A fault early warning method and system fusing SAW temperature and UHF partial discharge

Abstract

The application discloses a kind of fusion SAW temperature and UHF partial discharge fault early warning method and system, belong to power equipment state monitoring technical field, it includes obtaining power equipment acoustic surface wave temperature signal and ultra-high frequency partial discharge signal, generates original time series data, generates temperature change rate parameter and partial discharge activity intensity parameter respectively, and calculates state correlation coefficient;State correlation coefficient is compared with linkage threshold value, and generates hierarchical fusion early warning signal;Analysis hierarchical fusion early warning signal, generates and outputs equipment fault early warning information containing fault nature inference.The way that temperature and partial discharge signal are synchronously collected, the statistical correlation of the time series of two is calculated and analyzed, and linkage threshold value is constructed to generate hierarchical fusion early warning signal, can accurately identify thermal-electric coupling type fault, and infer the nature of fault, improve the reliability and intelligent level of early warning.

Description

Technical Field

[0001] This invention relates to the field of power equipment condition monitoring technology, and in particular to a fault early warning method and system that integrates SAW temperature and UHF partial discharge. Background Technology

[0002] Electrical equipment is the cornerstone of the safe and stable operation of a power system, and the health of its internal insulation and conductive connections directly affects the reliability of the power grid. During operation, equipment may experience localized overheating due to overload, poor contact, or material aging. Simultaneously, the insulating medium gradually deteriorates under the combined effects of electric fields, thermal fields, and mechanical stress, leading to partial discharge. Partial discharge is an early sign of insulation defects, and the ultra-high frequency electromagnetic wave signals it generates can be effectively monitored. Therefore, online monitoring of electrical equipment, especially real-time monitoring of the two key parameters of temperature and partial discharge, is of paramount importance for preventing sudden failures and ensuring power supply.

[0003] Currently, online monitoring technology for power equipment is widely used. In temperature monitoring, common methods include infrared thermometry and fiber optic thermometry; in partial discharge monitoring, ultra-high frequency (UHF) methods have become the mainstream technology due to their strong anti-interference capabilities and high sensitivity. In practical applications, maintenance personnel typically deploy independent temperature monitoring systems and partial discharge monitoring systems to monitor and trigger alarms for these two physical quantities respectively. When either monitored value exceeds a preset fixed threshold, the system will issue an alarm signal, prompting maintenance personnel to conduct repairs.

[0004] However, using separate monitoring systems makes it difficult to ensure precise temporal synchronization between temperature data and partial discharge data, creating challenges in data alignment for in-depth analysis of the correlation between the two. On the other hand, alarm methods based on fixed thresholds for a single physical quantity are rather crude, failing to distinguish between parameter fluctuations caused by normal operating conditions such as load and environment changes and abnormal changes caused by actual internal defects in the equipment, easily leading to frequent false alarms or missed alarms. Summary of the Invention

[0005] To address the aforementioned issues, this invention provides a fault early warning method and system that integrates SAW temperature and UHF partial discharge. By synchronously acquiring temperature and partial discharge signals, calculating and analyzing the statistical correlation between their time series, and constructing a linkage threshold to generate a graded fused early warning signal, this method can accurately identify thermoelectric coupling faults and infer the nature of the fault, thereby improving the reliability and intelligence level of the early warning.

[0006] The above objectives can be achieved through the following approach:

[0007] A fault early warning method integrating SAW temperature and UHF partial discharge includes acquiring surface acoustic wave temperature signal and UHF partial discharge signal of power equipment synchronously collected by a preset integrated sensor, generating raw time series data, and generating a temperature change rate parameter characterizing the abnormal thermal trend of the equipment and a partial discharge activity intensity parameter quantifying the insulation activity intensity based on the raw time series data.

[0008] Calculate the statistical correlation between the time series of the temperature change rate parameter and the time series of the partial discharge activity intensity parameter to generate a state correlation coefficient; construct a linkage threshold to characterize the physical coupling relationship between thermal state deterioration and insulation degradation, and compare the state correlation coefficient with the linkage threshold to generate a graded fusion early warning signal; parse the graded fusion early warning signal to generate and output equipment fault early warning information containing fault nature inference.

[0009] Optionally, generating the original time series data includes: transmitting a radio frequency interrogation signal to a preset integrated sensor to excite the surface acoustic wave element in the integrated sensor; receiving a reflected signal carrying temperature information returned by the integrated sensor, and simultaneously capturing a UHF partial discharge signal received by the integrated sensor to obtain a mixed signal; demodulating and digitizing the mixed signal, and marking it with a consistent timestamp to form the original time series data.

[0010] Optionally, generating the temperature change rate parameter characterizing the abnormal thermal trend of the equipment includes: establishing a dynamic temperature baseline that filters out environmental and load fluctuations based on the surface acoustic wave temperature signal in the original time series data; performing a differential calculation between the surface acoustic wave temperature signal and the dynamic temperature baseline to obtain a temperature difference value; and calculating and generating the temperature change rate parameter according to the changing trend of the temperature difference value over time.

[0011] Optionally, generating partial discharge activity intensity parameters for quantifying insulation activity intensity includes: performing pulse detection on the ultra-high frequency partial discharge signals in the original time-series data within a preset time window, and statistically obtaining the pulse repetition frequency; measuring the amplitude of all discharge pulses within the time window and sorting them from largest to smallest, selecting the amplitude of the discharge pulse ranked first to obtain the discharge amplitude; and weightedly fusing the pulse repetition frequency and the discharge amplitude to generate partial discharge activity intensity parameters.

[0012] Optionally, the step of calculating the statistical correlation between the time series of the temperature change rate parameter and the time series of the partial discharge activity intensity parameter, and generating the state correlation coefficient, includes: constructing a pair of time series samples of the temperature change rate parameter and the partial discharge activity intensity parameter within the same observation period; calculating the linear correlation strength of the time series sample pairs to obtain a correlation strength value; and calculating the state correlation coefficient based on the correlation strength value.

[0013] Optionally, generating the graded fusion early warning signal includes: constructing a linkage threshold to characterize the physical coupling relationship between thermal state deterioration and insulation degradation; determining whether the absolute value of the state correlation coefficient is greater than the linkage threshold; if so, generating the graded fusion early warning signal based on the state correlation coefficient and in combination with the temperature change rate parameter and the partial discharge activity intensity parameter.

[0014] Optionally, generating the graded fusion early warning signal based on the state correlation coefficient and in combination with the temperature change rate parameter and the partial discharge activity intensity parameter includes: calculating the change in the temperature change rate parameter and the change in the partial discharge activity intensity parameter, and fusing them to obtain the monitoring data change value; calculating a comprehensive risk index based on the state correlation coefficient and the monitoring data change value; mapping the comprehensive risk index to a preset risk level table to determine the early warning level; and constructing and outputting the graded fusion early warning signal based on the early warning level.

[0015] Optionally, the step of parsing the hierarchical fusion early warning signal to generate and output equipment fault early warning information containing fault nature inference includes: parsing the risk level from the hierarchical fusion early warning signal; and generating equipment fault early warning information containing fault nature inference and correlation analysis results based on the risk level and the state correlation coefficient.

[0016] Optionally, the method further includes: acquiring the frequency of electromagnetic wave energy generated by partial discharge; determining whether the electromagnetic wave energy frequency is within a preset target frequency band; if so, improving the detection sensitivity of the ultra-high frequency partial discharge signal.

[0017] Based on the same inventive concept, this invention also provides a fault early warning system integrating SAW temperature and UHF partial discharge. The system includes: a signal acquisition module for acquiring surface acoustic wave temperature signals and UHF partial discharge signals of power equipment synchronously acquired by a preset integrated sensor, generating raw time-series data; a parameter processing module for generating a temperature change rate parameter characterizing the abnormal thermal trend of the equipment and a partial discharge activity intensity parameter quantifying the insulation activity intensity based on the raw time-series data; a correlation analysis module for calculating the statistical correlation between the time series of the temperature change rate parameter and the time series of the partial discharge activity intensity parameter, generating a state correlation coefficient; an early warning generation module for constructing a linkage threshold characterizing the physical coupling relationship between thermal state deterioration and insulation degradation, comparing the state correlation coefficient with the linkage threshold, and generating a graded fused early warning signal; and an information output module for parsing the graded fused early warning signal, generating and outputting equipment fault early warning information containing fault nature inference.

[0018] Compared with the prior art, the present invention has the following advantages:

[0019] 1. The invention achieves synchronous acquisition of temperature and partial discharge signals through an integrated sensor, ensuring a strict correspondence between the two physical quantities in the time dimension from the data source, eliminating the timestamp deviation and data alignment problems caused by heterogeneous sensor systems; this highly faithful raw data provides a solid foundation for subsequent in-depth fusion analysis, making the mining of thermoelectric coupling fault characteristics more accurate and reliable, thereby improving the accuracy of fault early warning analysis.

[0020] 2. This invention elevates fault diagnosis from a single-parameter independent threshold judgment mode to a quantitative analysis level of the coupling relationship of multiple physical quantities. It can distinguish whether the synchronous changes in monitoring data are due to accidental factors or have an inherent physical causal relationship, filter out false alarms caused by external interference, and ensure that the warning is triggered only when the thermal state deterioration and insulation degradation show clear co-evolution characteristics, thereby improving the signal-to-noise ratio and reliability of the warning.

[0021] 3. This invention dynamically calculates a comprehensive risk index by integrating information from multiple dimensions such as correlation strength and parameter change trends. This method can not only determine whether a fault exists, but also assess its severity and development speed. The resulting graded early warning signals provide richer and more quantitative basis for operation and maintenance decisions, realizing an upgrade from simple alarms to risk classification.

[0022] 4. By analyzing the sign and magnitude of the state correlation coefficient, this invention can preliminarily determine the fault mode. This allows the early warning information to go beyond data presentation and be transformed into decision support information, improving the efficiency and pertinence of fault diagnosis and providing technical support for predictive maintenance and equipment life cycle health management.

[0023] Other features and advantages of the invention will be set forth in the description which follows, and will be apparent in part from the description, or may be learned by practicing the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures pointed out in the description, claims and drawings. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0025] Figure 1 This is a flowchart illustrating a fault early warning method that integrates SAW temperature and UHF partial discharge according to an embodiment of the present invention.

[0026] Figure 2 This is a schematic diagram comparing the SAW temperature signal with the dynamic temperature baseline in an embodiment of the present invention.

[0027] Figure 3 This is a schematic diagram of a fault early warning system that integrates SAW temperature and UHF partial discharge according to an embodiment of the present invention. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0029] Reference Figure 1 One embodiment of the present invention proposes a fault early warning method that integrates SAW temperature and UHF partial discharge. By synchronously collecting temperature and partial discharge signals, calculating and analyzing the statistical correlation between their time series, and constructing a linkage threshold to generate a hierarchical fusion early warning signal, the method can accurately identify thermoelectric coupling faults and infer the nature of the fault, thereby improving the reliability and intelligence level of the early warning.

[0030] The method described in this embodiment specifically includes:

[0031] Acquire surface acoustic wave temperature signals and ultra-high frequency partial discharge signals of power equipment synchronously collected by a preset integrated sensor, and generate raw time-series data;

[0032] Based on the original time-series data, a temperature change rate parameter characterizing the abnormal thermal trend of the equipment and a partial discharge activity intensity parameter quantifying the insulation activity intensity are generated respectively.

[0033] Calculate the statistical correlation between the time series of the temperature change rate parameter and the time series of the partial discharge activity intensity parameter to generate a state correlation coefficient;

[0034] A linkage threshold is constructed to characterize the physical coupling relationship between thermal state deterioration and insulation degradation, and the state correlation coefficient is compared with the linkage threshold to generate a hierarchical fusion early warning signal;

[0035] The hierarchical fusion early warning signal is analyzed to generate and output equipment fault early warning information containing fault nature inference.

[0036] This invention utilizes an integrated sensor to ensure completely synchronized acquisition of two key state information parameters—temperature and partial discharge—on a time base, guaranteeing the effectiveness of subsequent correlation analysis from the data source. These parameters are then transformed into characteristic parameters that characterize dynamic deterioration trends: the temperature change rate parameter and the partial discharge activity intensity parameter. Subsequently, statistical methods are used to calculate the correlation between the time series of these two dynamic parameters, generating a state correlation coefficient that quantitatively characterizes the synchronicity of the evolution of thermal and insulation states. Finally, by comparing this correlation coefficient with a linkage threshold reflecting the strength of physical coupling, a hierarchical early warning mechanism integrating correlation strength and parameter amplitude is triggered. Based on this, the fundamental nature of the fault is inferred, achieving a complete closed loop from data acquisition, feature extraction, correlation quantification to intelligent diagnosis.

[0037] This invention, by introducing state correlation analysis, effectively filters out parameter fluctuations caused by external factors such as environment and load that are unrelated to inherent equipment defects, reducing false alarms and missed alarms common in traditional single-parameter early warning methods. It can accurately identify concurrent thermal and electrical degradation processes driven by the same fault source, thus pinpointing the root cause of the fault in advance. Furthermore, the generated tiered early warning signals not only inform of the existence of risk but also reveal the severity and urgency of the fault through tier classification. The final output, including early warning information inferred from the nature of the fault, transforms complex monitoring data into intuitive and clear diagnostic conclusions, providing decision support for maintenance personnel and enabling them to conduct more targeted fault investigation and maintenance, thereby achieving early and accurate early warning and diagnosis of potential equipment faults.

[0038] Optionally, generating the raw time-series data includes:

[0039] A radio frequency interrogation signal is transmitted to a preset integrated sensor to excite the surface acoustic wave element inside the integrated sensor;

[0040] The system receives the reflected signal carrying temperature information returned by the integrated sensor and simultaneously captures the ultra-high frequency partial discharge signal received by the integrated sensor to obtain a mixed signal.

[0041] The mixed signal is demodulated and digitized to separate it, and marked with a consistent timestamp to form the original time-series data.

[0042] Specifically, the signal acquisition process begins by actively transmitting a specific frequency and coded radio frequency (RF) interrogation signal to an integrated sensor installed at the location to be tested on the power equipment. This integrated sensor is a composite passive wireless sensor that integrates a surface acoustic wave (SAW) temperature sensing element and an ultra-high frequency (UHF) partial discharge detection antenna. The RF interrogation signal excites the SAW element, converting its electromagnetic energy into mechanical waves on the piezoelectric substrate surface, i.e., SAW. Next, synchronous signal reception and acquisition are performed. As the SAW propagates on the piezoelectric substrate, its propagation speed changes due to temperature. After passing through a specially designed reflection structure, the mechanical wave is converted back into an electromagnetic wave, forming a reflected signal carrying temperature information, which is then transmitted back to the receiving end by the integrated sensor. Simultaneously, the UHF antenna of the integrated sensor continuously and passively captures UHF partial discharge signals radiated from potential partial discharges in the insulating medium of the power equipment. The receiving device synchronously receives the reflected signal carrying temperature information and the UHF partial discharge signal. Since both are captured simultaneously in the same or adjacent channels, they constitute an unprocessed mixed signal. Finally, the received mixed signal is processed to generate raw time-series data. The mixed signal is first fed into a demodulation unit, which separates the surface acoustic wave (SAW) reflection signal from its carrier wave. Subsequently, the entire signal stream undergoes high-speed analog-to-digital conversion and enters the digital processing stage. In the digital domain, specific signal recognition and separation algorithms completely separate the SAW pulse signal representing temperature from the UHF partial discharge pulse signal representing insulation state. For example, the temperature value is calculated by determining the time delay between the SAW reflection pulses, while the UHF partial discharge signal is preserved in its pulse waveform form. A crucial step is to assign a perfectly consistent timestamp to each set of successfully separated and calculated temperature data and corresponding UHF partial discharge data. This timestamp precisely records the instant the data was acquired. By continuously executing these steps, a time series set consisting of a series of temperature values ​​and partial discharge signals with precise synchronization timestamps is generated; this set constitutes the original time-series data.

[0043] Optionally, the temperature change rate parameter used to generate the characterization of abnormal thermal trends in the device includes:

[0044] Based on the surface acoustic wave temperature signal in the original time series data, a dynamic temperature baseline is established to filter out environmental and load fluctuations.

[0045] The surface acoustic wave temperature signal and the dynamic temperature baseline are differentially calculated to obtain the temperature difference value;

[0046] Based on the trend of the temperature difference value changing over time, the temperature change rate parameter is calculated and generated.

[0047] Specifically, such as Figure 2 As shown, a dynamic temperature baseline is first established to filter out environmental and load fluctuations. Unlike traditional methods that use fixed thresholds, this method aims to establish a reference value that reflects the expected temperature of the equipment under normal operating conditions in the current environment and load. To this end, it is necessary to synchronously collect load data such as current and ambient temperature data during equipment operation, and then correlate these data with the surface acoustic wave temperature signal obtained from long-term operation of the equipment in a healthy state. Multiple regression analysis or machine learning algorithms can be used to construct a mathematical model that can predict the normal operating temperature of the equipment based on real-time ambient temperature and load current. The output of this model is the dynamic temperature baseline. For example, the mathematical model can be expressed as:

[0048] ,

[0049] in, The dynamic temperature baseline at time t; Let t be the ambient temperature at time t; Let be the normalized value of the device load current at time t; , and The coefficients are model coefficients, obtained by fitting and training historical data during normal equipment operation.

[0050] Secondly, the real-time acquired surface acoustic wave temperature signal is differentially calculated with the dynamic temperature baseline to obtain the temperature difference value. This calculation aims to isolate normal fluctuations and highlight the net temperature rise caused by internal equipment anomalies. The calculation process is as follows:

[0051] ,

[0052] in, The temperature difference fraction at time t; This is the measured value of the surface acoustic wave temperature signal at time t, obtained from the original time-series data. This temperature difference value objectively reflects the abnormal temperature amount that exceeds normal expectations.

[0053] Finally, based on the trend of the temperature difference over time, a temperature change rate parameter is calculated. This parameter is used to quantify the development speed of abnormal thermal trends. This parameter is obtained by calculating the rate of change of the temperature difference within a preset time window, i.e., its time derivative. Specifically, it can be approximated as follows:

[0054] ,

[0055] in, The temperature change rate parameter at time t; and These are the temperature difference scores between the current time and the previous observation time, respectively. The length of the time window used to calculate the rate of change. The final temperature change rate parameter. This is a key indicator for characterizing abnormal thermal trends in equipment.

[0056] Optionally, the partial discharge activity intensity parameters used to generate the quantified insulation activity intensity include:

[0057] Within a preset time window, pulse detection is performed on the ultra-high frequency partial discharge signal in the original time series data, and the pulse repetition frequency is statistically obtained.

[0058] The amplitudes of all discharge pulses within the time window are measured and sorted from largest to smallest. The amplitude of the discharge pulse with the highest amplitude in the sorted order is selected to obtain the discharge amplitude.

[0059] The pulse repetition frequency and the discharge amplitude are weighted and fused to generate partial discharge activity intensity parameters.

[0060] Specifically, firstly, pulse detection is performed on the UHF partial discharge signal within a preset time window, and the pulse repetition frequency is statistically analyzed. A fixed time window length is set, such as one power frequency cycle or one second. Within this time window, the acquired UHF partial discharge signal sequence is processed. By setting an amplitude threshold higher than the background noise, all pulse signals exceeding this threshold are identified and counted. The number of pulses per unit time is the pulse repetition frequency.

[0061] Next, measure the amplitude of all discharge pulses within the time window and determine the maximum discharge amplitude. For each valid discharge pulse detected in the previous step, accurately measure its peak amplitude. Collect the amplitude data of all pulses within the time window and sort them from largest to smallest. Select the first value in the sorted list, i.e., the largest pulse amplitude, as the representative discharge amplitude of the time window.

[0062] Finally, the pulse repetition frequency and the discharge amplitude are weighted and fused to generate partial discharge activity intensity parameters. Since the pulse repetition frequency (in Hertz or times / second) and the discharge amplitude (in millivolts or decibels / milliwatts) are two physical quantities with different physical meanings and dimensions, they cannot be directly arithmetically calculated. Therefore, they need to be dimensionless first. Specifically, the measured pulse repetition frequency and discharge amplitude are normalized by dividing them by their respective reference values ​​or historical maximum values. Subsequently, based on expert experience or the different sensitivities of different fault types to these two parameters in specific application scenarios, corresponding weighting coefficients are set, and a weighted sum is performed to obtain the final partial discharge activity intensity parameters. This process can be expressed by the following formula:

[0063] ,

[0064] in, The final generated partial discharge activity intensity parameters; This refers to the pulse repetition frequency obtained statistically within a time window. This represents the maximum discharge amplitude measured within this window. and These are the normalized reference values ​​set for the pulse repetition frequency and the discharge amplitude, respectively. They can be obtained from the equipment's factory standard, the stable maximum value in historical operating data, or a preset alarm threshold. and These are the weighting coefficients corresponding to the pulse repetition frequency and the discharge amplitude, respectively, and their sum is 1. Their values ​​reflect the degree of emphasis placed on discharge frequency and discharge energy when assessing the insulation activity strength.

[0065] Optionally, the calculation of the statistical correlation between the time series of the temperature change rate parameter and the time series of the partial discharge activity intensity parameter, generating a state correlation coefficient, includes:

[0066] Construct time-series sample pairs of the temperature change rate parameter and the partial discharge activity intensity parameter within the same observation period;

[0067] Calculate the linear correlation strength of the time series sample pairs to obtain the correlation strength value;

[0068] The state correlation coefficient is calculated based on the aforementioned correlation strength value.

[0069] Specifically, a representative observation period is first selected, long enough to capture the dynamic characteristics of fault development, such as several hours or days. Within this observation period, the time series of the temperature change rate parameter calculated in the previous steps is aligned with the time series of the partial discharge activity intensity parameter. Since both originate from synchronously acquired raw time-series data, they naturally have consistent timestamps. Therefore, at any point within the observation period... Both can obtain a data pair The data pairs from all times within the observation period are then collected to form time series sample pairs for correlation analysis. Next, the linear correlation strength of these time series sample pairs is calculated to obtain the correlation strength value. Here, the Pearson correlation coefficient is used to quantify the linear dependence between the two time series. The Pearson correlation coefficient is a statistical indicator that measures the strength of the linear relationship between two variables. Its calculation formula is as follows:

[0070] ,

[0071] in, This refers to the calculated relevant intensity value; and These represent the values ​​of the temperature change rate parameter and the partial discharge activity intensity parameter at the i-th sampling point within the observation period, respectively. and These are the arithmetic mean of the time series of these two parameters over the entire observation period; Σ represents the summation over all sample points within the observation period.

[0072] Finally, the calculation obtained in the previous step It is directly used as the state correlation coefficient. This coefficient ranges from -1 to +1, and its magnitude and sign have clear physical meaning. A positive value indicates a positive correlation between the temperature change trend and the insulation activity strength change trend, while a negative value indicates a negative correlation. The absolute value indicates the strength of the correlation. This single value, the state correlation coefficient, highly condenses the information about the linkage between thermal state deterioration and insulation degradation over time.

[0073] Optionally, the generation of hierarchical fusion early warning signals includes:

[0074] Construct a linkage threshold to characterize the physical coupling relationship between thermal state deterioration and insulation degradation;

[0075] Determine whether the absolute value of the state correlation coefficient is greater than the linkage threshold;

[0076] If so, the graded fusion early warning signal is generated based on the state correlation coefficient, combined with the temperature change rate parameter and the partial discharge activity intensity parameter.

[0077] Specifically, the linkage threshold is a key criterion used to distinguish between accidental synchronous fluctuations between state parameters and strongly correlated evolution driven by a common fault source. This threshold is not a fixed empirical value, but rather based on the analysis of historical fault data for specific power equipment types, calculation results from multiphysics simulation models, or reasoning from expert system knowledge bases. For example, by analyzing numerous known fault cases where overheating leads to accelerated insulation aging and triggers partial discharge, the distribution of the state correlation coefficient between the temperature change rate parameter and the partial discharge activity intensity parameter during the fault development process is statistically analyzed. A critical value that can effectively distinguish between fault and normal states is selected as the linkage threshold for the equipment, denoted as... This threshold is a dimensionless pure number, ranging from 0 to 1. Next, the real-time calculated state correlation coefficient is compared with the linkage threshold. At the end of each observation cycle, the system updates the latest calculated state correlation coefficient. The absolute value, that is With the preset linkage threshold Compare and perform the judgment logic, i.e., make the judgment. Is it greater than .

[0078] Finally, if the above judgment result is yes, that is, the absolute value of the state correlation coefficient exceeds the linkage threshold, this indicates that there is a significant and strong correlation between the detected abnormal temperature trend and the intensity of partial discharge activity, which is highly likely to originate from the same physical fault chain. At this time, the system will trigger the early warning generation process and comprehensively utilize information from three dimensions—the state correlation coefficient itself, the current temperature change rate parameter, and the partial discharge activity intensity parameter—to jointly generate the hierarchical fusion early warning signal. Specifically, the fusion generation method involves feeding these three parameters as input into a preset hierarchical decision model. This model makes judgments based on a rule set or algorithm. For example, when the state correlation coefficient is significantly positive and both activity parameters show rapid growth, it is determined to be the highest risk level; when the correlation is strong but the parameter growth is slow, it is determined to be a medium level; other situations correspond to lower levels. The final output hierarchical fusion early warning signal is a data packet containing a specific early warning level code and the key parameter values ​​on which this early warning is based.

[0079] Optionally, generating the hierarchical fusion early warning signal based on the state correlation coefficient, combined with the temperature change rate parameter and the partial discharge activity intensity parameter, includes:

[0080] The changes in the temperature change rate parameter and the changes in the partial discharge activity intensity parameter are calculated and fused to obtain the monitoring data change value.

[0081] The comprehensive risk index is calculated based on the state correlation coefficient and the change value of the monitoring data;

[0082] The comprehensive risk index is mapped to a preset risk level table to determine the warning level;

[0083] Based on the aforementioned warning level, a hierarchical fusion warning signal is constructed and output.

[0084] Specifically, the changes in the temperature change rate parameter and the partial discharge activity intensity parameter are first calculated and then fused to obtain the monitoring data change value. This step aims to quantify the acceleration of fault development. First, the temperature change rate parameter within the current observation period is calculated. Parameters compared to the previous observation period The difference yields the change in the temperature change rate parameter. Similarly, calculate the partial discharge activity intensity parameters for the current cycle. Parameters from the previous cycle The difference yields the change in the partial discharge activity intensity parameter. ,in This represents the time interval between two consecutive observation periods. Since these two changes have different physical meanings and numerical ranges, they need to be normalized to make them dimensionless before being weighted and fused to generate the monitoring data change value. :

[0085] ,

[0086] in, and The weighting coefficients are preset, and the sum of the two is 1. Their values ​​reflect the degree of emphasis on thermal runaway trend and insulation degradation acceleration trend when assessing the failure development speed. and It serves as a normalized reference benchmark for each variable, which can be set through historical data statistics or expert experience.

[0087] Next, a comprehensive risk index is calculated based on the state correlation coefficient and the change value of the monitoring data. This index aims to integrate the certainty of the fault's existence, represented by the state correlation coefficient, with the severity of the fault's development, represented by the change value of the monitoring data. The comprehensive risk index is obtained by multiplying the absolute value of the state correlation coefficient by the change value of the monitoring data. :

[0088] ,

[0089] in, This is a comprehensive risk index; It is the absolute value of the state correlation coefficient, which reflects the strength of thermoelectric coupling. The absolute value is taken because, regardless of whether it is a positive or negative correlation, a high correlation strength means increased risk. It is the change value of the monitoring data obtained from the previous step.

[0090] Then, the comprehensive risk index is mapped to a preset risk level table to determine the warning level. This risk level table predefines the risk levels corresponding to different comprehensive risk index ranges, such as "Caution," "Warning," and "Critical," as shown in Table 1. By calculating the... By comparing the value with the threshold in the table, the specific warning level of the current device can be determined.

[0091] Table 1 Risk Level Table

[0092]

[0093] Finally, based on the determined warning level, a hierarchical fusion warning signal is constructed and output. This signal is a structured data packet, the core of which is the determined warning level code. It can also include key information on the warning criteria, such as the specific values ​​of the current comprehensive risk index, state correlation coefficient, temperature change rate parameter, and partial discharge activity intensity parameter, to facilitate detailed analysis by maintenance personnel.

[0094] Optionally, the step of parsing the hierarchical fusion early warning signal and generating and outputting equipment fault early warning information containing fault nature inference includes:

[0095] The risk level is extracted from the hierarchical fusion early warning signal;

[0096] Based on the risk level and the state correlation coefficient, equipment fault early warning information is generated, which includes fault nature inference and correlation analysis results.

[0097] Specifically, the risk level is first extracted from the graded fused early warning signal. After receiving the graded fused early warning signal data packet from the early warning generation module, the information output module decodes it. This data packet encapsulates the early warning level determined in the previous steps, such as "Caution," "Alarm," or "Critical" levels represented by codes. This step aims to directly obtain the severity assessment result of the current risk state of the equipment. Next, based on the extracted risk level and the synchronously extracted state correlation coefficient, the system calls a built-in fault diagnosis knowledge base or inference rule set to generate equipment fault early warning information containing fault nature inference and correlation analysis results. For example, the inference rule can be set as follows: if the state correlation coefficient is a significantly positive value, it indicates a strong positive correlation between the temperature change rate parameter and the partial discharge activity intensity parameter, physically pointing to "heat" as the main factor driving "electrical" degradation. Based on different risk levels, the system can generate different inferences. When the risk level is "Caution," it infers "early signs of insulation degradation caused by abnormal heating." When the risk level is "Critical," it infers "a serious overheating fault in the equipment has caused rapid and irreversible damage to the insulation structure, bringing it to the brink of breakdown." Conversely, if the state correlation coefficient is close to zero or negative, even with a high risk level, the system will infer that the fault nature is "thermal anomalies and electrical anomalies may be caused by independent fault sources, with no obvious physical coupling relationship between the two, requiring separate investigation," thus avoiding incorrect causal correlation judgments. Finally, the system integrates these inference results into a text description, which explicitly includes the risk level, the inference about the fundamental nature of the fault, and the quantitative results of the correlation analysis supporting the inference. For example, "Alarm level, inferred to be an overheating fault caused by poor contact accelerating insulation aging, based on the strong positive correlation between temperature and partial discharge activity, with a state correlation coefficient of 0.85." This information is the final output equipment fault warning information.

[0098] Optionally, the method further includes:

[0099] To obtain the frequency of electromagnetic wave energy generated by partial discharge;

[0100] Determine whether the frequency of the electromagnetic wave energy is within the preset target frequency band;

[0101] If so, the detection sensitivity of the ultra-high frequency partial discharge signal will be improved.

[0102] Specifically, firstly, while routinely acquiring UHF partial discharge signals, the system performs real-time spectrum analysis on the captured signals to obtain the electromagnetic wave energy frequencies generated by the partial discharge. Specifically, a Fast Fourier Transform (FFT) is performed on the time-domain waveform of the UHF partial discharge signal within a time window, converting it to a frequency-domain representation. By analyzing the obtained spectrum, one or more frequency points with the most concentrated energy are identified; these frequencies are the main electromagnetic wave energy frequencies of this partial discharge event. Secondly, the system compares the acquired electromagnetic wave energy frequencies with a preset target frequency band to determine whether they fall within that band. This preset target frequency band is determined based on prior knowledge and represents the typical spectral characteristics of partial discharge signals generated by specific types or particularly dangerous internal insulation defects such as metal particles or suspended potential bodies. This judgment process is a simple logical comparison, checking whether the measured peak energy frequency is between the upper and lower limits of the target frequency band. Finally, if the judgment result is yes, meaning the energy characteristic frequency of the partial discharge matches the characteristic spectrum of a high-risk fault, the system will automatically trigger a sensitivity enhancement mechanism. This mechanism improves the detection sensitivity of the ultra-high frequency partial discharge signal by adjusting the parameters of the signal acquisition hardware or software. Specific implementation methods may include dynamically lowering the amplitude detection threshold used for pulse identification, allowing weaker discharge pulses of the same type that might have been filtered out as noise to be effectively captured; or adjusting the gain of the variable gain amplifier at the signal acquisition front end to amplify the signal by a greater factor. This adjustment is temporary and aims to enhance monitoring of specific types of discharge activity currently occurring.

[0103] Based on the same inventive concept, such as Figure 3 As shown, the present invention also provides a fault early warning system integrating SAW temperature and UHF partial discharge, the system comprising:

[0104] The signal acquisition module is used to acquire surface acoustic wave temperature signals and ultra-high frequency partial discharge signals of power equipment synchronously collected by a preset integrated sensor, and generate raw time-series data.

[0105] The parameter processing module is used to generate, based on the original time-series data, a temperature change rate parameter characterizing the abnormal thermal trend of the equipment and a partial discharge activity intensity parameter quantifying the insulation activity intensity.

[0106] The correlation analysis module is used to calculate the statistical correlation between the time series of the temperature change rate parameter and the time series of the partial discharge activity intensity parameter, and generate a state correlation coefficient.

[0107] The early warning generation module is used to construct a linkage threshold to characterize the physical coupling relationship between thermal state deterioration and insulation degradation, and compare the state correlation coefficient with the linkage threshold to generate a hierarchical fusion early warning signal;

[0108] The information output module is used to parse the hierarchical fusion early warning signal, generate and output equipment fault early warning information containing fault nature inference.

[0109] To verify the feasibility of this invention in practice, it was applied to the online monitoring of a critical high-voltage switchgear in a substation. This switchgear is responsible for supplying power to important loads, and the health status of its internal busbar joints directly affects the safety and reliability of the power grid operation. Traditional monitoring methods typically operate temperature monitoring and partial discharge monitoring as two independent systems, making it difficult to detect the inherent correlation between them and easily missing progressive, coupled faults such as insulation degradation caused by overheating. This embodiment aims to achieve early and accurate warning of thermoelectric coupling faults inside the switchgear by applying this invention.

[0110] In this embodiment, an integrated surface acoustic wave (SAW) temperature and ultra-high frequency (UHF) partial discharge sensor is installed near the busbar connector inside the switchgear. The system transmits an radio frequency interrogation signal to the sensor through the signal acquisition module, simultaneously acquiring the SAW temperature signal and the UHF partial discharge signal, generating raw time-series data with precise synchronization timestamps. Subsequently, the various functional modules of the system process and analyze the acquired data.

[0111] To verify the beneficial effects of the present invention, continuous operating data of the switchgear during a certain period was recorded and analyzed. During this period, the bus joint became slightly loose due to long-term operation, gradually leading to poor contact and triggering a typical thermoelectric coupling fault evolution process.

[0112] In the early stages of the fault, the system monitored that the equipment load current fluctuated around 400A. The measured SAW temperature of the bus joint basically matched the dynamic temperature baseline established based on the load and ambient temperature, and the calculated temperature change rate parameter was close to 0. At the same time, the UHF partial discharge signal was weak, and the partial discharge activity intensity parameter remained below the background level of 0.1.

[0113] As the joint loosened further, its contact resistance increased, leading to abnormal heating. The system detected that the measured temperature began to deviate continuously from the dynamic temperature baseline, with the temperature difference gradually increasing. By 3:00 PM one afternoon, the temperature change rate parameter had risen to 0.5℃ / hour. The continuous overheating accelerated the aging of the insulation components near the joint, inducing partial discharge. The partial discharge activity intensity parameter also gradually climbed from the background level to 0.6. During this period, the correlation analysis module calculated a state correlation coefficient of 0.85 between the temperature change rate parameter and the partial discharge activity intensity parameter over a 24-hour observation period, which was higher than the preset linkage threshold of 0.7.

[0114] When the state correlation coefficient exceeds the linkage threshold, the early warning generation module is activated. The system combines the state correlation coefficient of 0.85 with the changing trends of the temperature change rate parameter and the partial discharge activity intensity parameter to calculate a comprehensive risk index of 0.72. Based on the preset risk level table, the risk level is determined to be "alarm". The information output module then parses this graded fused early warning signal, generates and pushes a clear early warning message to the operation and maintenance center: "Alarm Level: Inferred to be an overheating fault caused by poor internal contact of the equipment, which has led to accelerated insulation degradation. Basis: The temperature change rate and the partial discharge activity intensity show a strong positive correlation, with a state correlation coefficient of 0.85."

[0115] Furthermore, in another monitoring of other types of defects, when the system detected partial discharge caused by suspended metal particles, spectral analysis revealed that its energy was concentrated in the high-risk target frequency band of 1.2-1.5 GHz. The system then automatically lowered the amplitude threshold for UHF signal detection by 3 dB, successfully capturing more weak discharge pulses that were drowned out by noise, thus enabling earlier warning times compared to the fixed threshold method.

[0116] It should be noted that the electrical connections between the various units described above do not necessarily represent direct or indirect connections. Any indirect connection method can be applied to the embodiments of the present invention as long as it achieves the purpose of the present invention. The above descriptions are merely exemplary embodiments of the present invention and should not be construed as limiting the scope of the present invention.

[0117] All equivalent changes and modifications made in accordance with the teachings of this invention are still within the scope of this invention. Those skilled in the art will readily conceive of other embodiments of this invention upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this invention that follow the general principles of this invention and include common knowledge or conventional techniques in the art not described herein.

Claims

1. A fault early warning method integrating SAW temperature and UHF partial discharge, characterized in that, The method includes: The process involves acquiring surface acoustic wave (SAW) temperature signals and ultra-high frequency (UHF) partial discharge signals from power equipment, synchronously collected by a preset integrated sensor, to generate raw time-series data. This includes transmitting a radio frequency interrogation signal to the preset integrated sensor to excite the SAW element within the integrated sensor; receiving a reflected signal carrying temperature information returned by the integrated sensor and simultaneously capturing the UHF partial discharge signal received by the integrated sensor to obtain a mixed signal; demodulating and digitizing the mixed signal and marking it with a consistent timestamp to form the raw time-series data. Based on the original time-series data, a temperature change rate parameter characterizing the abnormal thermal trend of the equipment and a partial discharge activity intensity parameter quantifying the insulation activity intensity are generated. This includes establishing a dynamic temperature baseline based on the surface acoustic wave (SAW) temperature signal in the original time-series data, filtering out environmental and load fluctuations; performing a differential calculation between the SAW temperature signal and the dynamic temperature baseline to obtain a temperature difference value; calculating the temperature change rate parameter based on the temperature difference value's change trend over time; performing pulse detection on the ultra-high frequency partial discharge signal in the original time-series data within a preset time window to obtain the pulse repetition frequency; measuring the amplitude of all discharge pulses within the time window and sorting them from largest to smallest, selecting the amplitude of the first-ranked discharge pulse to obtain the discharge amplitude; and weighted fusing the pulse repetition frequency and the discharge amplitude to generate the partial discharge activity intensity parameter. The statistical correlation between the time series of the temperature change rate parameter and the time series of the partial discharge activity intensity parameter is calculated to generate a state correlation coefficient. This includes constructing time series sample pairs of the temperature change rate parameter and the partial discharge activity intensity parameter within the same observation period; calculating the linear correlation strength of the time series sample pairs to obtain the correlation strength value; and calculating the state correlation coefficient based on the correlation strength value. A linkage threshold is constructed to characterize the physical coupling relationship between thermal condition deterioration and insulation degradation. The state correlation coefficient is compared with the linkage threshold to generate a graded fusion early warning signal. This includes constructing the linkage threshold to characterize the physical coupling relationship between thermal condition deterioration and insulation degradation; determining whether the absolute value of the state correlation coefficient is greater than the linkage threshold; if so, calculating the change in the temperature change rate parameter and the change in the partial discharge activity intensity parameter, and fusing them to obtain the monitoring data change value; calculating a comprehensive risk index based on the state correlation coefficient and the monitoring data change value; mapping the comprehensive risk index to a preset risk level table to determine the early warning level; and constructing and outputting a graded fusion early warning signal based on the early warning level. The hierarchical fusion early warning signal is analyzed to generate and output equipment fault early warning information containing fault nature inference.

2. The fault early warning method integrating SAW temperature and UHF partial discharge as described in claim 1, characterized in that, The process of parsing the hierarchical fusion early warning signal and generating and outputting equipment fault early warning information containing fault nature inference includes: The risk level is extracted from the hierarchical fusion early warning signal; Based on the risk level and the state correlation coefficient, equipment fault early warning information is generated, which includes fault nature inference and correlation analysis results.

3. The fault early warning method integrating SAW temperature and UHF partial discharge as described in claim 1, characterized in that, The method further includes: To obtain the frequency of electromagnetic wave energy generated by partial discharge; Determine whether the frequency of the electromagnetic wave energy is within the preset target frequency band; If so, the detection sensitivity of the ultra-high frequency partial discharge signal will be improved.

4. A fault early warning system integrating SAW temperature and UHF partial discharge, characterized in that, The system includes: The signal acquisition module is used to acquire surface acoustic wave temperature signals and ultra-high frequency partial discharge signals of power equipment synchronously acquired by a preset integrated sensor, and generate raw time-series data. This includes transmitting a radio frequency interrogation signal to the preset integrated sensor to excite the surface acoustic wave element within the integrated sensor; receiving reflected signals carrying temperature information returned by the integrated sensor, and simultaneously capturing the ultra-high frequency partial discharge signals received by the integrated sensor to obtain a mixed signal; demodulating and digitizing the mixed signal, and marking it with a consistent timestamp to form the raw time-series data. The parameter processing module is used to generate, based on the original time-series data, a temperature change rate parameter characterizing the abnormal thermal trend of the equipment and a partial discharge activity intensity parameter quantifying the insulation activity intensity. This includes establishing a dynamic temperature baseline based on the surface acoustic wave (SAW) temperature signal in the original time-series data, filtering out environmental and load fluctuations; performing a differential calculation between the SAW temperature signal and the dynamic temperature baseline to obtain a temperature difference value; calculating and generating a temperature change rate parameter based on the temperature difference value's change trend over time; performing pulse detection on the ultra-high frequency partial discharge signal in the original time-series data within a preset time window, and statistically obtaining the pulse repetition frequency; measuring the amplitude of all discharge pulses within the time window and sorting them from largest to smallest, selecting the amplitude of the first-ranked discharge pulse to obtain the discharge amplitude; and weightedly fusing the pulse repetition frequency and the discharge amplitude to generate the partial discharge activity intensity parameter. The correlation analysis module is used to calculate the statistical correlation between the time series of the temperature change rate parameter and the time series of the partial discharge activity intensity parameter, and generate a state correlation coefficient. This includes constructing time series sample pairs of the temperature change rate parameter and the partial discharge activity intensity parameter within the same observation period; calculating the linear correlation strength of the time series sample pairs to obtain the correlation strength value; and calculating the state correlation coefficient based on the correlation strength value. An early warning generation module is used to construct a linkage threshold characterizing the physical coupling relationship between thermal condition deterioration and insulation degradation, and compare the state correlation coefficient with the linkage threshold to generate a graded fused early warning signal. This includes constructing the linkage threshold characterizing the physical coupling relationship between thermal condition deterioration and insulation degradation; determining whether the absolute value of the state correlation coefficient is greater than the linkage threshold; if so, calculating the change in the temperature change rate parameter and the change in the partial discharge activity intensity parameter, and fusing them to obtain the monitoring data change value; calculating a comprehensive risk index based on the state correlation coefficient and the monitoring data change value; mapping the comprehensive risk index to a preset risk level table to determine the early warning level; and constructing and outputting a graded fused early warning signal based on the early warning level. The information output module is used to parse the hierarchical fusion early warning signal, generate and output equipment fault early warning information containing fault nature inference.

Images