Oil temperature prediction and early warning method, system and equipment based on big data analysis and medium

Abstract

The invention discloses an oil temperature prediction and early warning method, system, equipment and medium based on big data analysis, which not only can discover oil temperature abnormity in time, but also can provide an early warning and processing scheme in time under an abnormal condition by collecting key data of a transformer and performing real-time calculation, thereby being beneficial to prolonging the service life of the transformer and improving the safety. The problems that at present, oil temperature monitoring is mostly based on a fixed model or an empirical formula, the real-time performance is poor, and complex working conditions or emergencies cannot be coped with in time are solved.

Description

Technical Field

[0001] The present invention relates to power equipment monitoring, and in particular to an oil temperature prediction and early warning method, system, equipment and medium based on big data analysis. Background Art

[0002] Existing transformer designs are usually based on standard ambient temperature and rated operating conditions, and their normal life is expected to be under continuous operating conditions. However, the key temperature indicators inside the transformer - oil temperature and winding hot spot temperature - are affected by many factors. Once the oil temperature or winding temperature exceeds the allowable value, the transformer will face potential safety risks and accelerate aging.

[0003] In the prior art, overheating problems inside transformers are usually caused by cooling system failures, reduced cooling efficiency, or increased loads. These problems will directly lead to increased oil temperature and increase the risk of winding overheating. Although many systems rely on monitoring these temperature parameters, it is difficult for existing technologies to accurately predict temperature changes in real time. Traditional monitoring methods are mostly based on fixed models or empirical formulas, which cannot respond to complex operating conditions or emergencies in a timely manner, and it is difficult to detect and warn of abnormal increases in oil temperature in a timely and accurate manner. The shortcomings of current similar technologies also include: insufficient real-time performance: it is impossible to achieve real-time and accurate prediction of oil temperature, and it often needs to rely on subsequent data analysis; imperfect prediction and warning mechanism: the ability to identify the trend of temperature increase is weak, especially when abnormal conditions occur, the warning is delayed; insufficient emergency handling capabilities: existing technologies have limited performance in evaluating the short-term overload capacity of transformers and the response to cooling device failures, and it is difficult to support rapid decision-making. Summary of the invention

[0004] Based on the problems raised by the above background technology, the purpose of the present invention is to provide an oil temperature prediction and early warning method, system, equipment and medium based on big data analysis. By collecting key data of the transformer and performing real-time calculations, not only can oil temperature abnormalities be discovered in time, but also early warnings and processing solutions can be provided in time under abnormal circumstances, which helps to extend the service life of the transformer and improve safety. It solves the problem that the current monitoring of oil temperature is mostly based on fixed models or empirical formulas, which have poor real-time performance and cannot respond to complex working conditions or emergencies in time.

[0005] The present invention is achieved through the following technical solutions:

[0006] The first aspect of the present invention provides an oil temperature prediction and early warning method based on big data analysis, comprising the following steps:

[0007] Obtain transformer monitoring data, and perform data cleaning and preprocessing on the transformer monitoring data; wherein the transformer monitoring data includes oil temperature data, load rate data, and ambient temperature data;

[0008] Based on the oil temperature data and the load rate data, the pre-processed transformer monitoring data is searched for the stable interval and the maximum load rate interval to obtain the stable interval data and the maximum load rate interval data;

[0009] Constructing an oil temperature model, and using historical transformer monitoring data to perform fitting optimization on the oil temperature model;

[0010] The rated oil temperature rise is calculated for the stable interval data and the maximum load rate interval data based on the fitted optimized oil temperature model, and the result of the rated oil temperature rise calculation is comprehensively calculated with the ambient temperature data to obtain the full load oil temperature.

[0011] In the above technical solution, transformer monitoring data including oil temperature data, load rate data and ambient temperature data are first obtained, wherein oil temperature data refers to the actual oil temperature value, which will be used as reference data for the oil temperature model; load rate data is usually used to calculate the trend of oil temperature changes, which is used to adjust the prediction in the subsequent oil temperature model; ambient temperature data is used to calculate the external influencing factors of oil temperature. Since the transformer monitoring data comes from the real-time monitoring system, the data from different sources are screened and preprocessed in time series to obtain a structured and continuous data set. Since oil temperature, load rate and ambient temperature are important parameters affecting the cooling performance of the transformer, the accuracy and completeness of the data are directly related to the reliability of subsequent analysis.

[0012] Then, the preprocessed transformer monitoring data is searched for intervals to obtain the stable interval and the maximum load rate interval, and the interval data of the two are used to perform the fitting optimization of the oil temperature model and the calculation of the full-load oil temperature.

[0013] Among them, the historical transformer monitoring data is used to fit and optimize the oil temperature model, which can obtain a more accurate rated oil temperature rise. Then, the rated oil temperature rise in the stable range and the rated oil temperature rise in the maximum load rate range are calculated using the fitted and optimized oil temperature model. The full-load oil temperature can be obtained by comprehensive calculation using the above two rated oil temperature rises and ambient temperature data. The full-load oil temperature can reflect the temperature inside the transformer in real time and promptly feedback the overheating problem inside the transformer. This solves the problem that the current monitoring of oil temperature is mostly based on fixed models or empirical formulas, which has poor real-time performance and cannot respond to complex working conditions or emergencies in a timely manner.

[0014] In an optional embodiment, data cleaning and preprocessing are performed on the transformer monitoring data, comprising the following steps:

[0015] Performing timestamp alignment on the transformer monitoring data to obtain synchronous transformer monitoring data;

[0016] Conditionally filtering the synchronous transformer monitoring data using a sliding window, and extracting features from the conditionally filtered synchronous transformer monitoring data to obtain change trend features;

[0017] Interpolation processing is performed on the change trend characteristics to complete data cleaning and preprocessing.

[0018] In an optional embodiment, searching for a stable interval of the pre-processed transformer monitoring data based on the oil temperature data and the load rate data includes the following steps:

[0019] Any sliding window is defined as a search sliding window, and the oil temperature data in the search sliding window are subjected to slope calculation by linear regression to obtain the oil temperature slope;

[0020] Calculating the standard deviation of the oil temperature data and the load rate data in the search sliding window to obtain the oil temperature standard deviation and the load rate standard deviation, and weighting the oil temperature standard deviation and the load rate standard deviation to obtain a combined standard deviation;

[0021] A load rate variation range threshold and an oil temperature slope threshold are preset. When the variation range of the load rate data in the search sliding window is less than the load rate variation range threshold, the oil temperature slope is greater than the oil temperature slope threshold and the combined standard deviation is the smallest in all sliding windows, the search sliding window is defined as a stable interval, and the preprocessed transformer monitoring data in the stable interval is defined as stable interval data.

[0022] In an optional embodiment, searching for a maximum load rate interval of the pre-processed transformer monitoring data based on the oil temperature data and the load rate data includes the following steps:

[0023] An arbitrary sliding window is defined as a search sliding window, and the average value of the load rate data in the search sliding window is calculated to obtain the average load rate;

[0024] When the average load rate of the search sliding window is the largest among all sliding windows, the search sliding window is defined as a maximum load rate interval, and the preprocessed transformer monitoring data within the maximum load rate interval is defined as maximum load rate interval data.

[0025] In an optional embodiment, an oil temperature model is constructed, and the oil temperature model is fitted and optimized using historical transformer monitoring data, including the following steps:

[0026] The oil temperature differential equation is constructed by using the nonlinear relationship between the load rate and the oil temperature change rate and the temperature relationship between the oil temperature change rate and the ambient temperature;

[0027] Extract oil temperature data, load rate data and ambient temperature data from historical transformer monitoring data and divide them into training set and validation set;

[0028] Bringing the training set into the oil temperature differential equation for fitting, and obtaining a training rated oil temperature rise and a training oil time constant;

[0029] The oil temperature corresponding to the validation set is predicted using the training rated oil temperature rise and the training oil time constant to obtain predicted oil temperature data;

[0030] Find the data point with the largest difference between the predicted oil temperature data and the oil temperature data of the validation set. If the data point with the largest difference does not exceed the error threshold, use the training rated oil temperature rise as the output data of the oil temperature model.

[0031] In an optional embodiment, the rated oil temperature rise calculation is performed on the stable interval data and the maximum load rate interval data based on the optimized oil temperature model, including the following steps:

[0032] The first rated oil temperature rise of the stable interval data is calculated using the optimized oil temperature model;

[0033] The second rated oil temperature rise of the highest load rate interval data is calculated using the optimized oil temperature model;

[0034] The ratio of the first rated oil temperature rise to the second rated oil temperature rise is calculated to obtain a cooling capacity ratio.

[0035] In an optional embodiment, the result of the rated oil temperature rise calculation is comprehensively calculated with the ambient temperature data, including: adding the quotient of the first rated oil temperature rise divided by the cooling capacity ratio and adding the ambient temperature data to obtain the full load oil temperature.

[0036] The second aspect of the present invention provides an oil temperature prediction and early warning system based on big data analysis, comprising:

[0037] A data acquisition module is used to acquire transformer monitoring data and perform data cleaning and preprocessing on the transformer monitoring data; wherein the transformer monitoring data includes oil temperature, load rate and ambient temperature;

[0038] An interval search module is used to search for a stable interval and a maximum load rate interval of the pre-processed transformer monitoring data based on the oil temperature data and the load rate data, and obtain stable interval data and maximum load rate interval data;

[0039] An oil temperature model module is used to construct an oil temperature model and to fit and optimize the oil temperature model using historical transformer monitoring data;

[0040] The full-load oil temperature calculation module is used to calculate the rated oil temperature rise for the stable interval data and the maximum load rate interval data based on the fitted optimized oil temperature model, and to comprehensively calculate the result of the rated oil temperature rise calculation and the ambient temperature data to obtain the full-load oil temperature.

[0041] A third aspect of the present invention provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements an oil temperature prediction and early warning method based on big data analysis when executing the program.

[0042] A fourth aspect of the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements an oil temperature prediction and early warning method based on big data analysis.

[0043] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0044] 1. An oil temperature fitting model based on big data is proposed, which can accurately predict the oil temperature of the transformer under full load, providing support for the implementation of oil temperature early warning work;

[0045] 2. The present invention takes the switching problem of coolers into consideration and proposes a cooling capacity ratio algorithm. By mining and analyzing historical data, the influence of the number of coolers on the oil temperature is converted into a mathematical model, which solves the problem that the current fixed model and empirical formula for transformer oil temperature prediction do not consider the number of coolers. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the following briefly introduces the drawings required for use in the embodiments. It should be understood that the following drawings only illustrate certain embodiments of the present invention and should not be regarded as limiting the scope. For those of ordinary skill in the art, other relevant drawings can be obtained based on these drawings without creative work. In the drawings:

[0047] Figure 1 A schematic diagram of a process flow for preprocessing transformer monitoring data provided in Embodiment 1 of the present invention;

[0048] Figure 2 A schematic diagram of a stable interval search result provided in Example 1 of the present invention;

[0049] Figure 3 A schematic diagram of the flow chart of oil temperature model fitting provided in Example 1 of the present invention;

[0050] Figure 4 A schematic diagram of the oil temperature model fitting results provided in Example 1 of the present invention;

[0051] Figure 5 A schematic diagram of the structure of an electronic device provided in Example 3 of the present invention. DETAILED DESCRIPTION

[0052] In order to make the objectives, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with embodiments and drawings. The exemplary embodiments of the present invention and their description are only used to explain the present invention and are not intended to limit the present invention.

[0053] Example 1

[0054] Embodiment 1 of the present invention provides an oil temperature prediction and early warning method based on big data analysis, which comprises the following steps:

[0055] Obtain transformer monitoring data, and perform data cleaning and preprocessing on the transformer monitoring data; wherein the transformer monitoring data includes oil temperature data, load rate data, and ambient temperature data;

[0056] Based on the oil temperature data and the load rate data, the pre-processed transformer monitoring data is searched for the stable interval and the maximum load rate interval to obtain the stable interval data and the maximum load rate interval data;

[0057] Constructing an oil temperature model, and using historical transformer monitoring data to perform fitting optimization on the oil temperature model;

[0058] The rated oil temperature rise is calculated for the stable interval data and the maximum load rate interval data based on the fitted optimized oil temperature model, and the result of the rated oil temperature rise calculation is comprehensively calculated with the ambient temperature data to obtain the full load oil temperature.

[0059] It should be noted that the existing technology has a weak ability to identify the trend of transformer temperature increase, and usually adopts a fixed model or empirical formula. The fixed model and empirical formula are difficult to deal with the internal overheating problem of the transformer under complex working conditions.

[0060] Based on the defects of the above-mentioned prior art, the present invention proposes an oil temperature prediction and early warning method based on big data analysis. First, transformer monitoring data including oil temperature data, load rate data and ambient temperature data are obtained, wherein oil temperature data refers to the actual oil temperature value, which will be used as reference data for the oil temperature model; load rate data is usually used to calculate the changing trend of oil temperature, which is used to adjust the prediction in the subsequent oil temperature model; ambient temperature data is used to calculate the external influencing factors of oil temperature. Since the transformer monitoring data comes from the real-time monitoring system, the data from different sources are screened and preprocessed according to the time series to obtain a structured and continuous data set. Since oil temperature, load rate and ambient temperature are important parameters affecting the cooling performance of the transformer, the accuracy and completeness of the data are directly related to the reliability of the subsequent analysis.

[0061] Then, the preprocessed transformer monitoring data is searched for intervals to obtain the stable interval and the maximum load rate interval, and the interval data of the two are used to perform the fitting optimization of the oil temperature model and the calculation of the full-load oil temperature.

[0062] Among them, the historical transformer monitoring data is used to fit and optimize the oil temperature model, which can obtain a more accurate rated oil temperature rise. Then, the rated oil temperature rise in the stable range and the rated oil temperature rise in the maximum load rate range are calculated using the fitted and optimized oil temperature model. The full-load oil temperature can be obtained by comprehensive calculation using the above two rated oil temperature rises and ambient temperature data. The full-load oil temperature can reflect the temperature inside the transformer in real time and promptly feedback the overheating problem inside the transformer. This solves the problem that the current monitoring of oil temperature is mostly based on fixed models or empirical formulas, which has poor real-time performance and cannot respond to complex working conditions or emergencies in a timely manner.

[0063] In an optional embodiment, data cleaning and preprocessing are performed on the transformer monitoring data, comprising the following steps:

[0064] Performing timestamp alignment on the transformer monitoring data to obtain synchronous transformer monitoring data;

[0065] Conditionally filtering the synchronous transformer monitoring data using a sliding window, and extracting features from the conditionally filtered synchronous transformer monitoring data to obtain change trend features;

[0066] Interpolation processing is performed on the change trend characteristics to complete data cleaning and preprocessing.

[0067] It should be noted that the process of data cleaning and preprocessing is as follows Figure 1 As shown, the transformer monitoring data is first read and parsed into a format suitable for processing, and then the time series from different sources are timestamp aligned to ensure the consistency of time points. In this embodiment, the alignment processing mainly involves oil temperature data, load rate data and ambient temperature data.

[0068] Perform initial settings for the sliding window, determine the sliding window size and step length, and slide the sliding window step by step. During the sliding process of the sliding window, calculate the mean, standard deviation and other statistical indicators of the oil temperature data, load rate data and ambient temperature data, and perform conditional filtering based on the calculated statistical indicators to select data that meets the conditions.

[0069] The key change trends and features of transformer monitoring data are extracted through trend analysis such as linear regression. In order to ensure the continuity and integrity of the pre-processed data, the missing parts of the data are supplemented through interpolation processing, and the data is adjusted to be consistent on the time axis. All processed data will be aligned again to provide standardized data for subsequent further analysis and processing.

[0070] In an optional embodiment, searching for a stable interval of the pre-processed transformer monitoring data based on the oil temperature data and the load rate data includes the following steps:

[0071] Any sliding window is defined as a search sliding window, and the oil temperature data in the search sliding window are subjected to slope calculation by linear regression to obtain the oil temperature slope;

[0072] Calculating the standard deviation of the oil temperature data and the load rate data in the search sliding window to obtain the oil temperature standard deviation and the load rate standard deviation, and weighting the oil temperature standard deviation and the load rate standard deviation to obtain a combined standard deviation;

[0073] A load rate variation range threshold and an oil temperature slope threshold are preset. When the variation range of the load rate data in the search sliding window is less than the load rate variation range threshold, the oil temperature slope is greater than the oil temperature slope threshold and the combined standard deviation is the smallest in all sliding windows, the search sliding window is defined as a stable interval, and the preprocessed transformer monitoring data in the stable interval is defined as stable interval data.

[0074] It should be noted that in this embodiment, 8 hours is used as a sliding window. The load rate range, combined standard deviation, and oil temperature slope are considered when searching for a stable interval. If the load rate fluctuation range within 8 hours is less than the preset threshold; the combined standard deviation is the smallest among all 8-hour windows in the current interval; and the slope of the oil temperature is greater than the preset slope threshold, then the time window is considered to be the most stable interval at the moment.

[0075] Due to measurement errors, there are oil temperature data that do not change in the original data, but this is not realistic. In order to avoid finding such data, it is necessary to measure the rate of change of the oil temperature. Therefore, this embodiment uses a linear regression method to calculate the slope of the oil temperature change over time.

[0076] Linear regression is essentially a method of fitting a straight line. It finds the best fit line by minimizing the sum of squared errors. In this scenario, the algorithm selects time as the independent variable (X) and oil temperature as the dependent variable (Y). The specific process is as follows: First, time is discretized into a number of equally spaced points (one point every 15 minutes), which form an arithmetic sequence. This sequence can be regarded as the independent variable X, which is a representation of time. Then, for each consecutive 8-hour time window, the corresponding oil temperature data is extracted, and these oil temperature values ​​constitute the dependent variable Y. The core goal of linear regression is to find a linear function Y=aX+b, where a is the slope, which is used in this embodiment to represent the rate of change of oil temperature per unit time.

[0077] Therefore, the process of calculating the oil temperature slope through linear regression can be briefly described as: matching each time point with the corresponding oil temperature, and determining a best-fitting straight line by minimizing the square error between the predicted value and the actual oil temperature value. Ultimately, the slope obtained by the algorithm reflects the rate of change of the oil temperature within the 8-hour window. If the slope is positive, it means that the oil temperature is on an upward trend during this period; if the slope is negative, the oil temperature is on a downward trend during this period.

[0078] In order to find the most stable interval, the present embodiment also considers the stability of the load rate and the oil temperature. In the present embodiment, the stability is measured by the load rate standard deviation and the oil temperature standard deviation.

[0079] Specifically, the screening process is as follows: In an 8-hour time window, the algorithm first calculates the standard deviation of the load rate within the interval. If the standard deviation of the load rate is small, it means that the load in the interval is relatively stable; conversely, a large standard deviation means that the load changes greatly and is unstable. The same applies to the standard deviation of the oil temperature.

[0080] The algorithm then weights the standard deviation of the load rate and the standard deviation of the oil temperature to obtain a combined standard deviation. This combined standard deviation reflects the overall fluctuation of the load rate and oil temperature. The smaller the combined standard deviation, the smaller the fluctuation of the load rate and oil temperature in this 8-hour time window, and the more stable the system.

[0081] In order to illustrate the effectiveness of the stable interval search method, the historical operation data of a transformer in an ultra-high voltage substation is selected for example analysis. The stable interval search results are shown in the figure below. Figure 2 As shown in the figure, the yellow interval represents the stable interval found, and the stability of this interval can be proved by analyzing multiple key indicators. First, the change trend of the load rate (blue line) and the oil temperature (red line) is stable, and the fluctuation range is small, which shows that the equipment's operating status has been continuously stable during this period. Secondly, the sum of the standard deviations of the load rate and the oil temperature (purple dotted line) is used to measure volatility. In the yellow interval, the sum of the standard deviations remains at a relatively low level, indicating that the fluctuations of the load rate and the oil temperature are small and the operating status is stable. In addition, the linear regression results of the load rate and the oil temperature in the stable interval (green line) show a relatively smooth trend, which further verifies that the volatility of this interval is low and relatively stable.

[0082] In an optional embodiment, searching for a maximum load rate interval of the pre-processed transformer monitoring data based on the oil temperature data and the load rate data includes the following steps:

[0083] An arbitrary sliding window is defined as a search sliding window, and the average value of the load rate data in the search sliding window is calculated to obtain the average load rate;

[0084] When the average load rate of the search sliding window is the largest among all sliding windows, the search sliding window is defined as a maximum load rate interval, and the preprocessed transformer monitoring data within the maximum load rate interval is defined as maximum load rate interval data.

[0085] In an optional embodiment, an oil temperature model is constructed, and the oil temperature model is fitted and optimized using historical transformer monitoring data, including the following steps:

[0086] The oil temperature differential equation is constructed by using the nonlinear relationship between the load rate and the oil temperature change rate and the temperature relationship between the oil temperature change rate and the ambient temperature;

[0087] Extract oil temperature data, load rate data and ambient temperature data from historical transformer monitoring data and divide them into training set and validation set;

[0088] Bringing the training set into the oil temperature differential equation for fitting, and obtaining a training rated oil temperature rise and a training oil time constant;

[0089] The oil temperature corresponding to the validation set is predicted using the training rated oil temperature rise and the training oil time constant to obtain predicted oil temperature data;

[0090] Find the data point with the largest difference between the predicted oil temperature data and the oil temperature data of the validation set. If the data point with the largest difference does not exceed the error threshold, use the training rated oil temperature rise as the output data of the oil temperature model.

[0091] It should be noted that the oil temperature differential equation is constructed as follows:

[0092]

[0093] Where k is the load factor, R is the loss ratio, X is the oil index, Δθ 0r is the rated oil temperature rise, k 11 τ o are thermal model constants and oil time constants, θ0 is oil temperature, θ a is the ambient temperature.

[0094] The square term k represents the nonlinear relationship between the load rate and the oil temperature change rate. 2 R reflects the contribution of load change to oil temperature change, and this contribution is adjusted by power X to accurately control the nonlinear change of oil temperature response.

[0095] It represents the temperature relationship between the oil temperature change rate and the ambient temperature, where the time constant τ o Indicates the time required for the oil temperature to gradually stabilize after being affected by load changes. Oil temperature change rate is the rate of change of oil temperature at the current moment, which is related to the time constant τ o Multiplying them together describes the response speed of the oil temperature to the load change. a Describes the difference between oil temperature and ambient temperature, indicating the impact of ambient temperature on oil temperature.

[0096] To ensure that the model can accurately predict the oil temperature under full load conditions, the key parameter Δθ in the model 0r and τ o Optimization is required through data fitting.

[0097] These two key parameters are trained and fitted using historical operating data to ensure that the model has high prediction accuracy under different load conditions. The fitting process determines the optimal ΔΔθ by minimizing the error between the predicted oil temperature and the actual oil temperature. 0r and τ o In practical applications, the optimized parameters will be dynamically adjusted according to different transformer types and environmental conditions to ensure the wide applicability of the model. The processing of these parameters enables the model to adapt to different operating conditions, further improving the accuracy and robustness of the prediction.

[0098] In this embodiment, the oil temperature model fitting flow chart is as follows: Figure 3 As shown, the 8-hour historical transformer monitoring data is divided into a training set and a validation set, where the data for the first 4 hours is used as a training set and the data for the last 4 hours is used as a validation set. The time differential of the oil temperature is calculated at a time interval of 15 minutes. Through the oil temperature model, according to the loss ratio R, load rate k and other parameters of the substation, the rated oil temperature rise and the oil time constant are fitted. The fitting parameters are used to calculate the oil temperature for the last 4 hours and compared with the actual oil temperature data. The data point with the largest difference is calculated to evaluate the quality of the fitting effect. If the maximum difference does not exceed 5 degrees Celsius, it is considered that the fitting effect meets expectations. The oil temperature value is obtained by adding the fitted rated oil temperature rise to the ambient temperature, such as Figure 4 As shown, it can be observed that the difference between the predicted oil temperature and the actual oil temperature is small, which indicates that the parameter fitting effect is good.

[0099] In an optional embodiment, the rated oil temperature rise calculation is performed on the stable interval data and the maximum load rate interval data based on the optimized oil temperature model, including the following steps:

[0100] The first rated oil temperature rise of the stable interval data is calculated using the optimized oil temperature model;

[0101] The second rated oil temperature rise of the highest load rate interval data is calculated using the optimized oil temperature model;

[0102] The ratio of the first rated oil temperature rise to the second rated oil temperature rise is calculated to obtain a cooling capacity ratio.

[0103] It should be noted that due to different load rates and different numbers of coolers, under full load conditions, the number of coolers is greater and the rated oil temperature rise is lower, so the rated oil temperature rise calculated by the oil temperature model cannot be used to calculate the final predicted oil temperature. Based on this, in the embodiment of the present invention, the cooling capacity ratio is used to calculate the final predicted oil temperature.

[0104] The specified cooling capacity ratio is the ratio of the rated oil temperature rise at two different load rates. It is used to measure the approximate cooling performance ratio of the two load rates. Since at a lower load rate, the number of coolers put into use is relatively small, and at a higher load rate, the number of coolers put into use increases accordingly, which is bound to affect the oil temperature. If we want to use the data of a lower load rate to predict the oil temperature under full load, and the number of coolers at a lower load rate is definitely less than the number at a high load rate, we need to use the cooling capacity ratio to reflect the impact of the number of coolers on the oil temperature.

[0105] Specifically, the load rate is divided into sections with 10% as a stage in a certain time period, for example, the load rate is from 0 to 0.1 as a section, 0.1 to 0.2 as a section, and so on, divided into 10 sections, and 8 8-hour data intervals of each load rate section of the transformer in this time period are found. The time interval between each data interval is as long as possible, and the time intervals available for selection are 5 days, 3 days, 1 day, 8 hours, 4 hours, and 1 hour. Read the continuous 8-hour intervals with high average load rate found before, use the oil temperature model to calculate their predicted oil temperatures respectively, and divide the former result by the latter result to obtain the cooling capacity ratio of each load rate section.

[0106] In this embodiment, it is assumed that the 0.8-0.9 section is found to be the highest load rate section, and the number of coolers in this section is regarded as the same as that at full load, so the cooling capacity ratio of this section is used as a reference value of 1, and a rated oil temperature rise B = 33.3°C is calculated from the data of the 0.8-0.9 section. A rated oil temperature rise A = 40.0°C is calculated from the stable interval data of the 0.4-0.5 section, so the cooling capacity ratio is C = A / B = 1.20. By analogy, the cooling capacity ratios of all other load rate sections compared to the highest load rate section can be calculated, and the cooling capacity ratio result table of a substation is shown in Table 1 below:

[0107]

[0108]

[0109] Table 1 Cooling capacity ratio results of a substation

[0110] When using the data of a certain section to predict the oil temperature in the future, query the cooling capacity ratio of the corresponding section in this table, divide the rated oil temperature rise by the cooling capacity ratio, and then add the ambient temperature to get the final predicted full-load oil temperature.

[0111] Predicting the full-load oil temperature can help operation and maintenance personnel effectively evaluate the operating status of the transformer under high load, identify potential overheating risks in advance, and ensure the safe and stable operation of the equipment.

[0112] In an optional embodiment, the result of the rated oil temperature rise calculation is comprehensively calculated with the ambient temperature data, including: adding the quotient of the first rated oil temperature rise divided by the cooling capacity ratio and adding the ambient temperature data to obtain the full load oil temperature.

[0113] Example 2

[0114] Embodiment 2 of the present invention provides an oil temperature prediction and early warning system based on big data analysis, including:

[0115] A data acquisition module is used to acquire transformer monitoring data and perform data cleaning and preprocessing on the transformer monitoring data; wherein the transformer monitoring data includes oil temperature, load rate and ambient temperature;

[0116] An interval search module is used to search for a stable interval and a maximum load rate interval of the pre-processed transformer monitoring data based on the oil temperature data and the load rate data, and obtain stable interval data and maximum load rate interval data;

[0117] An oil temperature model module is used to construct an oil temperature model and to fit and optimize the oil temperature model using historical transformer monitoring data;

[0118] The full-load oil temperature calculation module is used to calculate the rated oil temperature rise for the stable interval data and the maximum load rate interval data based on the fitted optimized oil temperature model, and to comprehensively calculate the result of the rated oil temperature rise calculation and the ambient temperature data to obtain the full-load oil temperature.

[0119] Example 3

[0120] Figure 5 A schematic diagram of the structure of an electronic device provided in Embodiment 3 of the present invention is shown in FIG. Figure 5 As shown, the electronic device includes a processor 21, a memory 22, an input device 23 and an output device 24; the number of processors 21 in the computer device can be one or more. Figure 5 A processor 21 is taken as an example; the processor 21, the memory 22, the input device 23 and the output device 24 in the electronic device can be connected by a bus or other means. Figure 5The example of connecting through bus is taken in the following.

[0121] The memory 22 is a computer-readable storage medium that can be used to store software programs, computer executable programs, and modules. The processor 21 executes various functional applications and data processing of the electronic device by running the software programs, instructions, and modules stored in the memory 22, that is, the oil temperature prediction and early warning method based on big data analysis of Example 1 is implemented.

[0122] The memory 22 may mainly include a program storage area and a data storage area, wherein the program storage area may store an operating system and at least one application required for a function; the data storage area may store data created according to the use of the terminal, etc. In addition, the memory 22 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some instances, the memory 22 may further include a memory remotely arranged relative to the processor 21, and these remote memories may be connected to the electronic device via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0123] The input device 23 can be used to receive the ID and password input by the user, etc. The output device 24 is used to output the network configuration page.

[0124] Example 4

[0125] Embodiment 4 of the present invention further provides a computer-readable storage medium, wherein the computer-executable instructions, when executed by a computer processor, are used to implement the oil temperature prediction and early warning method based on big data analysis as provided in Embodiment 1.

[0126] An embodiment of the present invention provides a storage medium containing computer executable instructions, and its computer executable instructions are not limited to the method operations provided in Example 1, and can also execute related operations in the oil temperature prediction and early warning method based on big data analysis provided in any embodiment of the present invention.

[0127] The specific implementation methods described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific implementation method of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. The oil temperature prediction and early warning method based on big data analysis is characterized by: The steps include: Obtain transformer monitoring data, and perform data cleaning and preprocessing on the transformer monitoring data; wherein the transformer monitoring data includes oil temperature data, load rate data, and ambient temperature data; Based on the oil temperature data and the load rate data, the pre-processed transformer monitoring data is searched for the stable interval and the maximum load rate interval to obtain the stable interval data and the maximum load rate interval data; Constructing an oil temperature model, and using historical transformer monitoring data to perform fitting optimization on the oil temperature model; The rated oil temperature rise is calculated for the stable interval data and the maximum load rate interval data based on the fitted optimized oil temperature model, and the result of the rated oil temperature rise calculation is comprehensively calculated with the ambient temperature data to obtain the full load oil temperature.

2. The oil temperature prediction and early warning method based on big data analysis according to claim 1 is characterized in that: The transformer monitoring data is cleaned and preprocessed, including the following steps: Performing timestamp alignment on the transformer monitoring data to obtain synchronous transformer monitoring data; Conditionally filtering the synchronous transformer monitoring data using a sliding window, and extracting features from the conditionally filtered synchronous transformer monitoring data to obtain change trend features; Interpolation processing is performed on the change trend characteristics to complete data cleaning and preprocessing.

3. The oil temperature prediction and early warning method based on big data analysis according to claim 2 is characterized in that: Based on the oil temperature data and the load rate data, the stable interval search of the pre-processed transformer monitoring data includes the following steps: Any sliding window is defined as a search sliding window, and the oil temperature data in the search sliding window are subjected to slope calculation by linear regression to obtain the oil temperature slope; Calculating the standard deviation of the oil temperature data and the load rate data in the search sliding window to obtain the oil temperature standard deviation and the load rate standard deviation, and weighting the oil temperature standard deviation and the load rate standard deviation to obtain a combined standard deviation; A load rate variation range threshold and an oil temperature slope threshold are preset. When the variation range of the load rate data in the search sliding window is less than the load rate variation range threshold, the oil temperature slope is greater than the oil temperature slope threshold and the combined standard deviation is the smallest in all sliding windows, the search sliding window is defined as a stable interval, and the preprocessed transformer monitoring data in the stable interval is defined as stable interval data.

4. The oil temperature prediction and early warning method based on big data analysis according to claim 2 is characterized in that: Based on the oil temperature data and the load rate data, the highest load rate interval is searched for the pre-processed transformer monitoring data, including the following steps: An arbitrary sliding window is defined as a search sliding window, and the average value of the load rate data in the search sliding window is calculated to obtain the average load rate; When the average load rate of the search sliding window is the largest among all sliding windows, the search sliding window is defined as a maximum load rate interval, and the preprocessed transformer monitoring data within the maximum load rate interval is defined as maximum load rate interval data.

5. The oil temperature prediction and early warning method based on big data analysis according to claim 1 is characterized in that: Constructing an oil temperature model and using historical transformer monitoring data to perform fitting optimization on the oil temperature model includes the following steps: The oil temperature differential equation is constructed by using the nonlinear relationship between the load rate and the oil temperature change rate and the temperature relationship between the oil temperature change rate and the ambient temperature; Extract oil temperature data, load rate data and ambient temperature data from historical transformer monitoring data and divide them into training set and validation set; Substitute the training set into the oil temperature differential equation for fitting, and obtain the training rated oil temperature rise and the training oil time constant; The oil temperature corresponding to the validation set is predicted using the training rated oil temperature rise and the training oil time constant to obtain predicted oil temperature data; Find the data point with the largest difference between the predicted oil temperature data and the oil temperature data of the validation set. If the data point with the largest difference does not exceed the error threshold, use the training rated oil temperature rise as the output data of the oil temperature model.

6. The oil temperature prediction and early warning method based on big data analysis according to claim 1 is characterized in that: The rated oil temperature rise calculation is performed on the stable interval data and the maximum load rate interval data based on the optimized oil temperature model, including the following steps: The first rated oil temperature rise of the stable interval data is calculated using the optimized oil temperature model; The second rated oil temperature rise of the highest load rate interval data is calculated using the optimized oil temperature model; The ratio of the first rated oil temperature rise to the second rated oil temperature rise is calculated to obtain a cooling capacity ratio.

7. The oil temperature prediction and early warning method based on big data analysis according to claim 6 is characterized in that: The result of the rated oil temperature rise calculation is comprehensively calculated with the ambient temperature data, including: adding the quotient of the first rated oil temperature rise divided by the cooling capacity ratio and the ambient temperature data to obtain the full load oil temperature.

8. The oil temperature prediction and early warning system based on big data analysis is characterized by: include: A data acquisition module is used to acquire transformer monitoring data and perform data cleaning and preprocessing on the transformer monitoring data; wherein the transformer monitoring data includes oil temperature, load rate and ambient temperature; An interval search module is used to search for a stable interval and a maximum load rate interval of the pre-processed transformer monitoring data based on the oil temperature data and the load rate data, and obtain stable interval data and maximum load rate interval data; An oil temperature model module is used to construct an oil temperature model and to fit and optimize the oil temperature model using historical transformer monitoring data; The full-load oil temperature calculation module is used to calculate the rated oil temperature rise for the stable interval data and the maximum load rate interval data based on the fitted optimized oil temperature model, and to comprehensively calculate the result of the rated oil temperature rise calculation and the ambient temperature data to obtain the full-load oil temperature.

9. An electronic device, characterized in that: It includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the oil temperature prediction and early warning method based on big data analysis as described in any one of claims 1 to 7 is implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the oil temperature prediction and early warning method based on big data analysis as described in any one of claims 1 to 7 is implemented.

Images