Oil level online monitoring method based on ultrasonic waves

Abstract

The invention relates to an online oil level monitoring method based on ultrasonic waves. The online oil level monitoring method specifically comprises the steps that ultrasonic probes are mounted at the bottom and the outer wall under an oil conservator in an adsorption mode through an ultrasonic coupling agent; ultrasonic waves enter the oil conservator through the coupling agent and are reflected when reaching the surface of the oil level in the oil conservator, a propagation path is calculated according to echo time, and then the height of the oil level is obtained; aiming at interference influence, measurement compensation is carried out, and a final result is corrected; the multi-frequency ultrasonic measuring sensor measures the height of the oil level in the oil conservator in real time and transmits the height to the system background; the multi-frequency ultrasonic measurement sensor compares the measured data with the reading of the traditional oil level gauge, and regularly performs online verification on the mechanical oil level; on the basis of online monitoring data, a corresponding relation graph of the oil level of the oil conservator, the environment and the working condition is established, and pre-judgment and early warning are achieved; based on the echo data information, the operation state is judged through a multivariate statistical analysis algorithm; the method has the advantages that the false oil level can be quickly distinguished, and the detection redundancy is increased.

Description

Technical Field

[0001] The invention belongs to the technical field of transformer oil level detection, and in particular relates to an ultrasonic-based online oil level monitoring method. Background Art

[0002] The oil pillow is an oil storage device for transformers. When the volume of transformer oil expands or shrinks with the change of oil temperature, the oil pillow plays the role of storing and replenishing oil, which can ensure that the oil tank is full of oil. At the same time, due to the equipment of the oil pillow, the contact area between the transformer and the air is reduced, and the moisture, dust and oxidized oil scale absorbed from the air are deposited in the sedimentation device at the bottom of the oil pillow, thereby greatly slowing down the deterioration rate of the transformer oil. In actual use, the oil pillow is equipped with a connected oil level gauge or a connecting rod oil level gauge. In daily inspections, the inspection personnel can directly see the liquid level height in the oil pillow through the oil level gauge. However, in actual use, the connected oil level gauge in the oil pillow has been blocked, resulting in the failure of the oil level gauge, and the capsule or diaphragm in the oil pillow has been stuck to the connecting rod oil level gauge, causing the oil level gauge to fail.

[0003] Normal and abnormal oil levels are judged by the same device. Faults may occur during operation. The transformer may have hidden oil leaks, such as slow leakage of insulating oil from the inlet and outlet oil pipes buried at the bottom. If the oil level gauge is stuck and cannot indicate the actual oil level of the transformer, the operator cannot judge whether the oil level is normal based on the oil level curve during daily inspections. When the transformer oil drops to the point where the capsule expands to its maximum volume, the most critical oil level abnormality signal cannot be sent. If it is allowed to develop, due to the thermal expansion and contraction effect of the transformer oil, atmospheric pressure will press air into the transformer from the oil leak point, and tiny air bubbles will randomly enter the bushing riser and the oil pillow, and enter the gas relay to send out a light gas action signal. However, since the oil level displayed by the oil level gauge at this time is within the "normal" range, the oil leak is relatively hidden, and it is difficult to determine the cause of the light gas action in a short period of time.

[0004] The false oil level of the transformer means that the oil level indicated by the oil level gauge is inconsistent with the actual oil level, making it difficult to discover the hidden dangers of the equipment. When the oil level on the oil level gauge is lower than the actual oil level, the maintenance personnel will add oil to the transformer oil conservator. After increasing the oil volume to the "normal" oil level based on the indication of the oil level gauge, the actual oil in the oil conservator will be much. At this time, if the ambient temperature rises or the load increases, the oil will expand due to heat and will spray out from the pressure releaser of the transformer; when the oil level on the oil level gauge is higher than the actual oil level, the oil level indicated by the oil level gauge of the oil conservator is "normal", but the actual oil level is already at a very low position. This situation will cause the insulation strength of the internal winding and the casing of the transformer to be insufficient. At this time, the first exposed bushing will discharge to the casing, causing unplanned shutdown of the equipment; therefore, it is very necessary to provide an ultrasonic-based online oil level monitoring method that can quickly identify "false oil levels", increase detection redundancy, and prevent unplanned shutdowns caused by abnormal oil levels. Summary of the invention

[0005] 1. Technical issues

[0006] In view of the above-mentioned existing technical status, this application mainly addresses the following technical problems:

[0007] 1. There may be hidden oil leakage during operation. If the oil level gauge is stuck and cannot indicate the actual oil level of the transformer, the operator cannot judge whether the oil level is normal based on the oil level curve during daily inspections.

[0008] 2. Both normal and abnormal oil level conditions are judged by the same device. The design redundancy is insufficient, and it is easy for the oil level indicated by the oil level gauge to be inconsistent with the actual oil level, making it difficult to discover hidden dangers in the equipment.

[0009] (II) Technical solution

[0010] The purpose of the present invention is to overcome the deficiencies of the prior art and to provide an ultrasonic-based online oil level monitoring method that can quickly identify "false oil levels", increase detection redundancy, and prevent unplanned shutdowns caused by abnormal oil levels.

[0011] The object of the present invention is achieved by: an ultrasonic-based online oil level monitoring method, the method comprising the following steps:

[0012] Step 1: Install the ultrasonic probe of the multi-frequency ultrasonic measurement sensor on the bottom and outer wall directly below the transformer oil pillow through ultrasonic coupling agent, and ensure that the ultrasonic transmitting and receiving planes are parallel to the liquid surface of the transformer oil pillow;

[0013] Step 2: The ultrasonic transmitting and receiving control unit of the multi-frequency ultrasonic measuring sensor transmits 2-10MHz ultrasonic waves, which enter the transformer oil pillow through the coupling agent. When the ultrasonic waves reach the oil level surface in the oil pillow, they are reflected. The propagation path is calculated by the echo time to obtain the oil level height.

[0014] Step 3: For the interference effect of ultrasonic wave in oil level measurement caused by oil temperature of oil pillow transformer, direct wave interference and random interference, measurement compensation is performed to correct the final result;

[0015] Step 4: The multi-frequency ultrasonic measurement sensor measures the height of the oil level in the transformer oil conservator in real time, and transmits the measurement data to the system background via wired or wireless means, so as to understand the height of the oil level in the transformer oil conservator in real time;

[0016] Step 5: The multi-frequency ultrasonic measurement sensor uses its measurement data to compare with the readings of the traditional connected or connecting rod oil level gauge, and regularly performs online calibration of the mechanical oil level;

[0017] Step 6: Based on the online monitoring data of the transformer oil conservator oil level, environmental data and the actual working condition of the transformer are obtained while monitoring the oil level, and a corresponding relationship diagram between the transformer oil conservator oil level and the transformer environment and working condition is established to achieve early prediction and early warning of the transformer oil conservator failure;

[0018] Step 7: Based on the echo data information during the ultrasonic detection in step 2, the moisture content, breakdown voltage, dielectric loss, volume resistivity, and oil interfacial tension parameters of the transformer oil are extracted through a multivariate statistical analysis algorithm to determine the operating status of the transformer.

[0019] Furthermore, the multi-frequency ultrasonic measurement sensor in step 1 includes an ultrasonic probe, a digital display, a display module, a data processor, a data receiving module and an analysis module; the digital processor is connected to the display module, the data receiving module and the analysis module respectively.

[0020] Furthermore, the multi-frequency ultrasonic measurement sensor also includes an ultrasonic transmitting and receiving control unit and a multi-frequency ultrasonic sensor; the ultrasonic probe is connected to the multi-frequency ultrasonic sensor; and the ultrasonic transmitting and receiving control unit is composed of an ultrasonic transmitting module and an ultrasonic receiving module.

[0021] Furthermore, the ultrasonic transmitting and receiving control unit is equipped with a power interface, a power switch, an Internet network interface, an ultrasonic signal output interface and a signal receiving interface.

[0022] Furthermore, a multi-frequency ultrasonic generating device is provided in the ultrasonic transmitting and receiving control unit, and the multi-frequency ultrasonic generating device is internally connected to the ultrasonic signal output interface of the ultrasonic transmitting and receiving control unit, and the output interface is connected to the multi-frequency ultrasonic sensor via a wire, and the output signal of the multi-frequency sensor is connected to the ultrasonic signal receiving interface of the ultrasonic transmitting and receiving control unit via a wire, and the ultrasonic signal receiving interface transmits the signal to the signal processing circuit of the ultrasonic transmitting and receiving control unit.

[0023] Furthermore, the ultrasonic transmitting module includes a power control module, a transmitting control module, a frequency control module, a transmitting driving module and a transducer; the ultrasonic receiving module includes a signal amplifying module, a frequency selecting module, a controller, a detector and a digital-to-analog conversion module.

[0024] Furthermore, the propagation path is calculated by echo time in step 2 to obtain the oil level height. Specifically, the ultrasonic wave is emitted to the inner wall at the bottom of the oil pillow by the measuring ultrasonic probe. When the ultrasonic wave contacts the inner wall, reflected energy is generated. When the reflected echo rebounds, it contacts the second medium, thereby generating a second reflected echo. The oil level value inside the oil pillow is obtained as follows: Wherein, T is the time difference between the first reflected energy and the second reflected energy; V is the propagation speed of ultrasonic wave in the measured liquid.

[0025] Furthermore, in step 3, the interference effect of ultrasonic wave on oil level measurement due to the oil temperature of the oil pillow transformer is measured and compensated by a temperature compensation method, specifically: the law of sound velocity in transformer oil changing with temperature is: V≈1390+4(T0-20), where T0 is the oil temperature.

[0026] Furthermore, step 7 adopts the principal component analysis algorithm in the multivariate statistical analysis algorithm, and uses the correlation between process variables to establish a principal component model under normal operating conditions. By analyzing the ultrasonic parameters, dimensionality reduction processing is performed on them, ultrasonic eigenvalues ​​are extracted, and the association between these eigenvalues ​​and oil sample parameters is established.

[0027] Furthermore, the principal component analysis algorithm in step 7 specifically includes the following steps:

[0028] Step 7.1: Principal Component Analysis Model: When constructing the data matrix, arrange the ultrasonic parameter samples row by row, that is, each row represents an observation sample x i * , each column represents an observed point Using the variable standardization method, Where, L n is an n-dimensional column vector with all elements equal to 1; is an n-dimensional row vector; At this time, the normalization matrix X = [ξ1,ξ2,...,ξ m ]; To perform principal component analysis on X, we first need to obtain the covariance matrix C of X, that is, In the formula, x i is the i-th element of the standardized matrix X; n is the number of elements; the m eigenvalues ​​λ of the covariance matrix C i and the corresponding unit orthogonal vector p i , where λ1>λ1>…>λ m >0; the standardized matrix X can be decomposed into: In the formula, represents the model value matrix of X; E is the modeling error vector matrix, that is: In the formula, t i is the score vector; p i is the load vector; l is the number of principal elements; is the principal component score matrix; is the principal component loading matrix; is the residual score matrix; is the residual loading matrix;

[0029] Step 7.2: Determination of statistics: Using the established principal component model, the new sample x can be expressed as: In the formula, is the projection of sample x on the principal element subspace; e is the error vector, which is the projection of x on the residual subspace; the statistic Q is also called the squared prediction error, which is the square of the projection of sample x on the residual subspace, that is: When Q≤Q α When Q>Q α When the system fails, Q α is the confidence limit under the test level α, that is: In the formula, C α is the critical value of the normal distribution at the test level α;

[0030] (III) Beneficial effects

[0031] 1. The present invention can quickly identify the real oil level of the main transformer of the capsule oil pillow without power outage, effectively avoiding the further development of defects caused by failure to find the oil level abnormality in time, resulting in unplanned equipment shutdown and even serious power grid accidents;

[0032] 2. During the maintenance of the main transformer, the method of the present invention can also be used to detect whether there is an "air cavity" in the main transformer or the casing riser. The ultrasonic probe is installed at the main transformer oil conservator to realize real-time monitoring of the oil level, increase the online oil level detection device, and increase the redundancy of the oil level representation;

[0033] 3. The present invention adopts the calculation principle of multivariate statistics to count and analyze the ultrasonic parameters carrying transformer oil parameter information to extract its characteristic quantity, further establish the association between the transformer oil operating state and the ultrasonic characteristic quantity, and then judge the operating state of the transformer;

[0034] 4. The present invention can detect transformer oil parameters such as micro-water content, breakdown voltage (BDV), active water, etc. online and in real time, solving the problems of cumbersome detection steps and inability to detect online of traditional detection devices, improving detection sensitivity and reducing detection costs. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 It is a block diagram of the composition of the present invention.

[0036] Figure 2 It is the detection principle diagram of the present invention.

[0037] Figure 3 It is a schematic diagram of the structure of the transmitting part of the present invention.

[0038] Figure 4It is a schematic structural diagram of the receiving part of the present invention.

[0039] Figure 5 This is a schematic diagram of the multi-frequency ultrasonic spectrum detection of the internal dynamic environment of a transformer according to the present invention. DETAILED DESCRIPTION

[0040] The present invention is further described below in conjunction with embodiments and / or drawings.

[0041] Example 1

[0042] like Figure 1-5 As shown, the method for online oil level monitoring based on ultrasonic wave comprises the following steps:

[0043] Step 1: Install the ultrasonic probe of the multi-frequency ultrasonic measurement sensor on the bottom and outer wall directly below the transformer oil pillow through ultrasonic coupling agent, and ensure that the ultrasonic transmitting and receiving planes are parallel to the liquid surface of the transformer oil pillow;

[0044] In the present invention, the ultrasonic probe adopts a "strong magnetic attraction" design and is magnetically fixed to the bottom and outer wall just below the transformer oil pillow through a strong magnetic adsorption component, which is suitable for application scenarios of various types of transformers.

[0045] The ultrasonic probe is in the center of the multi-frequency ultrasonic sensor. Ring magnets are installed around the sensor. When installing the sensor, you need to first apply ultrasonic coupling agent to the sensing part in the middle of the ultrasonic sensor. The applied coupling agent should be slightly higher than the surface around the sensor. When the sensor is close to the transformer oil pillow, the ring magnet of the sensor will be directly adsorbed to the transformer oil pillow.

[0046] The multi-frequency ultrasonic measurement sensor in step 1 includes an ultrasonic probe, a digital display, a display module, a data processor, a data receiving module and an analysis module; the digital processor is connected to the display module, the data receiving module and the analysis module respectively.

[0047] The multi-frequency ultrasonic measurement sensor further comprises an ultrasonic transmitting and receiving control unit and a multi-frequency ultrasonic sensor; the ultrasonic probe is connected to the multi-frequency ultrasonic sensor; the ultrasonic transmitting and receiving control unit is composed of an ultrasonic transmitting module and an ultrasonic receiving module.

[0048] The ultrasonic transmitting and receiving control unit is equipped with a power interface, a power switch, an Internet network interface, an ultrasonic signal output interface and a signal receiving interface.

[0049] A multi-frequency ultrasonic generating device is provided in the ultrasonic transmitting and receiving control unit, and the multi-frequency ultrasonic generating device is internally connected to the ultrasonic signal output interface of the ultrasonic transmitting and receiving control unit, and the output interface is connected to the multi-frequency ultrasonic sensor via a wire, and the output signal of the multi-frequency sensor is connected to the ultrasonic signal receiving interface of the ultrasonic transmitting and receiving control unit via a wire, and the ultrasonic signal receiving interface transmits the signal to the signal processing circuit of the ultrasonic transmitting and receiving control unit.

[0050] In the present invention, a control circuit based on DSP technology can accurately complete the calculation of ultrasonic parameters such as amplitude, phase and ultrasonic velocity from the ultrasonic original signal, and then transmit the ultrasonic parameters to a host computer based on a multivariate statistical analysis algorithm through a network. In the host computer, ultrasonic characteristic quantities are extracted from the ultrasonic parameters, and parameters such as the micro-water content, breakdown voltage, active water, dielectric loss, volume resistivity, and oil interfacial tension of the transformer oil are calculated through a series of intelligent algorithms. These transformer oil parameters are compared with data in a database to judge the operating status of the transformer or detect transformer faults.

[0051] The ultrasonic transmitting module includes a power control module, a transmitting control module, a frequency control module, a transmitting driving module and a transducer;

[0052] In the present invention, the power control module selects a suitable power signal instruction according to the type of transformer and the size of the transformer capacity, and sends it to the transmission control module for processing; the transmission control module generates a carrier pulse signal required for transducer transmission according to the output signals of the frequency control module and the power control module, and sends it to the transmission drive module, the frequency control determines the carrier frequency, and the power control determines the pulse width and amplitude, thereby achieving the function of controlling the transmission power of the transducer; the transmission drive module generates a high-power output signal after boosting the pulse signal sent by the transmission control module, and drives the transducer to transmit suitable multi-frequency ultrasonic waves.

[0053] The frequency control module can change the frequency of the transmitted ultrasonic wave, and the power control module can change the power of the transmitted ultrasonic wave. In this way, the frequency range and power of the multi-frequency ultrasonic wave can be controlled by changing the output of the frequency control module and the power control module, so that the equipment can be used normally under different transformer capacities, thereby improving the applicability of the product and reducing the cost of transformer detection.

[0054] The ultrasonic receiving module includes a signal amplifying module, a frequency selecting module, a controller, a detector and a digital-to-analog conversion module.

[0055] Among them, the signal amplification module amplifies the signal transmitted back by the ultrasonic sensor, and then sends it to the frequency selection module to screen the signal frequency; the frequency selection module filters out the interference signal by setting the appropriate frequency passing range to obtain accurate multi-frequency ultrasonic signal analog quantity; the multi-frequency ultrasonic signal analog quantity is transmitted to the digital-to-analog conversion module through the detector, and through AD conversion, the multi-frequency ultrasonic signal analog quantity is converted into a digital signal and transmitted to the microprocessor.

[0056] Step 2: The ultrasonic transmitting and receiving control unit of the multi-frequency ultrasonic measuring sensor transmits 2-10MHz ultrasonic waves, which enter the transformer oil pillow through the coupling agent. When the ultrasonic waves reach the oil level surface in the oil pillow, they are reflected. The propagation path is calculated by the echo time to obtain the oil level height. (The ultrasonic waves return along the transmitting path, and the ultrasonic receiving unit can receive the reflected ultrasonic waves. The propagation speed of ultrasonic waves in the transformer oil is basically constant. As long as the time difference from the ultrasonic wave emission to the reflected wave is measured, the height of the transformer oil level in the transformer oil pillow can be calculated);

[0057] In the present invention, the ultrasonic oil level measurement principle: the characteristic of ultrasonic wave is that the higher the frequency it emits, the weaker the diffraction ability, but in this state its reflection ability will be stronger. If an obstacle is encountered during the reflection or propagation process, the reflected wave will be partially absorbed by the obstacle, and a part of the energy will be reflected back based on the reflection obstacle. The obstacle is generally called a medium. The reflection intensity generated by ultrasonic wave in different media is different. Similarly, the back reflection wave intensity generated by different media is also different. Different media will use an existing formula for calculation to check the energy loss of ultrasonic wave generated by different media, as shown in the following formula: In the formula, α is the loss coefficient; η is the shear viscosity of the medium; from this we can know that when an ultrasonic wave with a certain energy propagates in a medium with a greater viscosity, its energy loss rate will be greater.

[0058] The ultrasonic measuring probe transmits ultrasonic waves to the inner wall at the bottom of the oil pillow. When the ultrasonic waves touch the inner wall, they will generate reflected energy. When the reflected echo rebounds, it will touch the second medium and generate a second reflected echo. The oil level inside the oil pillow can be obtained as follows: Wherein, T is the time difference between the first reflected energy and the second reflected energy; V is the propagation speed of ultrasonic wave in the measured liquid.

[0059] Step 3: For the interference effect of ultrasonic wave in oil level measurement caused by oil temperature of oil pillow transformer, direct wave interference and random interference, measurement compensation is performed to correct the final result;

[0060] In the present invention, under normal conditions, the diffusion speed of ultrasonic waves in the same propagation medium should remain constant, but the temperature and state of the measured object will constantly change, and it is impossible to maintain a constant state for each measurement. Therefore, when measuring under different conditions, the speed calculation formula should be adjusted according to the temperature and state of the measured object. When the transformer oil is at 20°C, the sound speed is about 1390m / s. When the transformer oil temperature changes by 1°C at room temperature, the ultrasonic speed changes by 4m / s. It can be seen that the influence on the measurement error is about 0.28%. When the temperature error exceeds 3°C, the measurement error of the liquid level exceeds 0.74%. The law of the sound speed in the transformer oil changing with the temperature is: V≈1390+4(T0-20), wherein T0 is the oil temperature. The temperature is measured by using a temperature sensor based on a temperature measurement compensation method, the sound speed is indirectly compensated, and the final calculation result is corrected to achieve the purpose of improving the measurement accuracy.

[0061] In addition, in order to ensure that the errors and adverse effects of interference signals on measurements are minimized, multiple measurements and averaging of the measurement results can be performed to ensure that the overall data is closer to the actual state; and in order to ensure the accuracy of the measurement results and their average values, two consecutive measurements will be taken, and the results of five measurements will be averaged. The data represented by the average value will be closer to the actual data; in the process of obtaining the final data, the maximum and minimum values ​​of the five measurements must be removed, and the average value of the remaining three data must be obtained. This can effectively ensure that the error value caused by the interference data is eliminated in the average value, thereby reducing the adverse effects of errors, interference data and various interference factors on the measurement results.

[0062] Step 4: The multi-frequency ultrasonic measurement sensor measures the height of the oil level in the transformer oil conservator in real time, and transmits the measurement data to the system background via wired or wireless means, so as to understand the height of the oil level in the transformer oil conservator in real time;

[0063] Step 5: The multi-frequency ultrasonic measurement sensor uses its measurement data to compare with the readings of the traditional connected or connecting rod oil level gauge, and regularly performs online calibration of the mechanical oil level;

[0064] Step 6: Based on the online monitoring data of the transformer oil conservator oil level, environmental data and the actual working condition of the transformer are obtained while monitoring the oil level, and a corresponding relationship diagram between the transformer oil conservator oil level and the transformer environment and working condition is established to achieve early prediction and early warning of the transformer oil conservator failure;

[0065] Step 7: Based on the echo data information during the ultrasonic detection in step 2, the moisture content, breakdown voltage, dielectric loss, volume resistivity, and oil interfacial tension parameters of the transformer oil are extracted through a multivariate statistical analysis algorithm to determine the operating status of the transformer.

[0066] The invention is an online oil level monitoring method based on ultrasonic wave, which can be widely used in 110kV, 220kV, 500kV capsule and bellows transformer oil pillows, and provides safer working guarantee for transformer oil pillows. In use, the invention adopts the principle of ultrasonic reflection, and utilizes the fact that the propagation speed of ultrasonic wave in the oil in the transformer oil pillow is stable. When the ultrasonic wave reaches the oil level surface, it forms a reflection, and the ultrasonic wave emitted through the surface will reach the oil level sensor again. As long as the time difference between ultrasonic wave emission and reception is calculated, the double oil level height value can be obtained. During installation, the invention is adsorbed directly below the transformer oil pillow through an ultrasonic coupling agent, and it is necessary to ensure that the ultrasonic wave emission and receiving planes are parallel to the transformer oil pillow liquid surface. The oil pillow oil level and transformer working condition relationship diagram established based on the transformer oil pillow online oil level monitoring data can be used to effectively analyze the daily operation of the transformer oil pillow, and establish a complete transformer oil pillow fault early warning system to improve the stability of transformer operation. The invention has the advantages of quickly identifying "false oil level", increasing detection redundancy, and preventing unplanned shutdown caused by abnormal oil level.

[0067] Example 2

[0068] like Figure 1-5 As shown, the method for online oil level monitoring based on ultrasonic wave comprises the following steps:

[0069] Step 1: Install the ultrasonic probe of the multi-frequency ultrasonic measurement sensor on the bottom and outer wall directly below the transformer oil pillow through ultrasonic coupling agent, or fix it on the bottom and outer wall directly below the transformer oil pillow through strong magnetic adsorption components, and ensure that the ultrasonic emission and receiving planes are parallel to the transformer oil pillow liquid surface;

[0070] Step 2: The ultrasonic transmitting and receiving control unit of the multi-frequency ultrasonic measuring sensor transmits 2-10MHz ultrasonic waves, which enter the transformer oil pillow through the coupling agent. When the ultrasonic waves reach the oil level surface in the oil pillow, they are reflected. The propagation path is calculated by the echo time to obtain the oil level height. (The ultrasonic waves return along the transmitting path, and the ultrasonic receiving unit can receive the reflected ultrasonic waves. The propagation speed of ultrasonic waves in the transformer oil is basically constant. As long as the time difference from the ultrasonic wave emission to the reflected wave is measured, the height of the transformer oil level in the transformer oil pillow can be calculated);

[0071] Step 3: For the interference effect of ultrasonic wave in oil level measurement caused by oil temperature of oil pillow transformer, direct wave interference and random interference, measurement compensation is performed to correct the final result;

[0072] Step 4: The multi-frequency ultrasonic measurement sensor measures the height of the oil level in the transformer oil conservator in real time, and transmits the measurement data to the system background via wired or wireless means, so as to understand the height of the oil level in the transformer oil conservator in real time;

[0073] Step 5: The multi-frequency ultrasonic measurement sensor uses its measurement data to compare with the readings of the traditional connected or connecting rod oil level gauge, and regularly performs online calibration of the mechanical oil level;

[0074] Step 6: Based on the online monitoring data of the transformer oil conservator oil level, environmental data and the actual working condition of the transformer are obtained while monitoring the oil level, and a corresponding relationship diagram between the transformer oil conservator oil level and the transformer environment and working condition is established to achieve early prediction and early warning of the transformer oil conservator failure;

[0075] Step 7: Based on the echo data information during the ultrasonic detection in step 2, the moisture content, breakdown voltage, dielectric loss, volume resistivity, and oil interfacial tension parameters of the transformer oil are extracted through a multivariate statistical analysis algorithm to determine the operating status of the transformer.

[0076] In the present invention, the principal component analysis algorithm in the multivariate statistical analysis algorithm is adopted, and the principal component model under normal working conditions is established by utilizing the correlation between process variables. By analyzing the ultrasonic parameters, the dimensionality reduction processing is performed on them, the ultrasonic eigenvalues ​​are extracted, and the association between these eigenvalues ​​and the oil sample parameters is established, which specifically includes the following steps:

[0077] Step 7.1: Principal Component Analysis Model: When constructing the data matrix, arrange the ultrasonic parameter samples row by row, that is, each row represents an observation sample (physical quantity) x i * , each column represents an observed point Considering that different observed quantities often have different dimensions, and different dimensions will cause the dispersion of the values ​​of each variable to vary greatly, in order to eliminate this effect, the variable standardization method is often used. Where, L n is an n-dimensional column vector with all elements equal to 1; is an n-dimensional row vector; At this time, the normalization matrix X = [ξ1,ξ2,...,ξ m ]; To perform principal component analysis on X, we first need to obtain the covariance matrix C of X, that is, In the formula, x i is the i-th element of the standardized matrix X; n is the number of elements; the m eigenvalues ​​λ of the covariance matrix C i and the corresponding unit orthogonal vector p i , where λ1>λ1>…>λ m >0; the standardized matrix X can be decomposed into: In the formula, represents the model value matrix of X; E is the modeling error vector matrix, that is: In the formula, t iis the score vector; p i is the load vector; l is the number of principal elements; is the principal component score matrix; is the principal component loading matrix; is the residual score matrix; is the residual loading matrix;

[0078] To determine the number of principal components l, the principal component cumulative contribution rate method is usually used, that is, the variance contribution rate l is used to calculate the principal component cumulative contribution rate l. CPV Calculate l according to the following formula: In the formula, l CL is the control limit, which is generally taken as 0.85l;

[0079] is the unit orthogonal matrix, is an orthogonal matrix; X is projected into two orthogonal complementary spaces: the principal component subspace and the residual subspace; when there is a certain linear correlation between variables, the principal component subspace can use a few variables of a small number of dimensions to describe most of the information of the process changes, while the residual subspace represents the noise part;

[0080] Step 7.2: Determination of statistics: Using the established principal component model, the new sample x can be expressed as: In the formula, is the projection of sample x on the principal element subspace; e is the error vector, which is the projection of x on the residual subspace; the statistic Q is also called the squared prediction error, which is the square of the projection of sample x on the residual subspace, that is: When Q≤Q α When Q>Q α When the system fails, Q α is the confidence limit under the test level α, that is: In the formula, C α is the critical value of the normal distribution at the test level α;

[0081] The invention is an online oil level monitoring method based on ultrasonic wave, which can be widely used in 110kV, 220kV, 500kV capsule and bellows transformer oil pillows, and provides safer working guarantee for transformer oil pillows. In use, the invention uses multi-frequency ultrasonic wave technology, and transmits and receives ultrasonic waves through a multi-frequency ultrasonic wave control unit and a multi-frequency ultrasonic wave sensor. The ultrasonic wave parameters transmitted back are analyzed by a host computer based on a multivariate statistical analysis algorithm, so as to obtain parameters such as micro-water content, breakdown voltage, active water, oil dielectric loss and the like of transformer oil. Finally, these parameters are compared with a database, and the operation state of the transformer can be judged in real time. The invention can realize online detection and fault diagnosis of transformer oil, solve the problems that traditional detection devices are susceptible to interference, cannot continuously detect in real time, and have complicated detection steps, and the invention can realize the adjustment of ultrasonic wave frequency range and power size, so as to facilitate the emission of ultrasonic waves of suitable frequency and power according to the size of transformer capacity, so as to reduce the cost of transformer online detection, and more accurately judge the real-time operation state of the transformer. The invention has the advantages of quickly distinguishing "false oil level", increasing detection redundancy, and preventing unplanned shutdown caused by abnormal oil level.

Claims

1. The method for online oil level monitoring based on ultrasonic wave is characterized by: The method comprises the following steps: Step 1: Install the ultrasonic probe of the multi-frequency ultrasonic measurement sensor on the bottom and outer wall directly below the transformer oil pillow through ultrasonic coupling agent, and ensure that the ultrasonic transmitting and receiving planes are parallel to the liquid surface of the transformer oil pillow; Step 2: The ultrasonic transmitting and receiving control unit of the multi-frequency ultrasonic measuring sensor transmits 2-10MHz ultrasonic waves, which enter the transformer oil pillow through the coupling agent. When the ultrasonic waves reach the oil level surface in the oil pillow, they are reflected. The propagation path is calculated by the echo time to obtain the oil level height. Step 3: For the interference effect of ultrasonic wave in oil level measurement caused by oil temperature of oil pillow transformer, direct wave interference and random interference, measurement compensation is performed to correct the final result; Step 4: The multi-frequency ultrasonic measurement sensor measures the height of the oil level in the transformer oil conservator in real time, and transmits the measurement data to the system background via wired or wireless means, so as to understand the height of the oil level in the transformer oil conservator in real time; Step 5: The multi-frequency ultrasonic measurement sensor uses its measurement data to compare with the readings of the traditional connected or connecting rod oil level gauge, and regularly performs online calibration of the mechanical oil level; Step 6: Based on the online monitoring data of the transformer oil conservator oil level, environmental data and the actual working condition of the transformer are obtained while monitoring the oil level, and a corresponding relationship diagram between the transformer oil conservator oil level and the transformer environment and working condition is established to achieve early prediction and early warning of the transformer oil conservator failure; Step 7: Based on the echo data information during the ultrasonic detection in step 2, the moisture content, breakdown voltage, dielectric loss, volume resistivity, and oil interfacial tension parameters of the transformer oil are extracted through a multivariate statistical analysis algorithm to determine the operating status of the transformer.

2. The method for online oil level monitoring based on ultrasonic wave according to claim 1, characterized in that: The multi-frequency ultrasonic measurement sensor in step 1 includes an ultrasonic probe, a digital display, a display module, a data processor, a data receiving module and an analysis module; the digital processor is connected to the display module, the data receiving module and the analysis module respectively.

3. The method for online oil level monitoring based on ultrasonic wave according to claim 2, characterized in that: The multi-frequency ultrasonic measurement sensor further comprises an ultrasonic transmitting and receiving control unit and a multi-frequency ultrasonic sensor; the ultrasonic probe is connected to the multi-frequency ultrasonic sensor; the ultrasonic transmitting and receiving control unit is composed of an ultrasonic transmitting module and an ultrasonic receiving module.

4. The method for online oil level monitoring based on ultrasonic wave according to claim 3, characterized in that: The ultrasonic transmitting and receiving control unit is equipped with a power supply interface, a power switch, an Internet network interface, an ultrasonic signal output interface and a signal receiving interface.

5. The method for online oil level monitoring based on ultrasonic wave according to claim 4, characterized in that: A multi-frequency ultrasonic generating device is provided in the ultrasonic transmitting and receiving control unit, and the multi-frequency ultrasonic generating device is internally connected to the ultrasonic signal output interface of the ultrasonic transmitting and receiving control unit, and the output interface is connected to the multi-frequency ultrasonic sensor via a wire, and the output signal of the multi-frequency sensor is connected to the ultrasonic signal receiving interface of the ultrasonic transmitting and receiving control unit via a wire, and the ultrasonic signal receiving interface transmits the signal to the signal processing circuit of the ultrasonic transmitting and receiving control unit.

6. The method for online oil level monitoring based on ultrasonic wave according to claim 3, characterized in that: The ultrasonic transmitting module includes a power control module, a transmitting control module, a frequency control module, a transmitting driving module and a transducer; the ultrasonic receiving module includes a signal amplifying module, a frequency selecting module, a controller, a detector and a digital-to-analog conversion module.

7. The method for online oil level monitoring based on ultrasonic wave according to claim 1, characterized in that: The step 2 uses the echo time to calculate the propagation path and then obtain the oil level. Specifically, the ultrasonic wave is emitted to the inner wall at the bottom of the oil pillow by the measuring ultrasonic probe. When the ultrasonic wave contacts the inner wall, it will generate reflected energy. When the reflected echo rebounds, it will contact the second medium and generate a second reflected echo. Therefore, the oil level value inside the oil pillow is: Wherein, T is the time difference between the first reflected energy and the second reflected energy; V is the propagation speed of ultrasonic wave in the measured liquid.

8. The method for online oil level monitoring based on ultrasonic wave according to claim 1, characterized in that: In step 3, the interference effect of ultrasonic wave on oil level measurement caused by oil temperature of oil pillow transformer is measured and compensated by temperature compensation method, specifically: the law of sound velocity in transformer oil changing with temperature is: V≈1390+4(T0-20), where T0 is oil temperature.

9. The method for online oil level monitoring based on ultrasonic wave according to claim 1, characterized in that: The step 7 adopts the principal component analysis algorithm in the multivariate statistical analysis algorithm, and uses the correlation between process variables to establish a principal component model under normal operating conditions. By analyzing the ultrasonic parameters, the ultrasonic parameters are subjected to dimensionality reduction processing, the ultrasonic eigenvalues ​​are extracted, and the association between these eigenvalues ​​and the oil sample parameters is established.

10. The method for online oil level monitoring based on ultrasonic wave according to claim 9, characterized in that: The principal component analysis algorithm in step 7 specifically includes the following steps: Step 7.1: Principal Component Analysis Model: When constructing the data matrix, arrange the ultrasonic parameter samples row by row, that is, each row represents an observation sample x i * , each column represents an observed point Using the variable standardization method, Where, L n is an n-dimensional column vector with all elements equal to 1; is an n-dimensional row vector; At this time, the normalization matrix X = [ξ1,ξ2,...,ξ m ]; To perform principal component analysis on X, we first need to obtain the covariance matrix C of X, that is, In the formula, x i is the i-th element of the standardized matrix X; n is the number of elements; the m eigenvalues ​​λ of the covariance matrix C i and the corresponding unit orthogonal vector p i , where λ1>λ1>…>λ m >0; the standardized matrix X can be decomposed into: In the formula, represents the model value matrix of X; E is the modeling error vector matrix, that is: Where, t i is the score vector; p i is the load vector; l is the number of principal elements; is the principal component score matrix; is the principal component loading matrix; is the residual score matrix; is the residual loading matrix; Step 7.2: Determination of statistics: Using the established principal component model, the new sample x can be expressed as: In the formula, is the projection of sample x on the principal element subspace; e is the error vector, which is the projection of x on the residual subspace; the statistic Q is also called the squared prediction error, which is the square of the projection of sample x on the residual subspace, that is: When Q≤Q α When Q>Q α When the system fails, Q α is the confidence limit under the test level α, that is: In the formula, C α is the critical value of the normal distribution at the test level α;

Images