CN109991519B - Partial discharge direction-finding method and system based on neural network and wireless sensor array - Google Patents
Partial discharge direction-finding method and system based on neural network and wireless sensor array Download PDFInfo
- Publication number
- CN109991519B CN109991519B CN201910175714.9A CN201910175714A CN109991519B CN 109991519 B CN109991519 B CN 109991519B CN 201910175714 A CN201910175714 A CN 201910175714A CN 109991519 B CN109991519 B CN 109991519B
- Authority
- CN
- China
- Prior art keywords
- wireless
- partial discharge
- data
- neural network
- directional
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01R—MEASURING ELECTRIC VARIABLES; MEASURING MAGNETIC VARIABLES
- G01R31/00—Arrangements for testing electric properties; Arrangements for locating electric faults; Arrangements for electrical testing characterised by what is being tested not provided for elsewhere
- G01R31/12—Testing dielectric strength or breakdown voltage ; Testing or monitoring effectiveness or level of insulation, e.g. of a cable or of an apparatus, for example using partial discharge measurements; Electrostatic testing
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S3/00—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received
- G01S3/02—Direction-finders for determining the direction from which infrasonic, sonic, ultrasonic, or electromagnetic waves, or particle emission, not having a directional significance, are being received using radio waves
- G01S3/14—Systems for determining direction or deviation from predetermined direction
Abstract
The invention discloses a partial discharge direction-finding method based on a neural network and a wireless sensing array, wherein the wireless sensing array comprises a plurality of wireless directional sensors, and the method comprises the following steps: acquiring amplitude data of a plurality of groups of partial discharge signals within a period of time through a wireless sensor array, wherein each group of data corresponds to a corresponding time point respectively, and each data in each group of data corresponds to a corresponding wireless directional sensor respectively; screening amplitude data of a plurality of groups of partial discharge signals based on a neural network, and leaving a data group with high reliability; the direction of the partial discharge source is determined based on the screened data set. The method can realize accurate positioning of the partial discharge source in the partial discharge direction finding, and has the advantages of low equipment cost, small equipment volume, good equipment portability, good environmental adaptability, high direction finding precision and accuracy. In addition, the invention also discloses a corresponding direction-finding system.
Description
Technical Field
The invention relates to a partial discharge direction-finding method and a partial discharge direction-finding system in the field of operation and maintenance of power transmission and transformation equipment of a power system, in particular to a partial discharge direction-finding method and a partial discharge direction-finding system based on a neural network and a wireless sensor array.
Background
The phenomenon that a Discharge occurs only in a local region of an insulator, but does not penetrate between conductors to which a voltage is applied, may occur in the vicinity of the conductors or elsewhere, and is called a Partial Discharge (PD). Partial discharge is a main cause of insulation failure of electrical equipment, particularly high-voltage electrical equipment, and strong partial discharge causes rapid reduction of insulation strength, which is an important factor causing insulation damage of high-voltage electrical equipment. Therefore, it is necessary to monitor the condition of an operating electrical device, particularly a high-voltage power device, based on partial discharge information.
Accurate positioning of the partial discharge source can provide important information for condition monitoring and maintenance of electrical equipment, particularly high voltage power equipment. The ultrahigh frequency detection technology is widely applied at home and abroad due to the advantages of high sensitivity, strong anti-interference performance and the like. In the total-station ultrahigh-frequency monitoring of partial discharge, a Time of Arrival (TOA) or Time Difference of Arrival (TDOA) based on a Time delay sequence is frequently used as a method for locating a partial discharge source. The method needs to keep nanosecond time synchronization among the sensors and at least a sampling rate of a few GSa/s, so that the hardware cost is high, the equipment volume is large, and the portability is poor.
In recent years, research and application of a local discharge source positioning technology based on Received Signal Strength at ultrahigh frequency (RSSI) are carried out at home and abroad, and compared with positioning methods such as TOA and TDOA, the RSSI has the characteristics of lower equipment cost, better environmental adaptability and the like. Therefore, the invention provides a partial discharge direction-finding method and system based on a neural network and a wireless sensor array, aiming at carrying out the direction determination of a partial discharge source by combining the RSSI technology, the neural network technology and the wireless sensor array technology, thereby realizing the accurate positioning of the partial discharge source in the partial discharge direction-finding direction, and simultaneously having lower equipment cost, smaller equipment volume, better equipment portability, better environmental adaptability and higher direction-finding precision and accuracy.
Disclosure of Invention
One of the purposes of the invention is to provide a partial discharge direction finding method based on a neural network and a wireless sensor array, which can realize accurate positioning of a partial discharge source in the partial discharge direction finding, and has the advantages of lower equipment cost, smaller equipment volume, better equipment portability, better environmental adaptability, higher direction finding precision and higher accuracy.
According to the above object, the present invention provides a partial discharge direction finding method based on a neural network and a wireless sensor array, which determines the direction of a partial discharge source, wherein: the wireless sensing array comprises a plurality of wireless directional sensors, each wireless directional sensor at least has a first specific receiving direction, and the received signal strength is strongest when the first specific receiving direction is over against the local discharge source; the first specific receiving directions of the wireless directional sensors in the wireless sensing array point to different directions respectively to receive partial discharge signals; the method comprises the following steps:
s100: acquiring amplitude data of a plurality of groups of partial discharge signals within a period of time through the wireless sensor array, wherein each group of data corresponds to a corresponding time point respectively, and each data in each group of data corresponds to a corresponding wireless directional sensor respectively;
s200: constructing a neural network, screening the amplitude data of the plurality of groups of partial discharge signals based on the neural network, and leaving a data group with high reliability;
s300: the direction of the partial discharge source is determined based on the data set filtered by step S200.
The invention provides a partial discharge direction-finding method based on a neural network and a wireless sensor array, which adopts a wireless directional sensor to construct the wireless sensor array, wherein first specific receiving directions of the wireless directional sensor in the wireless sensor array point to different directions, so that the same partial discharge signal is received in different directions and converted into amplitude data with different intensities, and the direction of a partial discharge source is determined according to the magnitude of the amplitude data and the pointing direction of the first specific receiving direction of the corresponding wireless directional sensor. Meanwhile, the invention also combines the neural network technology to screen the amplitude data, thereby correcting errors and improving the accuracy.
Typically, the wireless sensor array is arranged in a uniform circular array, that is, the wireless directional sensors are uniformly distributed on a circumference, and the first specific receiving direction of each wireless directional sensor points to the outside of the circumference along a radius of the circumference. The more the number of distributed wireless orientation sensors is, the more the number of parts is divided for 360-degree azimuth, and the higher the precision is. The wireless directional sensor is typically a very high frequency sensor, effectively detecting partial discharge signals.
Further, in the partial discharge direction finding method based on the neural network and the wireless sensor array, the wireless directional sensors in the wireless sensor array are circumferentially and uniformly arranged on a circumference to form a circular array, wherein the first specific receiving direction of each wireless directional sensor points to the outside of the circular array along a circumference radius.
The scheme is a conventional scheme and is convenient to implement and calculate.
Further, in the partial discharge direction finding method based on the neural network and the wireless sensor array, in step S300, a maximum value in each group of data is respectively searched, a direction pointed by a first specific receiving direction of the wireless directional sensor corresponding to the maximum value is a directional result corresponding to the group of data, and then an average value of all directional results is a direction of the partial discharge source measured within the period of time.
In the above scheme, the manner of measuring within a period of time is generally to continuously perform the acquisition of partial discharge signals within the period of time, then arrange and convert the acquired signals into corresponding amplitude data in groups according to a time sequence, and then perform the above steps on the grouped amplitude data. The principle of finding the maximum value of the amplitude data to determine the directional result is that the received signal strength is strongest when the first specific receiving direction is over against the local discharge source, so that the direction pointed by the first specific receiving direction of the corresponding wireless directional sensor is the direction of the local discharge source relative to the wireless sensor array, namely the directional result corresponding to the group of data.
Further, in the partial discharge direction finding method based on the neural network and the wireless sensor array, in step S300, an angle of the wireless directional sensors distributed on the circumference is used as an abscissa, an amplitude of a partial discharge signal received by the wireless directional sensors is used as an ordinate, curve fitting is performed on each group of data, a maximum value on the curve is found, an angle on the abscissa corresponding to the maximum value is a directional result corresponding to the group of data, and then an average value of all directional results is a direction of the partial discharge source measured within a period of time.
The scheme improves the positioning precision by curve fitting. In order to improve the positioning accuracy, one method is to increase the number of the wireless directional sensors in the wireless sensor array, so as to further uniformly divide the distribution directions of the wireless directional sensors based on the circumference, but this will cause the cost to be continuously increased, and the circumferential space is limited, so that the sensors cannot be increased without limit. Another method is to perform curve fitting according to the above scheme, so that a virtual wireless directional sensor position and corresponding received signal amplitude based on the fitted curve can be found on the coordinates, thereby improving the accuracy while saving the cost.
Furthermore, in the partial discharge direction-finding method based on the neural network and the wireless sensor array, the wireless directional sensor is also provided with a second specific receiving direction, an included angle of 180 degrees is fixed between the second specific receiving direction and the first specific receiving direction, and when the second specific receiving direction is over against the partial discharge source, the received signal intensity is weakest; in step S300, the minimum value in each group of data is respectively searched, the direction pointed by plus/minus 180 ° from the second specific receiving direction of the wireless directional sensor corresponding to the minimum value is the directional result corresponding to the group of data, and then the average value of all directional results is the direction of the local discharge source measured in the period of time.
Considering that the sensing characteristic of the wireless directional sensor is usually that the received signal strength is strongest, i.e. the gradient near the maximum amplitude is small, and the received signal strength is weakest, i.e. the gradient near the minimum amplitude is large, in order to quickly and accurately find the extreme value, the above-mentioned scheme adopts the steps of finding the minimum value of each group of data, and taking the direction pointed by plus/minus 180 degrees in the second specific receiving direction of the wireless directional sensor corresponding to the minimum value as the corresponding directional result of the group of data.
Furthermore, in the partial discharge direction-finding method based on the neural network and the wireless sensor array, the wireless directional sensor is also provided with a second specific receiving direction, an included angle of 180 degrees is fixed between the second specific receiving direction and the first specific receiving direction, and when the second specific receiving direction is over against the partial discharge source, the received signal intensity is weakest; in step S300, the angle of the wireless directional sensors distributed on the circumference is used as an abscissa, the amplitude of the local discharge signal received by the wireless directional sensors is used as an ordinate, curve fitting is performed on each group of data, and a minimum value on the curve is found, an angle plus/minus 180 ° on the abscissa corresponding to the minimum value is an orientation result corresponding to the group of data, and then an average value of all orientation results is the direction of the local discharge source measured within the period of time.
According to the scheme, the positioning accuracy is improved by curve fitting, the minimum value of the fitting curve is searched, and the direction pointed by plus/minus 180 degrees in the second specific receiving direction of the virtual wireless orientation sensor corresponding to the minimum value is used as the orientation result corresponding to the group of data.
Further, in the partial discharge direction finding method based on the neural network and the wireless sensor array, the wireless directional sensor comprises a single-pole PCB antenna.
The scheme is used for manufacturing the wireless directional sensor based on a single-pole PCB antenna serving as a partial discharge ultrahigh frequency sensor. Usually the monopole type PCB antenna is mounted in a metal container and encapsulated in an electromagnetic wave permeable material in the first specific receiving direction, so that due to electromagnetic shielding effects the received signal strength is maximal when the first specific receiving direction is facing the partial discharge source, the received signal strength is minimal when the first specific receiving direction is turned 180 ° away from the partial discharge source, i.e. when the second specific receiving direction is facing the partial discharge source, and a smooth transition is usually formed between the two extremes during the rotation of the sensor.
Further, in the partial discharge direction finding method based on the neural network and the wireless sensor array, the neural network is a BP neural network.
The bp (back propagation) neural network is a multi-layer feedforward neural network trained according to an error back propagation algorithm, and is the most widely used neural network at present. The scheme utilizes the learning capacity of the neural network on the complex random nonlinear process to screen out a data set with high reliability.
Furthermore, in the partial discharge direction finding method based on the neural network and the wireless sensor array, the curve fitting adopts cubic spline interpolation fitting.
In the above scheme, the Cubic Spline Interpolation (Cubic Spline Interpolation) is a process of obtaining a curve function set mathematically by solving a three bending moment equation set through a smooth curve of a series of shape value points.
Another objective of the present invention is to provide a partial discharge direction-finding system based on a neural network and a wireless sensor array, which can accurately locate a partial discharge source in a partial discharge direction-finding manner, and has the advantages of low equipment cost, small equipment volume, good equipment portability, good environmental adaptability, and high direction-finding precision and accuracy.
According to the above object, the present invention provides a partial discharge direction-finding system based on a neural network and a wireless sensor array, the system includes a wireless sensor array and a data processing device, which are connected by data, and the direction of the partial discharge source is determined by any one of the above methods, wherein the data processing device executes the steps S200 and S300.
According to the partial discharge direction finding system based on the neural network and the wireless sensor array, the direction of the partial discharge source is determined by adopting any one of the partial discharge direction finding methods, so that according to the principle, the system can accurately position the partial discharge source in the partial discharge direction finding direction, and meanwhile, the system has the advantages of low equipment cost, small equipment volume, good equipment portability, good environmental adaptability, high direction finding precision and high direction finding accuracy.
Compared with the partial discharge source positioning method based on the time delay sequence, such as TOA or TDOA, the partial discharge direction-finding method based on the neural network and the wireless sensor array, provided by the invention, comprises the following steps: the local discharge source positioning method based on the time delay sequence, such as TOA or TDOA, needs to keep nanosecond time synchronization among sensors and at least a few GSa/s sampling rates, so that the hardware cost is high, the equipment volume is large, and the portability is poor. The method of the invention combines the RSSI technology, the neural network technology and the wireless sensor array technology to carry out the orientation measurement of the partial discharge source, thereby realizing the accurate positioning of the partial discharge source in the partial discharge direction, and simultaneously having lower equipment cost, smaller equipment volume, better equipment portability, better environmental adaptability and higher direction-finding precision and accuracy.
The partial discharge direction-finding system based on the neural network and the wireless sensing array also has the advantages and beneficial effects.
Drawings
Fig. 1 is a schematic flow chart of a partial discharge direction finding method based on a neural network and a wireless sensor array according to the present invention.
Fig. 2 is a schematic flow chart of a partial discharge direction finding method based on a neural network and a wireless sensor array according to an embodiment of the present invention.
Fig. 3 is a schematic structural diagram of a monopole type PCB antenna used in an embodiment of the partial discharge direction-finding method based on a neural network and a wireless sensor array according to the present invention.
Fig. 4 is a schematic bottom perspective view of a metal container of a monopole type PCB antenna used in an embodiment of a partial discharge direction finding method based on a neural network and a wireless sensor array according to the present invention.
Fig. 5 is a schematic top perspective view of a metal container of a monopole type PCB antenna used in one embodiment of a partial discharge direction finding method based on a neural network and a wireless sensor array according to the present invention.
Fig. 6 is a schematic structural diagram of a wireless sensor array used in an embodiment of a partial discharge direction finding method based on a neural network and the wireless sensor array according to the present invention.
Fig. 7 is a schematic diagram of a typical topology of a BP neural network used in an embodiment of a partial discharge direction finding method based on a neural network and a wireless sensor array according to the present invention.
FIG. 8 is a schematic diagram of a cubic spline interpolation fitting curve of a set of magnitude data in a coordinate system in a field verification example.
Fig. 9 is a schematic diagram of the orientation results corresponding to each set of data obtained based on the amplitude data without being screened by the BP neural network in the field verification example.
Fig. 10 is a schematic diagram of the orientation results corresponding to each set of data obtained based on the amplitude data screened by the BP neural network in the field verification example.
Detailed Description
The partial discharge direction-finding method and system based on the neural network and the wireless sensor array according to the present invention will be further described in detail with reference to the drawings and specific embodiments of the specification.
Fig. 1 illustrates a flow of a partial discharge direction finding method based on a neural network and a wireless sensing array.
As shown in fig. 1, the flow of the partial discharge direction finding method based on the neural network and the wireless sensor array according to the present invention includes:
s100: acquiring amplitude data of a plurality of groups of partial discharge signals within a period of time through a wireless sensor array, wherein each group of data corresponds to a corresponding time point respectively, and each data in each group of data corresponds to a corresponding wireless directional sensor respectively;
s200: constructing a neural network, screening amplitude data of a plurality of groups of partial discharge signals based on the neural network, and leaving a data group with high reliability;
s300: the direction of the partial discharge source is determined based on the data set filtered by step S200.
The wireless sensor array comprises a plurality of wireless directional sensors, each wireless directional sensor at least has a first specific receiving direction, and the received signal strength is strongest when the first specific receiving direction is over against the local discharge source; the first specific receiving directions of the wireless directional sensors in the wireless sensing array are respectively pointed to different directions to receive the partial discharge signals.
In some embodiments, the wireless orientation sensors in the wireless sensor array are circumferentially and uniformly arranged on a circumference to form a circular array, wherein the first specific receiving direction of each wireless orientation sensor points to the outside of the circular array along a circumference radius.
In some embodiments, in step S300, a maximum value in each set of data is respectively searched, a direction pointed by the first specific receiving direction of the wireless orientation sensor corresponding to the maximum value is an orientation result corresponding to the set of data, and then an average value of all the orientation results is a direction of the local discharge source measured in the period of time.
In some embodiments, in step S300, an angle of the wireless directional sensors distributed on the circumference is taken as an abscissa, an amplitude of the partial discharge signal received by the wireless directional sensor is taken as an ordinate, curve fitting is performed on each group of data, a maximum value on the curve is found, an angle on the abscissa corresponding to the maximum value is an orientation result corresponding to the group of data, and then an average value of all the orientation results is a direction of the partial discharge source measured in a period of time.
In some embodiments, the wireless orientation sensor further has a second specific receiving direction having a fixed 180 ° angle with the first specific receiving direction, the received signal strength being weakest when the second specific receiving direction is opposite to the partial discharge source; in step S300, the minimum value in each group of data is respectively searched, the direction pointed by plus/minus 180 ° from the second specific receiving direction of the wireless directional sensor corresponding to the minimum value is the directional result corresponding to the group of data, and then the average value of all directional results is the direction of the local discharge source measured within a period of time.
In some embodiments, the wireless orientation sensor further has a second specific receiving direction having a fixed 180 ° angle with the first specific receiving direction, the received signal strength being weakest when the second specific receiving direction is directly opposite to the partial discharge source; in step S300, the angle of the wireless directional sensors distributed on the circumference is used as an abscissa, the amplitude of the partial discharge signal received by the wireless directional sensors is used as an ordinate, curve fitting is performed on each group of data, and a minimum value on the curve is found, the angle plus/minus 180 ° on the abscissa corresponding to the minimum value is the directional result corresponding to the group of data, and then the average value of all the directional results is the direction of the partial discharge source measured within a period of time.
In some embodiments, the wireless directional sensor includes a monopole-type PCB antenna.
In certain embodiments, the neural network is a BP neural network.
In some embodiments, the curve fitting uses a cubic spline interpolation fit.
Fig. 2 illustrates a flow of a partial discharge direction finding method based on a neural network and a wireless sensor array in one embodiment. Fig. 3 illustrates a schematic structural diagram of a monopole type PCB antenna used in one embodiment of a partial discharge direction finding method based on a neural network and a wireless sensing array. Fig. 4 illustrates a bottom perspective structure of the metal container of the monopole-type PCB antenna in this embodiment. Fig. 5 illustrates a top perspective view of the metal container of the monopole-type PCB antenna in this embodiment. Fig. 6 illustrates the structure of the wireless sensor array in this embodiment. Fig. 7 illustrates a typical topology of the BP neural network in this embodiment.
As shown in fig. 2, the flow of the partial discharge direction-finding method based on the neural network and the wireless sensor array according to an embodiment of the present invention includes the following steps 1 to 11:
wherein the wireless sensing array comprises a plurality of wireless directional sensors, the wireless directional sensors comprise a monopole type PCB antenna and a corresponding signal processing circuit, the monopole type PCB antenna (a rectangular frame in FIG. 3 schematically represents a circuit board of the monopole PCB antenna, and an ellipse schematically represents a metal antenna) as shown in FIG. 3 is installed at a bottom space B of the metal container as shown in FIG. 4, the bottom space A is filled with an encapsulation medium, wherein the first specific receiving direction is encapsulated by a material which can transmit electromagnetic waves, and the other directions are encapsulated by a material which can not transmit electromagnetic waves. The direction having an angular difference of 180 ° with respect to the first specific receiving direction is the second specific receiving direction of the wireless orientation sensor. The received signal strength is strongest when the first specific receiving direction is over against the partial discharge source, the received signal strength is weakest when the second specific receiving direction is over against the partial discharge source, and the received signal strength smoothly transitions when the direction between the first specific receiving direction and the second specific receiving direction is over against the partial discharge source. A signal processing circuit for converting a signal received by the monopole-type PCB antenna into amplitude data output is installed at the head space a of the metal container as shown in fig. 5.
Wireless sensor array as shown in fig. 6, a total of 12 wireless directional sensors S1-S12 are arranged uniformly in the circumferential direction on a circular frame 1 to form a circular array, wherein the first specific receiving direction of each wireless directional sensor is respectively directed to the outside of the circular array along the radius of the circumference to receive the partial discharge signal. Taking the wireless orientation sensor S1 as an example, the first specific receiving direction D is directed to the outside of the circular array along the radius of the circumference, and the direction having an angle difference of 180 ° with the first specific receiving direction D is the second specific receiving direction E of the wireless orientation sensor. In consideration of the fact that the gradient near the maximum value of the signal strength is small and the gradient near the minimum value of the signal strength is large in the transition characteristic, the embodiment finds the minimum value of the signal strength received by the wireless directional sensor and the azimuth angle of the corresponding wireless directional sensor in the wireless sensing array, and determines the direction of the local discharge source based on the azimuth angle.
Step 1: and starting.
Step 2: the sensors collect and transmit data. In this step, a total of 12 wireless directional sensors S1-S12 continuously collect partial discharge signals at the same time, and the partial discharge signals are grouped in a time series arrangement, converted into corresponding amplitude data, and transmitted to a data processing device.
And step 3: discharge amplitude information for 1 minute was acquired. In the step, the partial discharge lasts for 1 minute, and the partial discharge is generated for 1 time per second, the data processing device acquires amplitude data of 60 groups of partial discharge signals within 1 minute through the wireless sensor array, wherein each group of data corresponds to a corresponding time point, and each data in each group of data corresponds to a corresponding wireless directional sensor.
And 4, step 4: and (6) normalizing. And the step of putting each group of amplitude data into a coordinate system through the data processing device and carrying out normalization processing. The angle of the wireless directional sensors distributed on the circumference is used as an abscissa, the amplitude of the partial discharge signals received by the wireless directional sensors is used as an ordinate, and the amplitude data of the 60 groups of partial discharge signals are put into a coordinate system to form 60 groups of coordinate data. Since the angle of the 1 st wireless orientation sensor S1 can be either 0 ° or 360 °, we will consider the 1 st wireless orientation sensor S1 to be also the 13 th wireless orientation sensor.
And 5: and (5) training a neural network. In the step, a BP neural network is constructed through the data processing device, amplitude data of the 60 groups of partial discharge signals are screened based on the BP neural network, and a data group with high reliability is left.
In this step, for 60 sets of data, some data may be inaccurate and need to be screened, otherwise a large amount of errors are caused in the subsequent orientation. The embodiment utilizes the learning ability of the BP neural network to a complex random nonlinear process to carry out data screening. The BP neural network adopts a three-layer structure: an input layer, a hidden layer, and an output layer. The input layer is 13 pieces of ordinate data of each group of amplitude data, the output layer only contains one output node, the result is expressed by 1 or 0, and the hidden layer is connected with other layers through the weight.
When the difference between a certain numerical value in the input ordinate data corresponding to the 13 wireless directional sensors and the average value of all 60 discharges corresponding to the wireless directional sensor is large, the reliability of the data is considered to be low, the final output result is represented by 0, and the data in the group is considered to be large in deviation and large directional errors can be caused; when the fluctuation is small, the reliability is considered to be high, and the output result is 1. A typical structure of BP neural network is shown in FIG. 7, where X isInput layer, W is hidden layer, Y is output layer, x1-xmAs an input node, m is 13, y in this embodiment1-ynIn this embodiment, n is 1, w is an output nodekiAnd wijIs the connection weight. After the neural network screening, a data set with a result of 1 is left, so that a data set with small data fluctuation can be obtained. And subsequently operating on the left data group.
Step 6: and (6) performing interpolation fitting on the data. The step of performing interpolation fitting on the screened data sets respectively through the data processing device.
In the step, each set of the coordinate data left after screening is subjected to curve fitting respectively to form a fitting curve corresponding to each set of the coordinate data.
In this step, the curve fitting adopts cubic spline interpolation fitting.
Cubic spline interpolation, is performed at every two adjacent abscissas [ x ]j,xj+1]Let s (x)j)＝yj，s(xj+1)＝yj+1Using a cubic function in the subinterval [ x ]j,xj+1]To get s (x)j)、s(xj+1) And connecting, wherein the expression is as follows:
s(x)＝ajx3+bjx2+cjx+dj (1)
where x ∈ (x)j,xj+1)，j＝1,2,…,12，aj,bj,cj,djIs the undetermined coefficient.
The same cubic spline function is established in each subinterval, and 12 subintervals are total, so that 12 unknown cubic functions and 48 unknown numbers exist, and the smoothness condition at the connecting node can be obtained:
adding the equations of 13 points for 46 equations and 48 unknowns, two boundary conditions need to be supplemented, and the present embodiment selects the natural boundary condition, i.e., the second order of the boundary pointsDerivative s' (x)1) And s' (x)13) The value is 0.
And (4) after the simultaneous equation is solved to obtain an unknown number, connecting the 13 points by utilizing a cubic spline function to obtain a fitting curve.
And 7: and searching an angle corresponding to the lowest point of the curve. In this step, the data processing device searches for the minimum value on each fitting curve, and determines the angle on the abscissa corresponding to the minimum value.
And 8: the angle is +/-180 degrees. In this step, the data processing device adds/subtracts 180 ° to the angle on the abscissa corresponding to each minimum value, which is the orientation result corresponding to each set of data.
And step 9: and obtaining a positioning result. In this step, the average of all the orientation results by the data processing device is the direction of the partial discharge source measured in the above 1 minute.
Step 10: and judging whether the positioning is finished or not. In the step, the data processing device judges whether the positioning is finished according to the direction-finding instruction, if so, the step 11 is executed, and if not, the step 2 is executed again.
Step 11: and (6) ending.
The wireless sensing array and the data processing device are in data connection with each other to form the partial discharge direction-finding system based on the neural network and the wireless sensing array. The system determines the direction of the partial discharge source by adopting the partial discharge direction-finding method.
The above embodiments were verified by field testing as follows.
FIG. 8 illustrates a cubic spline interpolation fit curve of a set of magnitude data in a coordinate system. Fig. 9 illustrates the orientation results corresponding to each set of data obtained based on the amplitude data without BP neural network screening. Fig. 10 illustrates the orientation results corresponding to each set of data obtained based on the amplitude data filtered by the BP neural network. The abscissa of fig. 8 is an angle, the ordinate is a normalized amplitude, the legend F is a fitting curve, the legend G is an actually obtained amplitude data coordinate, and the legend H is an abscissa indicator line corresponding to a minimum ordinate on the found fitting curve. The circumferential coordinate of fig. 9 and 10 is an angle and the radial coordinate is a normalized amplitude.
In a high pressure lobby, the direction determination of the partial discharge source is performed in accordance with the system and method of the above-described embodiment. 12 wireless directional sensors are selected to construct a wireless sensing array, so that the resolution is 30 degrees, and the azimuth angle of the partial discharge source is set to be 240 degrees. The partial discharge lasted 1 minute, producing 1 partial discharge per second. Discharge amplitude information for 1 minute was acquired. Connecting the amplitude data into a continuous smooth curve by utilizing a cubic spline interpolation method, and calculating the lowest point between 0 degrees and 360 degrees on the curve. A schematic of cubic spline interpolation on the data is shown in fig. 8. The orientation results without BP neural network screening are shown in figure 9, and most of the orientation results are between 240 degrees and 300 degrees, and in addition, a small amount of scatter points exist, so that the dispersity is high. After the screening of the BP neural network, the directional result is greatly reduced, but the concentration of the directional result is greatly improved, as shown in FIG. 10, the calculated average value is 245.9 degrees, the angle is the direction of the local discharge source measured in the 1 minute, and the error between the direction of the actual local discharge source and the direction is 5.9 degrees.
It is to be noted that the above lists only specific embodiments of the present invention, and it is obvious that the present invention is not limited to the above embodiments, and many similar variations follow. All modifications which would occur to one skilled in the art and which are, therefore, directly derived or suggested from the disclosure herein are deemed to be within the scope of the present invention.
Claims (7)
1. A partial discharge direction-finding method based on a neural network and a wireless sensor array is used for determining the direction of a partial discharge source, and is characterized in that: the wireless sensing array comprises a plurality of wireless directional sensors, each wireless directional sensor at least has a first specific receiving direction, the received signal strength is strongest when the first specific receiving direction is over against a local discharge source, each wireless directional sensor also has a second specific receiving direction, an included angle of 180 degrees is fixed between each wireless directional sensor and the first specific receiving direction, and the received signal strength is weakest when the second specific receiving direction is over against the local discharge source; the first specific receiving directions of the wireless directional sensors in the wireless sensing array point to different directions respectively to receive partial discharge signals; the wireless directional sensors in the wireless sensing array are uniformly arranged on a circumference in the circumferential direction to form a circular array, wherein the first specific receiving direction of each wireless directional sensor points to the outside of the circular array along the radius of the circumference; the method comprises the following steps:
s100: acquiring amplitude data of a plurality of groups of partial discharge signals within a period of time through the wireless sensor array, wherein each group of data corresponds to a corresponding time point respectively, and each data in each group of data corresponds to a corresponding wireless directional sensor respectively;
s200: constructing a neural network, screening the amplitude data of the plurality of groups of partial discharge signals based on the neural network, and leaving a data group with high reliability;
s300: determining the direction of the partial discharge source based on the data set screened in step S200: respectively searching for the maximum value in each group of data, wherein the direction pointed by the first specific receiving direction of the wireless directional sensor corresponding to the maximum value is the directional result corresponding to the group of data, and then averaging all the directional results to obtain the direction of the local discharge source measured in the period of time; respectively searching for the minimum value in each group of data, wherein the direction pointed by plus/minus 180 degrees of the second specific receiving direction of the wireless directional sensor corresponding to the minimum value is the directional result corresponding to the group of data, and then averaging all the directional results, namely the direction of the local discharge source measured in the period of time.
2. The method according to claim 1, wherein in step S300, an angle of the wireless directional sensors distributed on a circumference is used as an abscissa, an amplitude of the local discharge signal received by the wireless directional sensors is used as an ordinate, curve fitting is performed on each set of data, a maximum value on the curve is found, an angle on the abscissa corresponding to the maximum value is an orientation result corresponding to the set of data, and then an average of all the orientation results is a direction of the local discharge source measured in the period of time.
3. The partial discharge direction-finding method based on the neural network and the wireless sensor array as claimed in claim 1, wherein the wireless directional sensor further has a second specific receiving direction which has a fixed 180 ° angle with the first specific receiving direction, and the received signal strength is weakest when the second specific receiving direction is opposite to the partial discharge source; in step S300, the angle of the wireless directional sensors distributed on the circumference is used as an abscissa, the amplitude of the local discharge signal received by the wireless directional sensors is used as an ordinate, curve fitting is performed on each group of data, and a minimum value on the curve is found, an angle plus/minus 180 ° on the abscissa corresponding to the minimum value is an orientation result corresponding to the group of data, and then an average value of all orientation results is the direction of the local discharge source measured within the period of time.
4. The neural network and wireless sensor array-based partial discharge direction-finding method of claim 1, wherein the wireless directional sensor comprises a monopole-type PCB antenna.
5. The partial discharge direction-finding method based on the neural network and the wireless sensing array as claimed in claim 1, wherein the neural network is a BP neural network.
6. The partial discharge direction finding method based on the neural network and the wireless sensing array as claimed in claim 2 or 3, wherein the curve fitting adopts cubic spline interpolation fitting.
7. A partial discharge direction-finding system based on a neural network and a wireless sensing array, which is characterized in that the system comprises the wireless sensing array and a data processing device which are connected by data, and the direction of a partial discharge source is determined by adopting the partial discharge direction-finding method of any one of claims 1 to 6, wherein the data processing device executes the steps S200 and S300.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910175714.9A CN109991519B (en) | 2019-03-08 | 2019-03-08 | Partial discharge direction-finding method and system based on neural network and wireless sensor array |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201910175714.9A CN109991519B (en) | 2019-03-08 | 2019-03-08 | Partial discharge direction-finding method and system based on neural network and wireless sensor array |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109991519A CN109991519A (en) | 2019-07-09 |
| CN109991519B true CN109991519B (en) | 2021-11-16 |
Family
ID=67130309
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201910175714.9A Active CN109991519B (en) | 2019-03-08 | 2019-03-08 | Partial discharge direction-finding method and system based on neural network and wireless sensor array |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109991519B (en) |
Families Citing this family (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110501620B (en) * | 2019-09-18 | 2022-05-17 | 国网陕西省电力公司咸阳供电公司 | Signal source characteristic-based station domain partial discharge positioning method |
| CN113589116A (en) * | 2021-08-11 | 2021-11-02 | 山东通广电子有限公司 | Method and system for detecting partial discharge position |
| CN113791318A (en) * | 2021-09-01 | 2021-12-14 | 上海交通大学 | Partial discharge fault identification method and system based on multispectral detection array |
| CN116165599B (en) * | 2023-04-24 | 2023-06-27 | 武汉能钠智能装备技术股份有限公司四川省成都市分公司 | Ultrashort wave direction finding system and integrated ultrashort wave direction finding equipment |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101290346A (en) * | 2008-04-29 | 2008-10-22 | 江苏维尔科技有限公司 | Wrist strap type positioning apparatus |
| CN106019098A (en) * | 2016-06-15 | 2016-10-12 | 国网北京市电力公司 | Partial discharge positioning method, device and system |
| CN106771933A (en) * | 2017-01-23 | 2017-05-31 | 天津大学 | Power cable shelf depreciation high frequency electric monitoring system based on wireless network |
| CN106970354A (en) * | 2017-04-10 | 2017-07-21 | 南京邮电大学 | A kind of 3-D positioning method based on multiple light courcess and photosensor array |
| CN107656180A (en) * | 2017-09-25 | 2018-02-02 | 国网黑龙江省电力有限公司电力科学研究院 | A kind of shelf depreciation intelligence sensor detection method |
Family Cites Families (9)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101268761B1 (en) * | 2011-09-19 | 2013-05-29 | 한국전력공사 | Method for deciding defect type of partial discharge and apparatus thereof |
| CN102665273A (en) * | 2012-04-19 | 2012-09-12 | 中国科学技术大学苏州研究院 | Wireless sensor network node positioning method |
| CN102967807B (en) * | 2012-11-10 | 2015-04-22 | 西安交通大学 | Insulation defect locating method |
| CN103713242A (en) * | 2013-12-27 | 2014-04-09 | 上海交通大学 | Novel ultrahigh frequency sensor for positioning local discharge source space and array thereof |
| CN103969559A (en) * | 2014-05-15 | 2014-08-06 | 国家电网公司 | Method and system for positioning local discharge space of power equipment of transformer substation |
| CN106291281A (en) * | 2016-08-08 | 2017-01-04 | 国网上海市电力公司 | A kind of substation equipment shelf depreciation alignment system and method thereof |
| CN107085171B (en) * | 2017-04-01 | 2019-10-11 | 西安交通大学 | A kind of optimization placement method of the superfrequency phased array for substation's shelf depreciation direction finding |
| CN107179491A (en) * | 2017-06-21 | 2017-09-19 | 国家电网公司 | A kind of partial discharge positioning method and system based on pattern-recognition |
| CN107942212A (en) * | 2017-11-17 | 2018-04-20 | 国网天津市电力公司 | A kind of substation's partial discharge positioning method without blur estimation based on spatial spectrum |
-
2019
- 2019-03-08 CN CN201910175714.9A patent/CN109991519B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN101290346A (en) * | 2008-04-29 | 2008-10-22 | 江苏维尔科技有限公司 | Wrist strap type positioning apparatus |
| CN106019098A (en) * | 2016-06-15 | 2016-10-12 | 国网北京市电力公司 | Partial discharge positioning method, device and system |
| CN106771933A (en) * | 2017-01-23 | 2017-05-31 | 天津大学 | Power cable shelf depreciation high frequency electric monitoring system based on wireless network |
| CN106970354A (en) * | 2017-04-10 | 2017-07-21 | 南京邮电大学 | A kind of 3-D positioning method based on multiple light courcess and photosensor array |
| CN107656180A (en) * | 2017-09-25 | 2018-02-02 | 国网黑龙江省电力有限公司电力科学研究院 | A kind of shelf depreciation intelligence sensor detection method |
Non-Patent Citations (2)
| Title |
|---|
| Partial Discharge Angle of Arrival Estimation Using UHF Wireless Sensor Array;Li Ren;《IOP Conference Series: Materials Science and Engineering》;20190331;全文 * |
| RSSI-Power-Based Direction of Arrival Estimation of Partial Discharges in Substations;Fan Wu;《energies》;20190906;全文 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109991519A (en) | 2019-07-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109991519B (en) | Partial discharge direction-finding method and system based on neural network and wireless sensor array | |
| CN106226761B (en) | A kind of relevant higher-frequency radar multifrequency detection method of high-performance | |
| CN107181543B (en) | Three-dimensional indoor passive positioning method based on propagation model and position fingerprint | |
| CN101424737B (en) | Wireless sensor network signal chart positioning method based on signal strength distance | |
| CN109946573B (en) | Local discharge direction finding method and system based on clustering and wireless sensor array | |
| CN102944866A (en) | Interferometer system based secondary surveillance radar response signal direction-finding method | |
| CN107390097A (en) | A kind of acoustoelectric combined shelf depreciation simulation detection system of GIS and its detection method | |
| CN111948598B (en) | Method and device for detecting space domain interference signal | |
| CN109507635A (en) | Utilize the array amplitude phase error evaluation method of two unknown orientation auxiliary sources | |
| CN110837079B (en) | Target detection method and device based on radar | |
| CN103000996A (en) | Uniform circular array direction-finder antenna receiving mutual impedance test and mutual coupling compensation system | |
| CN107682822B (en) | Compressed sensing outdoor positioning method based on electromagnetic field intensity | |
| CN113960523A (en) | Universal ultra-wideband correction direction-finding method and system based on FPGA | |
| CN111505566B (en) | Ultrahigh frequency radio frequency signal DOA estimation method | |
| JP4294634B2 (en) | Calibration table creation method and arrival direction estimation method | |
| CN109946572B (en) | Partial discharge signal direction finding method and system based on circular ultrahigh frequency sensor array | |
| CN109752633B (en) | Method and system for positioning partial discharge signal of transformer substation | |
| CN110907925B (en) | Weight positioning method under high-frequency ground wave radar double-station model | |
| CN113567916B (en) | Correlation interferometer direction finding method, system, equipment and storage medium | |
| CN105008950A (en) | Measuring device and measuring method for direction finding and direction uncertainty determination | |
| CN113777556B (en) | Radio signal three-dimensional amplitude comparison direction finding method and device | |
| CN111505391B (en) | Method for detecting mismatched antenna unit in array antenna | |
| CN115835467B (en) | Plasma three-dimensional positioning system and method based on active and passive fusion | |
| CN113050036B (en) | GIS beam forming positioning method based on multi-resonance point microphone array | |
| CN113640737B (en) | Two-dimensional power distribution-based small array element array high-resolution azimuth estimation method |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |