Method for improving sound measurement precision of transformer
Technical Field
The invention relates to a method for improving the sound measurement precision of a transformer, belonging to the sound detection technology of the transformer.
Background
The operation faults of the transformer are key reasons for causing large-area power failure of the power system, and various defects such as overload, unbalanced load, harmonic load, severe overheat, direct current magnetic bias, partial discharge, loosening of a winding iron core, loosening of accessories and the like of the transformer are related to the operation sound of the transformer according to DL/T573-2010 power transformer maintenance guidelines; the accurate measurement of the running sound of the transformer has important significance for the running state evaluation and fault diagnosis of the transformer; however, the transformer sound detection process is easily interfered by external environmental factors, including driving vehicles, bird singing, insect singing, frog singing and the like, so that the problems of difficult development of transformer sound detection, low detection efficiency, poor measurement accuracy and the like are caused, and the transformer running state evaluation and fault diagnosis process is seriously influenced.
Disclosure of Invention
The invention provides a method for improving the sound measurement precision of a transformer, and aims to solve the problem that the transformer is easy to be interfered by external sound in the sound measurement process in the prior art.
The technical solution of the invention is as follows: a method of improving the accuracy of transformer sound measurement, the method comprising the steps of:
1) Setting sampling frequency and sampling time, and detecting a transformer acoustic signal S 1;
2) Selecting a first cut-off frequency, designing a Bart Wo Sigao-pass digital Filter1, and obtaining a transformer acoustic signal S 2 after wind noise is eliminated;
3) Selecting a second cut-off frequency, designing a Butterworth low-pass digital Filter2, and obtaining a transformer frequency band acoustic signal S 3;
4) Selecting a third cut-off frequency, and designing a Butterworth low-pass Filter3 and a Butterworth Wo Sigao-pass Filter4 to respectively obtain a low-frequency band sound signal S 4 and a high-frequency band sound signal S 5;
5) Weiner filtering enhancement processing is carried out, and enhancement signals of the low-frequency band acoustic signal S 4 and the high-frequency band acoustic signal S 5 are obtained;
6) And calculating the sound pressure level of the enhanced signal and calculating the total sound pressure level of the transformer.
Preferably, the Butterfly35 pass digital Filter1 is a 6 th order Butterfly Wo Sigao pass digital Filter1, the Butterworth low pass digital Filter2 is a 6 th order Butterworth low pass digital Filter2, the Butterworth low pass Filter3 is a 6 th order Butterworth low pass Filter3, and the Butterworth Wo Sigao pass Filter4 is a 6 th order Butterworth Wo Sigao pass Filter4.
Preferably, the sampling frequency in said step 1) is greater than or equal to 48kHz, capable of covering the audible frequency band; the sampling time exceeds 2S, and detecting and acquiring a transformer sound signal S 1; the transformer sound signal S 1 refers to the sound pressure of the transformer at different times, and is a waveform with a horizontal axis representing the time and a vertical axis representing the sound pressure.
Preferably, the first cut-off frequency in the step 2) is preferably 80Hz, the direct method is adopted to design a Butterworth (Butterworth) high-pass digital Filter1 of order 6, and the transformer sound signal S 1 is sent into the Butterworth Wo Sigao-pass digital Filter1 of order 6 to obtain a transformer sound signal S 2 after wind noise is eliminated.
Preferably, in the step 3), the second cut-off frequency is preferably 2kHz, a direct method is adopted to design a Butterworth (Butterworth) low-pass digital Filter2, and the Butterworth (Butterworth) low-pass digital Filter2 is used to process the transformer acoustic signal S 2 after wind noise is eliminated, so as to obtain the transformer frequency band acoustic signal S 3.
Preferably, the third cut-off frequency in the step 4) is preferably 700Hz, and a direct method is adopted to design a Butterworth (Butterworth) low-pass Filter3 of order 6 and a Butterworth (Butterworth) high-pass Filter4 of order 6; processing the transformer frequency band acoustic signal S 3 by using a 6-order Butterworth low-pass Filter3 to obtain a low-frequency band acoustic signal S 4; the transformer band acoustic signal S 3 is processed with a6 th order bute Wo Sigao pass Filter4 to obtain a high band acoustic signal S 5.
Preferably, in the step 5), weiner filtering enhancement processing is performed to obtain enhancement signals of the low-frequency band acoustic signal S 4 and the high-frequency band acoustic signal S 5, specifically: weiner filtering enhancement processing is carried out on the low-frequency band acoustic signal S 4, and an enhancement signal of the low-frequency band acoustic signal S 4 is obtained; weiner filtering enhancement processing is performed on the high-frequency band acoustic signal S 5 to obtain an enhancement signal of the high-frequency band acoustic signal S 5.
Preferably, in the step 5), when Weiner filtering enhancement processing is performed on the low-band acoustic signal S 4 and the high-band acoustic signal S 5, frame division processing is required, the frame length is 4096, and the frame shift is 2048.
Preferably, in the step 6), the sound pressure level calculation is performed on the enhanced signal, and specifically includes:
1) Performing sound pressure level calculation on the enhanced signal of the low-frequency band sound signal S 4 to obtain a first sound pressure level L eq1;
2) And (3) performing sound pressure level calculation on the enhanced signal of the high-frequency band sound signal S 5 to obtain a second sound pressure level L eq2.
Preferably, the method for calculating the total sound level of the transformer in the step 6) is as shown in formula (2):
Where L eq is the total sound pressure level of the transformer, L eq1 is the first sound pressure level, and L eq2 is the second sound pressure level.
Preferably, the function expressions of the 6-order buttery Wo Sigao-pass digital Filter1, the 6-order butterworth low-pass digital Filter2, the 6-order butterworth low-pass Filter3 and the 6-order butterworth Wo Sigao-pass Filter4 are shown as the formula (1):
Wherein: m-1 … m-m a are all filter orders; m and m-1 … m-m b are also filter orders; m a =6 is the feedback filter order; m b =6 is the feedforward filter order; b i(i=0,1,……,mb) is the feedforward filter coefficient; a i(i=0,1,……,ma) is a feedback filter coefficient; the feedforward Filter coefficient b i and the feedback Filter coefficient a i in the function expressions of the 6-order Butterworth Wo Sigao-pass digital Filter1, the 6-order Butterworth low-pass digital Filter2, the 6-order Butterworth low-pass Filter3 and the 6-order Butterworth Wo Sigao-pass Filter4 are set according to actual needs.
The invention has the advantages that:
1) The invention sets sampling frequency and sampling time, detects the transformer sound signal, eliminates wind noise in the transformer sound signal and high-frequency noise outside the frequency of the transformer sound signal through designing a filter, acquires a main frequency band sound signal, and divides the main frequency band sound signal into a low frequency band sound signal and a high frequency band sound signal, can further eliminate external interference sound in the main frequency band range of the transformer sound through Weiner filtering enhancement processing, and calculates the total sound pressure level of the transformer through a noncoherent superposition principle;
2) The invention can effectively restrain various external interference sounds possibly occurring inside and outside the sound frequency band range of the transformer, and has the advantages of convenient use and high measurement precision.
Drawings
FIG. 1 is a schematic diagram of the basic flow of the method according to the embodiment of the invention.
FIG. 2 is a graph showing the amplitude-frequency response of a high-pass digital filter with a cut-off frequency of 80Hz according to an embodiment of the present invention.
FIG. 3 is a graph showing the amplitude-frequency response of a low-pass digital filter with a cut-off frequency of 2kHz in accordance with an embodiment of the present invention.
FIG. 4 is a graph showing the amplitude-frequency response of a low-pass digital filter with a cutoff frequency of 700Hz according to an embodiment of the present invention.
FIG. 5 is a graph showing the amplitude-frequency response of a high-pass digital filter with a cutoff frequency of 700Hz according to an embodiment of the present invention.
Detailed Description
Referring to fig. 1, the implementation steps of the method for improving the sound measurement accuracy of the transformer in this embodiment include:
1) Setting sampling frequency and sampling time, and detecting a transformer acoustic signal S 1;
2) Selecting a first cut-off frequency, designing a 6-order Baud Wo Sigao-pass digital Filter1, and obtaining a transformer acoustic signal S 2 after wind noise is eliminated;
3) Selecting a second cut-off frequency, designing a 6-order Butterworth low-pass digital Filter2, and obtaining a transformer frequency band acoustic signal S 3;
4) Selecting a third cut-off frequency, and designing a 6-order Butterworth low-pass Filter3 and a 6-order Butterworth Wo Sigao-pass Filter4 to process the transformer frequency band acoustic signal S 3 to respectively obtain a low-frequency band acoustic signal S 4 and a high-frequency band acoustic signal S 5;
5) Weiner filtering enhancement processing is respectively carried out on the low-frequency band acoustic signal S 4 and the high-frequency band acoustic signal S 5, and enhancement signals of the low-frequency band acoustic signal S 4 and enhancement signals of the high-frequency band acoustic signal S 5 are obtained;
6) And respectively carrying out sound pressure level calculation on the enhancement signals of the low-frequency band sound signal S 4 and the high-frequency band sound signal S 5, and calculating the total sound pressure level of the transformer by utilizing the incoherent superposition principle.
In this embodiment, the input signal of the 6 th-order bat Wo Sigao-pass digital Filter1 is set to x (m), the output signal of the 6 th-order bat Wo Sigao-pass digital Filter1 is set to y (m), and the function expression of the direct form 6 th-order bat Wo Sigao-pass digital Filter1 is shown in formula (1):
Wherein: m-1 … m-m a are all filter orders; m and m-1 … m-m b are also filter orders; m a =6 is the feedback filter order; m b =6 is the feedforward filter order; b i(i=0,1,……,mb) is the feedforward filter coefficient; a i(i=0,1,……,ma) is a feedback filter coefficient; the specific designs of the feedforward filter coefficient b i and the feedback filter coefficient a i are as follows:
b0=0.985292071325286;
b1=-5.91175242795171;
b2=14.7793810698793;
b3=-19.7058414265057;
b4=14.7793810698793;
b5=-5.91175242795171;
b6=0.985292071325286;
a0=1;
a1=-5.97036578104684;
a2=14.8522676796558;
a3=-19.7054087928995;
a4=14.7062781369370;
a5=-5.85357170846266;
a6=0.970800465816472。
The functional expression of the 6-order Butterworth low-pass digital Filter Filter2 is also shown in the formula (1); the specific designs of the feedforward filter coefficient b i and the feedback filter coefficient a i are as follows:
b0=5.457735721692281e-07;
b1=3.274641433015368e-06;
b2=8.186603582538421e-06;
b3=1.091547144338456e-05;
b4=8.186603582538421e-06;
b5=3.274641433015368e-06;
b6=5.457735721692281e-07;
a0=1;
a1=-5.259364609410227;
a2=11.566129391042576;
a3=-13.610057554787442;
a4=9.035975203089848;
a5=-3.208665800580283;
a6=0.476018300154148。
The functional expression of the 6 th order butterworth low-pass Filter3 is also shown in a formula (1); the specific designs of the feedforward filter coefficient b i and the feedback filter coefficient a i are as follows:
b0=1.256801246407591e-09;
b1=7.540807478445544e-09;
b2=1.885201869611386e-08;
b3=2.513602492815181e-08;
b4=1.885201869611386e-08;
b5=7.540807478445544e-09;
b6=1.256801246407591e-09;
a0=1;
a1=-5.740709993005902;
a2=13.736951993835140;
a3=-17.537999982678507;
a4=12.599532988608864;
a5=-4.829319169164712;
a6=0.771544242840395。
The functional expression of the 6-order Butt Wo Sigao-pass Filter4 is also shown in formula (1); the specific designs of the feedforward filter coefficient b i and the feedback filter coefficient a i are as follows:
b0=0.878375912033339;
b1=-5.270255472200034;
b2=13.175638680500086;
b3=-17.567518240666782;
b4=13.175638680500086;
b5=-5.270255472200034;
b6=0.878375912033339;
a0=1;
a1=-5.740709993005909;
a2=13.736951993835167;
a3=-17.537999982678564;
a4=12.599532988608921;
a5=-4.829319169164739;
a6=0.771544242840400。
In this embodiment, the transformer band acoustic signal S 3 is an acoustic signal greater than 80Hz and less than or equal to 2 kHz; the low-frequency band sound signal S 4 is a sound signal which is more than 80Hz and less than or equal to 700 Hz; the high-frequency band acoustic signal S 5 is an acoustic signal of 2kHz or less and greater than 700 Hz.
In this embodiment, the sampling frequency in step 1) is 65536Hz, which can cover the audible frequency band, and the sampling time is 2S, so as to obtain the transformer acoustic signal S 1.
In this embodiment, in order to avoid wind noise interference in the transformer sound detection process, step 2) adopts a direct method to design a 6 th order Butterworth (Butterworth) high-pass digital Filter1 with a cut-off frequency of 80Hz, as shown in fig. 2; the method comprises the steps of processing a transformer acoustic signal S 1 by using a 6-order Butterworth high-pass digital Filter1 with the frequency of 80Hz according to a formula (1) to obtain a transformer acoustic signal S 2 after wind noise elimination:
Wherein: m-1 … m-m a are all filter orders; m and m-1 … m-m b are also filter orders; m a =6 is the feedback filter order; m b =6 is the feedforward filter order; b i(i=0,1,……,mb) is the feedforward filter coefficient; a i(i=0,1,……,ma) is the feedback filter coefficient.
In this embodiment, since the sound of the transformer body and the cooling fan thereof are quasi-steady-state sounds, the frequency spectrum is generally concentrated in the range of 2 kHz; therefore, in the step 3), the cut-off frequency is 2kHz, and a direct method is adopted to design a 6-order Butterworth low-pass digital Filter2, as shown in figure 3; and processing the transformer acoustic signal S 2 after wind noise elimination according to the mode of a formula (1) by using a 6-order Butterworth low-pass digital Filter2 to obtain a transformer frequency band acoustic signal S 3.
In this embodiment, since the frequency spectrum of the sound signal of the transformer body is mainly located at the frequency of integral multiple of 50Hz below 700Hz, and the cooling fan noise is distributed more uniformly in the frequency band range within 2kHz, it is considered that the external interference frequency band may occur in the 2kHz range; therefore, in order to eliminate external interference as much as possible, it is necessary to divide the frequency of the signal within 2kHz into a low-frequency band acoustic signal S 4 and a high-frequency band acoustic signal S 5; in step 4), a direct method is adopted to design a 6 th order Butterworth low-pass Filter3 (as shown in fig. 4) and a 6 th order Butterworth high-pass Filter4 (as shown in fig. 5) with the cut-off frequency of 700Hz, and the transformer band acoustic signals S 3 are processed by the 6 th order Butterworth low-pass Filter3 and the 6 th order Butterworth high-pass Filter4 according to the formula (1) to obtain the low-band acoustic signals S 4 and the high-band acoustic signals S 5.
In this embodiment, first, in step 5), the low-frequency band acoustic signal S 4 and the high-frequency band acoustic signal S 5 are subjected to frame division, the frame length is 4096, the frame is 2048, and then the low-frequency band acoustic signal S 4 and the high-frequency band acoustic signal S 5 are respectively processed by using a Weiner filtering enhancement method, so as to obtain the low-frequency band acoustic signal enhancement signal S 6 and the high-frequency band acoustic signal enhancement signal S 7 respectively.
In this embodiment, the sound pressure level calculation is performed on the low-frequency band sound signal enhancement signal S 6 and the high-frequency band sound signal enhancement signal S 7, so as to obtain a first sound pressure level L eq1 and a second sound pressure level L eq2 respectively; the total sound pressure level L eq of the transformer is calculated according to the formula (2) by utilizing the incoherent superposition principle:
the invention comprehensively considers the sound frequency band range of the transformer and the frequency band where external noise possibly appears, effectively inhibits various external interference sounds possibly appearing inside and outside the sound frequency band range of the transformer through a plurality of column filter combinations and Weiner filtering enhancement methods, and has the advantages of convenient use and high measurement precision.