阅读说明：本技术 基于近场宽带波束形成的大型电抗器故障检测方法及系统 (Large-scale reactor fault detection method and system based on near-field broadband beam forming ) 是由 叶中付 杨会超 王鹏宇 于 2020-04-23 设计创作，主要内容包括：本发明公开了一种基于近场宽带波束形成的大型电抗器故障检测方法及系统,通过比较电抗器的声音强度或者能量图像与其正常工作时的声音强度或者能量图像,判断电抗器设备是否存在故障需要工作人员维护,进而实现了电抗器设备工作状态的在线检测。通过近场宽带波束形成的声音强度或者能量图像来判断电抗器工作状态,是一种在线检测方案,不仅降低了维护的成本,而且还提高了人们用电的质量,给大型电抗器设备的工作状态检测提供了一种新的方法。(The invention discloses a method and a system for detecting faults of a large reactor based on near-field broadband beam forming, wherein whether the reactor equipment has faults or not is judged by comparing the sound intensity or the energy image of the reactor with the sound intensity or the energy image of the reactor during normal work, and the maintenance of workers is required, so that the on-line detection of the working state of the reactor equipment is realized. The working state of the reactor is judged through the sound intensity or energy image formed by the near-field broadband wave beams, the method is an online detection scheme, the maintenance cost is reduced, the electricity utilization quality of people is improved, and a new method is provided for detecting the working state of large-scale reactor equipment.)

1. A large-scale reactor fault detection method based on near-field broadband beam forming is characterized by comprising the following steps:

a microphone array is arranged outside the reactor;

scanning a reactor in work by using a microphone array and adopting a near-field broadband beam forming method, combining a received signal with a pre-designed weighting coefficient, and calculating a sound intensity image or a sound energy image of the reactor;

and comparing the sound intensity image or the sound energy image under the preset normal working condition with the sound intensity image or the sound energy image during detection, judging whether the reactor has a fault according to the comparison result, and sending out an early warning when the fault is judged.

2. The method for detecting the fault of the large reactor based on the near-field broadband beam forming as claimed in claim 1, wherein the step of arranging a microphone array outside the reactor comprises:

the distance between the plane of the microphone array and the surface of the reactor is R, and M × N microphones are spaced by d in the X-axis direction₁The spacing d in the Y-axis direction₂Standing to finally form [ (M-1) × d₁]×[(N-1)×d₂]And (3) a uniform area array with the size, wherein an array element with coordinates of (0, 0) is used as a reference array element.

3. The method for detecting the fault of the large reactor based on the near-field broadband beam forming of claim 2 is characterized in that when a microphone array is used for scanning the reactor in operation by adopting the near-field broadband beam forming method, a near-field broadband receiving signal model is established firstly: taking a coordinate origin array element of the microphone array as a reference array element, and taking an X axis and a Y axis as reference lines of the array; suppose there are K signal sourcesAt two-dimensional angles respectivelyIncident on a microphone array, whereinθ_kRespectively representing an incident azimuth angle and an incident pitch angle of a kth signal source, wherein t is an incident moment; let the coordinate of the m-th row and n-th column array element be (x)_m,y_n) The distance d of the k-th signal incident on the array element relative to the n-th array element of the m-th row_mnkExpressed as:wherein r is_ks＝R/cosθ_kDenotes the distance, r, from the kth signal source to the reference array element_mnThe distance from the mth row and the nth column array element to the reference array element is represented, wherein M is 1, 2.

Dividing a microphone array receiving signal in a statistical time period T into L subsections by taking the time length T as a statistical time period, wherein each subsection comprises I time sampling points, the overlapping rate between adjacent subsections is gamma, and the gamma is more than or equal to 0 and less than 1;

then, performing I-point discrete fourier transform on each sub-segment broadband signal received by the microphone array to obtain I sub-band narrowband array signals, and for the ith sub-band, representing the steering vector of the kth signal source as:

wherein the content of the first and second substances,c represents the sound velocity magnitude; 1,2, 1, f_iRepresents the frequency of the ith sub-band;

the discrete fourier transform of the I point is performed on the array received signal of each sub-segment as:

X_l(f_i)＝A_l(f_i)S_l(f_i)+N_l(f_i),i＝1,2,...,I，l＝1,2,...,L

wherein the content of the first and second substances,the first sub-segment representing the received signal of the array is passed through point IThe i subband narrow band array signal obtained by discrete Fourier transform, wherein subscript number of x is the serial number of the microphone;the method comprises the steps that the ith sub-band narrowband signal obtained by I-point discrete Fourier transform of the ith sub-band of a source signal is represented, and subscript numbers of s are serial numbers of a signal source;the first subband narrow-band array noise is obtained by I-point discrete Fourier transform of the first subband array received noise, and the subscript number of n is the serial number of a microphone; a (f)_i)＝[a₁(f_i,r_1s),a₂(f_i,r_2s),…,a_K(f_i,r_Ks)]And a guide vector matrix of the ith subband signal of the ith subsection of the array, wherein subscript numbers of a are serial numbers of signal sources.

4. The method for detecting the fault of the large reactor based on the near-field broadband beam forming as claimed in claim 1 or 3, wherein the mode of combining the received signals with the pre-designed weighting coefficients and calculating the sound intensity image or the sound energy image of the reactor comprises the following steps:

within a statistical time period T, for the array receiving signal of the ith sub-band, the frequency f of the ith sub-band_iFor each amplitude weighting performed in the X-axis and Y-axis, the amplitude weighting is performed by different types of window functions, and the weighting matrix is:whereinIs the weighted value of the window function in the X-axis direction in the ith sub-frequency,is as followsThe weight of the window function in the Y-axis direction in the i sub-frequencies,andeach term in (a) represents a weighted value of a window function of an amplitude; at the same time, orderWherein vec represents a matrix or vector straightening operation;

for the array receiving signal of the ith subsegment, the frequency f of the ith sub-band_iAbove, the formula for calculating the sound intensity image is:where | represents the modulo of a complex number,to representThe conjugate transpose of (1); sound intensity image q in (x, y) coordinates within a statistical period of time T_(x,y)The sum of the sound intensity of each sub-band frequency of each sub-segment is expressed asIn a statistical time period T, the array scans the complete angle area of the reactorThe obtained reactor sound intensity image is:

in a statistical periodImage of sound energy in (x, y) coordinates within TThe sum of the sound intensity of each sub-band frequency of each sub-segment is expressed asIn a statistical time period T, the array scans the complete angle area of the reactorThe obtained reactor sound energy image is:

wherein, subscripts X and Y are complete angle areasThe middle microphone array acquires the quantity of all signals in an X axis and a Y axis;

based on the above principle, azimuth angle by step scanningAnd acquiring a sound intensity image or a sound energy image of the reactor in real time in a pitch angle delta theta mode.

5. The method for detecting the fault of the large reactor based on the near-field broadband beam forming of claim 1, wherein the comparing presets a sound intensity image or a sound energy image under a normal working condition and during detection, judges whether the reactor has a fault according to a comparison result, and sends out an early warning when the fault is judged to occur, and comprises the following steps:

according to the priori knowledge, the sound intensity image in the statistical time period T under the normal working condition of the reactor can be obtained in advance and recorded asCalculating to obtain a sound intensity image of the reactor in the statistical time period TComparing the absolute value of the difference between the two complete angle regions:where | | represents the norm of the matrix ifThen the angle area is consideredWhen a fault occurs, an early warning is sent out (for example, the early warning is carried out on the staff in the modes of light, sound and the like); wherein the content of the first and second substances,for a preset full angle regionSound intensity image ofA difference threshold of (c);

or, according to the priori knowledge, the sound energy image in the statistical time period T under the normal working condition of the reactor can be obtained in advance and recorded asCalculating to obtain a sound intensity image of the reactor in the statistical time period TComparing the difference between the two complete angle regionsFor the value:if it isThen the angle area is consideredWhen a fault occurs, an early warning is sent out; wherein the content of the first and second substances,for a preset full angle regionSound energy image of andis detected.

6. A large reactor fault detection system based on near-field broadband beam forming, which is used for realizing the method of any one of claims 1-5 and comprises the following steps:

a microphone array disposed outside the reactor;

the sampling processing device scans a reactor in work by using the microphone array by adopting a near-field broadband beam forming method and acquires a receiving signal of the microphone array;

the detection processing device is used for combining the received signals with a pre-designed weighting coefficient and calculating a sound intensity image or a sound energy image of the reactor; comparing the calculated sound intensity image or sound energy image of the reactor with the sound intensity image or sound energy image which is obtained in advance and corresponds to the sound intensity image or sound energy image under the condition that the reactor normally works, and judging whether the reactor has a fault according to the comparison result;

and the abnormal state processing device is used for sending out early warning when the fault is judged to occur.

Technical Field

The invention relates to the field of large-scale reactor fault detection, in particular to a large-scale reactor fault detection method and system based on near-field broadband beam forming.

Background

At present, the power demand is continuously promoted, the power consumption quality problem is also concerned widely, and the stable supply of electric energy and the national economy have an inseparable relationship. A large number of practices prove that the electric reactor always has certain latent faults before serious accidents occur. Latent faults such as partial discharge, local overheating, winding deformation, loosening of mechanical parts, and deterioration of equipment insulation inside the reactor are caused by accumulation over time. When the reactor runs, due to mutual movement between the machine body and the firmware, the parts or between the parts, the equipment can make a sound, and when the running state changes, the sound made by the equipment also changes. Meanwhile, the winding and the iron core in the reactor have important functions of electromagnetic exchange, and different faults can occur in the high-voltage and strong-electromagnetic environment, so that the running sound can be changed accordingly.

At present, the protection method of the electric reactor mainly carries out relay protection through electric parameters such as voltage, current and the like during fault, and the method can reduce the power consumption quality and improve the economic cost of a power supply company. And the related latent fault is difficult to detect, and an effective online detection method, technology and device are lacked.

Disclosure of Invention

The invention aims to provide a method and a system for detecting faults of a large reactor based on near-field broadband beam forming, which can detect latent faults and give early warning to related personnel in advance to process the condition of equipment, thereby improving the electricity utilization quality of people and reducing the economic cost.

The purpose of the invention is realized by the following technical scheme:

a large-scale reactor fault detection method based on near-field broadband beam forming comprises the following steps:

a microphone array is arranged outside the reactor;

scanning a reactor in work by using a microphone array and adopting a near-field broadband beam forming method, combining a received signal with a pre-designed weighting coefficient, and calculating a sound intensity image or a sound energy image of the reactor;

and comparing the sound intensity image or the sound energy image under the preset normal working condition with the sound intensity image or the sound energy image during detection, judging whether the reactor has a fault according to the comparison result, and sending out an early warning when the fault is judged.

A large-scale reactor fault detection system based on near-field broadband beam forming is used for the method and comprises the following steps:

a microphone array disposed outside the reactor;

the sampling processing device scans a reactor in work by using the microphone array by adopting a near-field broadband beam forming method and acquires a receiving signal of the microphone array;

the detection processing device is used for combining the received signals with a pre-designed weighting coefficient and calculating a sound intensity image or a sound energy image of the reactor; comparing the calculated sound intensity image or sound energy image of the reactor with the sound intensity image or sound energy image which is obtained in advance and corresponds to the sound intensity image or sound energy image under the condition that the reactor normally works, and judging whether the reactor has a fault according to the comparison result;

and the abnormal state processing device is used for sending out early warning when the fault is judged to occur.

According to the technical scheme provided by the invention, the sound intensity or the energy image of the reactor is compared with the sound intensity or the energy image of the reactor in normal working, whether the reactor equipment has a fault or not is judged, and the maintenance of a worker is required, so that the online detection of the working state of the reactor equipment is realized. The working state of the reactor is judged through the sound intensity or energy image formed by the near-field broadband wave beams, the method is an online detection scheme, the maintenance cost is reduced, the electricity utilization quality of people is improved, and a new method is provided for detecting the working state of large-scale reactor equipment.

Drawings

In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings needed to be used in the description of the embodiments are briefly introduced below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art to obtain other drawings based on the drawings without creative efforts.

Fig. 1 is a flowchart of a method for detecting a fault of a large reactor based on near-field broadband beam forming according to an embodiment of the present invention;

fig. 2 is a schematic diagram of a signal model of a planar microphone array according to an embodiment of the present invention;

fig. 3 is a schematic diagram of a system configuration for implementing the method according to the embodiment of the present invention;

the parts corresponding to each mark in the figure are: 1-a microphone array; 2-a reactor; 3-a sampling processing device; 4-detection processing means; 5-abnormal state processing means; 6-target; 7-a support for the microphone array; 8-microphone elements; 9-reference array element.

Detailed Description

The technical solutions in the embodiments of the present invention are clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments of the present invention without making any creative effort, shall fall within the protection scope of the present invention.

The embodiment of the invention provides a method for detecting a fault of a large reactor based on near-field broadband beam forming, which mainly comprises the following steps of:

and step 1, arranging a microphone array outside the reactor.

In the embodiment of the invention, the microphone array is parallel to the surface of the reactor, the distance between the plane of the microphone array and the surface of the reactor is R, and M × N microphones are spaced at an interval d in the X-axis direction₁(horizontal direction), Y-axis direction (vertical direction) by an interval d₂Standing to finally form [ (M-1) × d₁]×[(N-1)×d₂]And (3) a uniform area array with the size, wherein an array element with coordinates of (0, 0) is used as a reference array element.

And 2, scanning the reactor in work by using a microphone array and adopting a near-field broadband beam forming method, combining a received signal with a pre-designed weighting coefficient, and calculating a sound intensity image or a sound energy image of the reactor.

In the embodiment of the invention, a near-field broadband beam forming method is adopted to scan the reactor in work, so that the scanning plane is scanned in all directions; firstly, establishing a near-field broadband receiving signal model: as shown in fig. 2, the coordinate origin array element of the microphone array is used as a reference array element, and the X axis and the Y axis are used as reference lines of the array; suppose there are K signal sourcesAt two-dimensional angles respectivelyIncident on a microphone array, whereinθ_kRespectively representing an incident azimuth angle and an incident pitch angle of a kth signal source, wherein t is an incident moment; let the coordinate of the m-th row and n-th column array element be (x)_m,y_n) The distance d of the k-th signal incident on the array element relative to the n-th array element of the m-th row_mnkExpressed as:

wherein r is_ks＝R/cosθ_kRepresenting the distance from the kth signal source to the reference array element, R representing the distance between the microphone array plane and the reactor surface, R_mnThe distance from the mth row and the nth column array element to the reference array element is represented, wherein M is 1, 2.

Dividing a microphone array receiving signal in the statistical time period T into L subsections, wherein each subsection comprises I time sampling points, the overlapping rate between adjacent subsections is gamma, and gamma is more than or equal to 0 and less than 1;

and then, performing I-point discrete Fourier transform on each subsection broadband signal received by the microphone array to obtain I subband narrowband array signals, wherein I is not necessarily an integer power of 2, and if I is the integer power of 2, the operation time of the process can be saved by using a fast algorithm. For the ith subband, the steering vector for the kth signal source is represented as

Wherein the content of the first and second substances,c represents the sound velocity magnitude; 1,2, 1, f_iRepresents the frequency of the ith sub-band;

the discrete fourier transform of the I point is performed on the array received signal of each sub-segment as:

X_l(f_i)＝A_l(f_i)S_l(f_i)+N_l(f_i),i＝1,2,...,I，l＝1,2,...,L

wherein the content of the first and second substances,i-th subband narrowband array signals obtained by performing I-point discrete fourier transform on the ith subband of the array received signal, wherein the index number of x is the serial number of the microphone, for example,the first sub-band signal is an ith sub-band narrow-band signal obtained by I-point discrete Fourier transform of the first sub-band of the 1 st microphone receiving signal;i-th subband narrowband signals obtained by I-point discrete fourier transform of the I-th subband of the signal source are represented, and the subscript number of s is the serial number of the signal source, for example,obtaining an ith sub-band narrowband signal for the ith sub-band of the 1 st signal source through I-point discrete Fourier transform;the ith subband narrowband array noise obtained by I-point discrete fourier transform of the ith subband array received noise is shown, and the subscript number of n is the serial number of the microphone, for example,the method comprises the steps that 1, the ith sub-band narrow-band array noise obtained by I-point discrete Fourier transform of the 1 st microphone receiving noise is represented; a (f)_i)＝[a₁(f_i,r_1s),a₂(f_i,r_2s),…,a_K(f_i,r_Ks)]A matrix of steering vectors for the ith subband signal of the ith subband of the array, based on the assumption that the location of the source does not substantially change within a statistical time period T, the index number of a being the serial number of the source, e.g., a₁(f_i,r_1s) The vector matrix of the signal source of the ith sub-band and 1 st signal source.

The embodiment of the invention designs the following weighting coefficients, and calculates the sound intensity image or the sound energy image according to the weighting coefficients.

Within a statistical time period T, for the array receiving signal of the ith sub-band, the frequency f of the ith sub-band_iFor near-field broadband beamforming, amplitude weighting is performed from left to right and from top to bottom in both the X-axis and Y-axis directions (or from left to right and from top to bottom in the azimuthal-elevation-angle two-bit direction); for each amplitude weighting performed in the X-axis and the Y-axis, the amplitude weighting is performed by different types of window functions, and the weighting matrix is:whereinIs the weighted value of the window function in the X-axis direction in the ith sub-frequency,is the weighted value of the window function in the Y-axis direction in the ith sub-frequency,andeach term in (a) represents a weighted value of a window function of an amplitude; at the same time, orderWhere vec represents the operation of straightening of a matrix or vector.

For the array receiving signal of the ith subsegment, the frequency f of the ith sub-band_iAbove, the formula for calculating the sound intensity image is:where | represents the modulo of a complex number,to representThe conjugate transpose of (1); sound intensity image q in (x, y) coordinates within a statistical period of time T_(x,y)The sum of the sound intensity of each sub-band frequency of each sub-segment is expressed asIn a statistical time period T, the array scans the complete angle area of the reactorThe obtained reactor sound intensity image is:

sound energy image in (x, y) coordinates within a statistical time period TThe sum of the sound intensity of each sub-band frequency of each sub-segment is expressed asIn a statistical time period T, the array scans the complete angle area of the reactorThe obtained reactor sound energy image is:

based on the above principle, azimuth angle by step scanningAcquiring a sound intensity image or a sound energy image of the reactor in real time in a pitch angle delta theta mode; wherein, subscripts X and Y are complete angle areasThe number of all signals acquired by the medium microphone array in the X axis and the Y axis, the size of X and Y and step selectionRelating to delta theta, i.e. to the full angular areaThere is a relationship.

And 3, comparing the sound intensity image or the sound energy image under the preset normal working condition with the sound intensity image or the sound energy image during detection, judging whether the reactor has a fault according to the comparison result, and sending out an early warning when the fault is judged.

According to the priori knowledge, the sound intensity image in the statistical time period T under the normal working condition of the reactor can be obtained in advance and recorded asCalculation systemThe sound intensity image of the reactor during the counting period T is recorded asComparing the absolute value of the difference between the two complete angle regions:where | | represents the norm of the matrix ifThen the angle area is consideredWhen a fault occurs, an early warning is sent out (for example, the early warning is carried out on the staff in the modes of light, sound and the like); wherein the content of the first and second substances,for a preset full angle regionSound intensity image ofIs detected.

Or, according to the priori knowledge, the sound energy image in the statistical time period T under the normal working condition of the reactor can be obtained in advance and recorded asCalculating to obtain a sound intensity image of the reactor in the statistical time period TComparing the absolute value of the difference between the two complete angle regions:if it isThen the angle area is consideredWhen a fault occurs, an early warning is sent out; wherein the content of the first and second substances,for a preset full angle regionSound energy image of andis detected.

Another embodiment of the present invention further provides a system for detecting a fault of a large reactor based on near-field broadband beam forming, which is used to implement the foregoing method, and as shown in fig. 3, the system mainly includes:

the microphone array 1 comprises M × N array elements 8 (reference array elements 9), a support of the microphone array is 7, the microphone array is arranged outside the reactor 2, and the microphone array receives a sound signal 6 emitted by the reactor 2;

the sampling processing device 3 scans a reactor in work by using a microphone array and adopting a near-field broadband beam forming method, and obtains a receiving signal of the microphone array; specifically, the method comprises the following steps: collecting a sound signal 6 emitted by the reactor 2 by using the microphone array 1; in a statistical time period T, sound signals 6 of the reactor 2 collected by the microphone array 1 are converted into sub-band signals of all the subsections through subsection and discrete Fourier transform;

a detection processing device 4 for combining the received signal with a pre-designed weighting coefficient and calculating a sound intensity image or a sound energy image of the reactor; specifically, for each sub-band signal of each sub-segment, each group of angles of the complete angle area of the array scanning reactor is subjected to near-field broadband beam forming by using a pre-designed weighting vector; calculating a sound intensity image or a sound energy image of the reactor 2 within a statistical period; comparing the calculated sound intensity image or sound energy image of the reactor with the sound intensity image or sound energy image which is obtained in advance and corresponds to the sound intensity image or sound energy image under the normal working condition of the reactor 2, and judging whether the reactor 2 has a fault according to the comparison result;

and the abnormal state processing device 5 is used for sending out early warning when the fault is judged to occur.

The specific technical details of the system have been described in detail in the previous embodiments, and thus are not described again.

It will be clear to those skilled in the art that, for convenience and simplicity of description, the foregoing division of the functional modules is merely used as an example, and in practical applications, the above function distribution may be performed by different functional modules according to needs, that is, the internal structure of the system is divided into different functional modules to perform all or part of the above described functions.

Through the above description of the embodiments, it is clear to those skilled in the art that the above embodiments can be implemented by software, and can also be implemented by software plus a necessary general hardware platform. With this understanding, the technical solutions of the embodiments can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a usb disk, a removable hard disk, etc.), and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods according to the embodiments of the present invention.

The above description is only for the preferred embodiment of the present invention, but the scope of the present invention is not limited thereto, and any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope of the present invention are included in the scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.