

[0031] Further, when receiving the second-level alarm command, the following operations are performed: replace high-strength materials to increase the support stiffness to , add dampers in high-risk areas to absorb vibration energy, automatically reduce pump speed and suspend system operation, switch to the backup cooling system to ensure the continuity of equipment operation.

[0032] (III) Beneficial effects

[0033] The present invention provides a multi-level safety response control method for an industrial and commercial liquid cooling energy storage system, which has the following beneficial effects:

[0034] 1. Locate the time when abnormal vibration occurs on the time axis, provide accurate time-frequency distribution information, capture abnormal frequency and amplitude, which helps to further distinguish normal working conditions from fault characteristics and improve the recognition efficiency of abnormal patterns; assign the most likely cause label to the vibration anomaly based on the posterior probability model, quickly lock the cause of the fault, improve diagnostic efficiency, and use the prediction results to guide the subsequent optimization of the cooling system and reduce the time for troubleshooting.

[0035] 2. The simulation outputs the distribution of natural frequency and resonant frequency, which can quantify the vibration risk range. In order to optimize the pipeline structure support, the abnormal frequency is matched with the simulation results. It can be clear whether the abnormality is caused by the structural natural frequency or the fluid-induced resonance, so as to implement targeted improvements.

[0036] 3. The resonance risk of the liquid cooling system is dynamically modeled through the long short-term memory network (LSTM), and the timing analysis algorithm is used to capture the precursors of resonance, accurately predict the time range of resonance, the possible amplitude range and its impact; by calculating the abnormal vibration energy, the matching degree of the spectrum characteristic distribution and the natural frequency of the structure, the risk level of resonance caused by abnormal vibration is quantified, and the distribution of high-risk areas is evaluated in combination with parameters such as flow rate and pressure to identify the risk concentration points.

[0037] 4. The resonance prediction results are divided into multiple levels of warning, which can provide operators with clear and executable response strategies to prevent system vibration out of control or operation interruption in advance; by real-time monitoring of the system vibration frequency, combined with the relationship curve between the coolant flow rate and the resonance frequency, the flow rate or pump speed is automatically adjusted to dynamically avoid the resonance frequency, so that the system is away from the natural frequency range and the risk of vibration coupling is reduced.

[0038] 5. The flow rate adjustment process is based on closed-loop control and real-time detection of the adjustment effect. This feedback mechanism ensures the accuracy and controllability of flow rate adjustment, avoids unnecessary vibration caused by misadjustment, balances the relationship between cooling efficiency and safety, and reduces unnecessary power loss.

[0039] 6. By optimizing the distribution and spacing of support points, the natural frequency of the piping system is adjusted to keep it away from the dangerous frequency area caused by the dynamic coupling vibration of the coolant, thereby fundamentally reducing the possibility of resonance; using the particle swarm optimization (PSO) algorithm, based on the multivariable simulation results of the system, the optimal combination of support point stiffness and position is calculated to ensure that both structural stability is met and excessive cost or installation complexity is avoided. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 It is a schematic flow chart of the multi-level safety response control method of the industrial and commercial liquid cooling energy storage system of the present invention. DETAILED DESCRIPTION

[0041] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0042] See also Figure 1 The present invention provides a multi-level safety response control method for an industrial and commercial liquid cooling energy storage system, comprising:

[0043] Step 1: Deploy sensor networks at key nodes of the power energy storage system, and use the sensor networks to collect the time series matrix of operating status data , for the time series matrix Clean up abnormal data;

[0044] The step 1 includes the following contents:

[0045] Step 101: After deploying different types of sensors at key nodes of the liquid cooling system, a sensor network is formed, including: vibration sensors: deployed at pipe bends (high-risk areas for resonance) and pump outlets to capture dynamic changes in vibration frequency and amplitude, and obtain vibration frequency and amplitude. and amplitude ; Flow rate sensor, installed at the pump outlet and pipeline bifurcation, real-time monitoring of coolant flow rate ; Pressure sensor: deployed at the coolant inlet, radiator outlet and other pipeline nodes with significant pressure gradient to monitor the pressure distribution of the coolant ; Temperature sensors are deployed at the coolant inlet, outlet and radiator surface to capture changes in coolant temperature and ambient temperature ; Acoustic wave detector: used to detect microbubbles and cavitation in the pipeline. It is placed at the pump outlet and the lowest point of local pressure in the pipeline to obtain the characteristic frequency and signal energy of the acoustic signal;

[0046] Sampling frequency of all sensors Keep consistent, set Hz to meet the capture requirements of high-frequency dynamic processes. After data acquisition, the sampling data of all sensors are synchronized according to the timestamp to form a complete time series matrix :

[0047] Apply preset constraint rules to each data series to remove abnormal points that are beyond the physical reasonable range, for example: coolant flow rate: Should meet , vibration frequency Should meet the pipeline natural frequency range ,temperature Should meet the coolant working range wait;

[0048] By arranging multiple types of sensors at key locations such as bends, high-pressure areas, and pump outlets in the liquid cooling system, it is ensured that the multi-dimensional dynamic characteristics of the system operation are captured; vibration sensors are arranged in high resonance risk areas to effectively capture the dynamic changes in vibration frequency and amplitude, providing basic data for resonance analysis; using pressure sensors and flow rate sensors, the pressure gradient and flow state of the coolant in different pipe sections can be grasped in real time, and abnormal flow patterns (such as eddy currents or turbulence) can be quickly identified; the acoustic wave detector can promptly detect microbubbles or cavitation in the pipeline to avoid pipeline vibration and long-term structural damage caused by cavitation.

[0049] All sensors operate at a consistent sampling frequency to ensure the accuracy of capturing high-frequency dynamic processes, avoid data misalignment problems caused by asynchronous sampling, and provide protection for time series data analysis.

[0050] Step 102: Use Kalman filter algorithm to filter the time series matrix Perform real-time smoothing to eliminate random noise during data acquisition; Fluctuation of flow rate Perform joint detection and obtain correlation factors through correlation analysis , according to the correlation factor Determine whether there is an abnormal correlation between the two. (If set ), then the two are judged to be highly correlated and recorded as potential anomalies;

[0051] The Kalman filter algorithm can effectively remove random noise from sensor data, enhance the smoothness and credibility of data, and ensure the true reflection of dynamic characteristics; through the correlation factor analysis of vibration frequency and flow velocity fluctuation, it can accurately locate potential fluid-induced vibration or system coupling anomalies, and avoid missing multivariable coupling anomalies;

[0052] Eliminate data points that are beyond the physically reasonable range (such as outliers in flow rate and temperature) based on preset rules to avoid misleading effects of outliers on subsequent modeling and analysis;

[0053] Step 103: label and classify all verified data, including:

[0054] Normal data: meets physical rules and has no abnormal correlation; single variable anomaly: a physical quantity exceeds the reasonable range; multivariate correlation anomaly: highly correlated anomalies between different physical quantities. Normal data is used as input for subsequent spectrum analysis and risk assessment, and abnormal data is used for abnormal feature extraction;

[0055] When using, combine the contents in steps 101 to 103:

[0056] Subdividing data into normal data, single variable anomalies, and multivariate correlation anomalies can provide a more accurate input data set for subsequent spectrum analysis and risk prediction. By classifying and analyzing abnormal data and extracting abnormal feature patterns, it is helpful to deeply explore the potential causes of system failures; accurately distinguish single variable anomalies (such as a physical quantity exceeding the limit) and complex multivariate coupling anomalies, and improve the ability to locate faults and diagnose causes.

[0057] Step 2: Preprocess the vibration signal of the cooling system, extract the characteristic data of abnormal vibration through Fourier transform and energy distribution analysis, and analyze the abnormal area of ​​time-frequency distribution. To locate;

[0058] The step 2 includes the following contents:

[0059] Step 201: Process the vibration signal using a finite impulse response (FIR) bandpass filter to retain the vibration signal within the natural frequency range of the system and remove low-frequency drift and high-frequency noise; in order to avoid interference of sudden vibration signals on spectrum analysis, a detection method based on autoregressive residual is used to remove transient signals;

[0060] The use of bandpass filters to retain signals within the system's inherent frequency range can effectively remove low-frequency drift and high-frequency noise, focus on key vibration characteristics, eliminate transient signals and reduce the impact of abnormal sudden interference on spectrum analysis, improve the reliability of analysis results, provide purer signal input for subsequent Fourier transform and wavelet analysis, and enhance feature extraction accuracy.

[0061] Step 202: Perform Fourier transform on the normalized signal to obtain the amplitude distribution of the vibration signal in the frequency domain; divide the signal into multiple time windows, perform spectrum analysis on each time window and form a time-frequency diagram, perform wavelet transform on the vibration signal, and capture abnormal frequencies in the signal. and abnormal amplitude ;

[0062] The frequency domain characteristics of the signal extracted by Fourier transform can reveal the main frequency components and energy distribution of the vibration signal, locate the time when abnormal vibration occurs on the time axis, provide accurate time-frequency distribution information, capture abnormal frequency and abnormal amplitude, and help to further distinguish normal operating conditions from fault characteristics, thereby improving the efficiency of abnormal pattern recognition.

[0063] Step 203: extract abnormal features and identify abnormal patterns based on statistical analysis of the spectrum baseline and time-frequency distribution of normal working conditions, wherein the spectrum distribution baseline is established using the normal working condition data. :

[0064] ;

[0065] If the spectrum intensity Exceeding the reference spectrum value ,in, , then mark the frequency point as an abnormal frequency point , detect whether there are significantly abnormal frequency components in the spectrum on the time-frequency diagram, and calculate the local energy density to identify the abnormal areas of the time-frequency distribution To locate:

[0066] ;

[0067] in, For time and frequency The power spectral density of

[0068] If the signal at time and frequency of Amplitude higher than the reference spectrum value , it is determined to be an abnormal area of ​​time-frequency distribution ;

[0069] When using, combine the contents in steps 201 to 203:

[0070] Establishing a spectrum baseline based on normal working conditions, quickly locating abnormal components in the signal through the detection of abnormal frequency points, calculating local energy density and locating abnormal areas of time-frequency distribution, and clarifying the spatial distribution range of abnormal signals can provide a basis for further fault cause analysis. By extracting abnormal patterns, the diagnostic capability of complex vibration problems can be significantly improved.

[0071] Step 3: Generate a time series multivariate data set from the characteristic data of abnormal vibration , with the fused time series multivariate dataset Determine the cause label and combine simulation analysis to obtain the natural frequency of the cooling system and resonant frequency , match and determine the cause of abnormal vibration;

[0072] The step three includes the following contents:

[0073] Step 301: Import characteristic data of abnormal vibration, including abnormal frequency , abnormal amplitude , abnormal time-frequency distribution area , combined with the collected and verified coolant flow rate , pressure gradient and temperature changes , synchronize and standardize the above data; fuse them to generate a unified time series multivariate data set :

[0074] ;

[0075] Analyze whether the abnormal vibration is caused by one of the following causes: fluid-induced vibration : Vibration caused by uneven flow rate or pressure fluctuation; external environmental disturbance :Influence from external mechanical vibration or unstable foundation, structural looseness or fatigue : Natural frequency shift caused by loose pipe support or fatigue;

[0076] The abnormal vibration characteristic data is synchronously integrated with the flow velocity, pressure gradient, temperature change and other data to generate a unified multivariate data set, which can improve data consistency. It supports the analysis of the causes of vibration anomalies from multiple dimensions, such as fluid-induced vibration, external environmental disturbance, structural looseness, etc., and locates the specific source of the anomaly.

[0077] Step 302: Using the fused time series multivariate data set As input, the output is the cause label ,in,

[0078] ;

[0079] in is the posterior probability of each cause; is the optimal prediction value;

[0080] Based on the posterior probability model, the most likely cause label is assigned to the vibration anomaly, the fault cause is quickly identified, the diagnosis efficiency is improved, and the prediction results are used to guide the subsequent optimization of the cooling system, reducing the time for troubleshooting.

[0081] Step 303: Simulate the coolant flow characteristics of the pipeline system based on fluid mechanics simulation software and calculate the natural frequency of the cooling system under normal working conditions. and resonant frequency :

[0082] The simulation process includes: geometric modeling: building a geometric model of the pipeline of the liquid cooling system, refining the structure of the bends and support points; fluid property setting: setting the density of the coolant and viscosity , and flow rate Boundary conditions; Natural frequency calculation: Use modal analysis to solve the natural frequency of the structure ; Resonance frequency calculation: Calculate different flow rates through frequency response analysis and pressure gradient Resonance frequency under conditions ;

[0083] Simulation output natural frequency and the resonance frequency under different fluid conditions Distribution, the abnormal frequency The natural frequency obtained from simulation and resonant frequency To match:

[0084] like , then the abnormal vibration may be caused by the natural frequency of the structure; if , then the abnormal vibration may be caused by fluid-induced resonance;

[0085] When using, combine the contents in steps 301 to 303:

[0086] The simulation outputs the distribution of natural frequency and resonant frequency, which can quantify the range of vibration risk. In order to optimize the pipeline structure support and match the abnormal frequency with the simulation results, it can be determined whether the abnormality is caused by the structural natural frequency or the fluid-induced resonance, so as to implement targeted improvements.

[0087] Step 4: Use the LSTW neural network to predict the risk of abnormal resonance, and take corresponding countermeasures for the cooling network based on the prediction results, including one or more of dynamic adjustment of the coolant flow rate, adjustment of the arrangement of the support points, automatic reduction of the pump speed, and suspension of the system operation and switching to the backup cooling system;

[0088] The step 4 includes the following contents:

[0089] Step 401: Use the LSTW neural network to perform multivariate modeling on the historical vibration frequency and amplitude data and the current working condition data to obtain an abnormal resonance risk prediction model: use the multi-source data fusion analysis results as input, including the coolant flow rate , pressure gradient and temperature changes , vibration frequency , pressure gradient Sliding window sequence with equal characteristics ; Predict the probability of obtaining resonance risk , Risk impact range ,like , send a first-level alarm command to the outside. If , send a secondary alarm command to the outside;

[0090] The LSTW neural network is used to model multivariate historical data, which can achieve high-precision prediction of resonance risk. The risk classification alarm mechanism (level 1 and level 2 alarms) provides decision support for timely response of the cooling system to avoid the expansion of accidents.

[0091] Based on historical multivariate data and real-time data, the resonance risk of the liquid cooling system is dynamically modeled through the long short-term memory network LSTM, and the time series analysis algorithm is used to capture the precursors of resonance, accurately predict the time range of resonance, the possible amplitude range and its impact; by calculating the abnormal vibration energy, the matching degree of the spectrum characteristic distribution and the natural frequency of the structure, the risk level of resonance caused by abnormal vibration is quantified, and the distribution of high-risk areas is evaluated in combination with parameters such as flow rate and pressure to identify the risk concentration points;

[0092] Adjust parameters according to the actual operating status to adapt to different flow, pressure and external disturbance scenarios, improve the applicability and accuracy of the prediction, and avoid misjudgment due to changes in boundary conditions; divide the resonance prediction results into a multi-level early warning system, which can provide system operators with clear and executable response strategies to prevent system vibration out of control or operation interruption in advance.

[0093] Step 402: When a first-level alarm instruction is received, the coolant flow rate is dynamically adjusted, and the flow rate is adjusted to avoid the dangerous resonance frequency range based on the relationship between the resonance frequency and the flow rate. , combined with the current vibration frequency and flow rate , calculate the target flow rate to avoid resonance :Based on fluid mechanics theory, the natural frequency of coolant flow is Approximately satisfies the following relationship:

[0094] ;

[0095] in: is the coolant flow rate, is the characteristic length of the pipeline, is the bulk elastic modulus of the coolant, is the coolant density; the parameters are calibrated using the fluid mechanics simulation results to establish the resonant frequency-flow velocity curve of the actual system ;

[0096] ;

[0097] in is the target frequency offset (calibrated by experiment), is the system's resonant frequency-flow velocity mapping relationship;

[0098] Dynamically adjust the pump output flow rate and gradually adjust the flow rate to the target value , monitor the changes in vibration amplitude and frequency through real-time feedback to verify the adjustment effect;

[0099] Dynamically adjusting the flow rate based on the relationship between the resonant frequency and the flow rate can effectively avoid dangerous resonant frequency ranges and reduce resonance risks. The real-time feedback mechanism monitors the adjustment effect to ensure the actual effectiveness of the adjustment measures.

[0100] By real-time monitoring of the system vibration frequency and combining the relationship curve between the coolant flow rate and the resonant frequency, the flow rate or pump speed is automatically adjusted to dynamically avoid the resonant frequency, so that the system is away from the natural frequency range and the risk of vibration coupling is reduced.

[0101] The flow rate adjustment process is based on closed-loop control, and the adjustment effect (such as vibration intensity, frequency offset, etc.) is detected in real time. This feedback mechanism ensures the accuracy and controllability of flow rate adjustment, avoids unnecessary vibration caused by misadjustment, balances the relationship between cooling efficiency and safety, and reduces unnecessary power loss. It can be adjusted according to different flow requirements and external disturbance environments (such as high temperature and pressure fluctuations) to ensure that the solution is flexible and adaptable to a variety of work scenarios.

[0102] Step 403: If the coolant flow rate adjustment effect is insufficient, adjust the arrangement of the support points, increase the number of support points in the high-risk area or adjust the support point spacing to ,

[0103] Since the arrangement of pipeline support points will directly affect its natural frequency and vibration modes, and establish a support point optimization model by combining fluid mechanics simulation and vibration analysis to calculate the ideal support spacing and stiffness :

[0104] ;

[0105] in, is the support point stiffness, is the pipe mass per unit length, is the support point spacing;

[0106] Determine the optimal parameter combination through particle swarm optimization algorithm , so that the natural frequency and vibration frequency Avoid dangerous resonance ranges;

[0107] By optimizing the distribution and spacing of support points, the natural frequency of the piping system is adjusted to keep it away from the dangerous frequency area caused by the dynamic coupling vibration of the coolant, thereby fundamentally reducing the possibility of resonance. The particle swarm optimization (PSO) algorithm is used to calculate the optimal combination of support point stiffness and position based on the multivariable simulation results of the system. This ensures that both structural stability is met and excessive cost or installation complexity is avoided.

[0108] When the cooling system is running for a long time and the support points may become fatigued or their stiffness may change, the support point layout can be re-evaluated in combination with real-time vibration data, and the structural stability of the system can be maintained by adjusting or adding support points. The optimized support points form a damping effect in the vibration propagation path, reducing the transmission of high-frequency vibration energy, and improving the anti-resonance ability of the entire cooling system.

[0109] Step 404: When receiving the secondary alarm command, perform the following operations: Replace high-strength materials (such as carbon fiber composite materials) to increase the support stiffness to , add dampers (such as rubber vibration isolation pads or hydraulic damping devices) in high-risk areas to absorb vibration energy, automatically reduce pump speed and suspend system operation, switch to the backup cooling system, and ensure the continuity of equipment operation.

[0110] By replacing traditional piping materials with high-strength materials (such as high-modulus composite materials), the system's anti-resonance ability can be significantly improved. At the same time, dampers or vibration absorbers are added to absorb or attenuate vibration energy in high-vibration areas to reduce the impact on other parts of the system; in scenarios where the resonance risk level is high or long-term continuous operation is required, backup cooling paths or backup equipment are designed. When an abnormality occurs, it automatically switches to the backup system to ensure continuous operation of the system and avoid operation interruptions or shutdowns caused by vibration abnormalities. Based on historical data analysis after multiple vibration abnormalities, the overall design of the cooling system (such as flow channel layout, pipeline material selection, support point location distribution, etc.) is further optimized to improve the long-term stability of the system.

[0111] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0112] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0113] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the units is only some logical function divisions. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0114] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0115] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A multi-level safety response control method for an industrial and commercial liquid cooling energy storage system, characterized in that: include, A sensor network is deployed at the key nodes of the power storage system, and the sensor network collects the time series matrix of the operating status data. , for the time series matrix Clean up abnormal data; Preprocess the vibration signal of the cooling system, extract the characteristic data of abnormal vibration through Fourier transform and energy distribution analysis, and analyze the abnormal area of ​​time-frequency distribution. To locate; Generate a time series multivariate dataset from the characteristic data of abnormal vibration , with the fused time series multivariate dataset Determine the cause label and combine simulation analysis to obtain the natural frequency of the cooling system and resonant frequency , match and determine the cause of abnormal vibration; Use the LSTW neural network to predict the risk of abnormal resonance, and take corresponding countermeasures for the cooling network based on the prediction results, including one or more of dynamic adjustment of coolant flow rate, adjustment of the arrangement of support points, automatic reduction of pump speed, and suspension of system operation and switching to a backup cooling system; Import characteristic data of abnormal vibration, including abnormal frequency , abnormal amplitude , abnormal time-frequency distribution area , combined with the collected and verified coolant flow rate , pressure gradient and temperature changes , fused to generate a unified time series multivariate dataset : ; Using the fused time series multivariate dataset As input, the output is the cause label ,in, ； in is the posterior probability of each cause; is the best prediction value, fluid induced vibration : Vibration caused by uneven flow rate or pressure fluctuation; external environmental disturbance :Influence from external mechanical vibration or unstable foundation, structural looseness or fatigue : Natural frequency shift caused by loose pipe support or fatigue.

2. The multi-level safety response control method for a liquid-cooled energy storage system according to claim 1, characterized in that: After deploying different types of sensors at the key nodes of the liquid cooling system, a sensor network is formed. The sampling frequency of all sensors is kept consistent, and the sampling data of all sensors are synchronized according to the timestamp to form a complete time series matrix. , apply preset constraint rules to each data sequence and remove abnormal points that exceed the physical reasonable range; For the time series matrix Perform real-time smoothing to eliminate random noise during data acquisition; The vibration frequency Fluctuation of flow rate Conduct joint testing to determine whether there is any abnormal correlation between the two; All verified data are labeled and classified, including normal data, single variable anomalies, and multivariate correlation anomalies.

3. The multi-level safety response control method for a liquid-cooled energy storage system according to claim 2, characterized in that: Process the vibration signal, retain the vibration signal within the natural frequency range of the system, remove low-frequency drift and high-frequency noise, and use the detection method based on autoregressive residual to eliminate transient signals; Focus on key vibration characteristics, eliminate transient signals, reduce the impact of abnormal sudden interference on spectrum analysis, and improve the reliability of analysis results.

4. The multi-level safety response control method for a liquid-cooled energy storage system according to claim 3, characterized in that: After normalizing the vibration signal, perform Fourier transform to obtain the amplitude distribution of the vibration signal in the frequency domain; divide the signal into multiple time windows, perform spectrum analysis on each time window and form a time-frequency diagram, perform wavelet transform on the vibration signal, and capture abnormal frequencies in the signal and abnormal amplitude ; Extract abnormal features and identify abnormal patterns based on statistical analysis of the spectrum baseline and time-frequency distribution of normal operating conditions; If the spectrum intensity of the vibration signal exceeds the reference spectrum value, the corresponding frequency point will be marked as an abnormal frequency point. By calculating the local energy density, the time-frequency distribution of abnormal areas to locate.

5. The multi-level safety response control method for a liquid-cooled energy storage system according to claim 4, characterized in that: Based on the fluid mechanics simulation software, the coolant flow characteristics of the pipeline system are simulated and the natural frequency is output by simulation. and the resonance frequency under different fluid conditions Distribution, the abnormal frequency The natural frequency obtained from simulation and resonant frequency Matching, the cause of abnormal vibration.

6. The multi-level safety response control method for a liquid-cooled energy storage system according to claim 5, characterized in that: The multi-source data fusion analysis results are used as input, including the coolant flow rate , pressure gradient and temperature changes , vibration frequency , pressure gradient , predict the probability of obtaining resonance risk Scope of risk impact ,like , send a first-level alarm command to the outside. If , send a secondary alarm command to the outside.

7. The multi-level safety response control method for a liquid-cooled energy storage system according to claim 6, characterized in that: When a level 1 alarm command is received, the coolant flow rate is dynamically adjusted. Combined with the relationship between the resonance frequency and the flow rate, the flow rate is adjusted to avoid the dangerous resonance frequency range. The output flow of the pump is dynamically adjusted to gradually adjust the flow rate to the target value. The changes in vibration amplitude and frequency are monitored through real-time feedback to verify the adjustment effect.

8. The multi-level safety response control method for a liquid-cooled energy storage system according to claim 7, characterized in that: If the coolant flow rate adjustment effect is insufficient, adjust the layout of the support points, increase the number of support points in the high-risk area or adjust the support point spacing to , because the arrangement of pipeline support points will directly affect its natural frequency and vibration modes, and establish a support point optimization model by combining fluid mechanics simulation and vibration analysis to calculate the ideal support spacing and stiffness : ； in, is the support point stiffness, is the pipe mass per unit length, is the support point spacing; Determine the optimal parameter combination through particle swarm optimization algorithm , so that the natural frequency and vibration frequency Avoid dangerous resonance ranges.

9. The multi-level safety response control method for a liquid-cooled energy storage system according to claim 8, characterized in that: When receiving the second-level alarm command, perform the following operations: Replace high-strength materials to increase the support stiffness to , add dampers in high-risk areas to absorb vibration energy, automatically reduce pump speed and suspend system operation, switch to the backup cooling system to ensure the continuity of equipment operation.

Images