CN116184182A - A GIS Disconnector Mechanical State Recognition Method Based on Curve Similarity - Google Patents

Abstract

The invention relates to a method for identifying the mechanical state of a GIS isolating switch based on curve similarity, and belongs to the technical field of power systems. The present invention obtains the power curve of the drive motor in each typical state, first realizes the segmentation of the stroke curve based on the meshing point recognition method of the multi-window slope, and secondly uses the Fresche distance to describe the shape of the curve, and designs the mechanical state of the GIS isolating switch based on the matching of the curve shape Identify the process, and finally verify the practicability of the method by combining experiments and measured data. The method of the invention can quickly identify the mechanical state of the GIS isolating switch, has high accuracy and is easy to popularize and apply.

Description

A method for identifying mechanical state of GIS disconnector based on curve similarity

Technical Field

The invention belongs to the technical field of power systems, and in particular relates to a method for identifying the mechanical state of a GIS disconnector based on curve similarity.

Background Art

The disconnector is an important node in the power system and the most widely used switchgear in the substation. Its mechanical state has an important impact on the stable operation of the power system. In recent years, disconnectors have frequently failed. Common faults include inadequate opening and closing, jamming, and three-phase dyssynchronism. For example, inadequate opening and closing, operating the grounding switch when the disconnector on the busbar side is not in place, and a short-circuit grounding fault occurs, which is a serious accident in the power system. Therefore, the detection of the mechanical state of the disconnector is an important concern of the power system. Disconnectors are divided into two types, open disconnectors and gas insulated disconnectors (GIS). The detection methods for the mechanical state of open disconnectors include direct detection methods such as pressure sensors, attitude sensors, and micro switches, and two state sensing methods such as motor current and vibration signal detection. The first several direct detection methods are not suitable for GIS disconnectors, because the contact fingers of GIS disconnectors are sealed in the gas cavity, and it is difficult to arrange sensors in direct detection methods. If sensors are arranged in advance, there are also problems such as difficulty in leading out signal lines, which is completely unacceptable in industry. There is also a way to install the travel switch in the mechanism box of GIS to detect its opening and closing travel, but this method is difficult to detect serious faults such as jamming and connecting rod breakage. Therefore, when analyzing the shortcomings of existing technologies, the shortcomings of state perception methods are mainly analyzed:

1. Drive motor current sensing:

Chinese patent CN113933567A discloses an online monitoring system for the opening and closing status of a GIS disconnector, including: a Hall sensor for obtaining the motor current value of the disconnector to be tested; the motor current value includes: the motor current value of the disconnector to be tested when it moves from the open position to the closed position, and the motor current value of the disconnector to be tested when it moves from the closed position to the open position; a monitoring device connected to the Hall sensor through a shielded cable and an aviation connector; the monitoring device receives the motor current value and converts it into a current time waveform; a data analysis and diagnostic instrument wirelessly connected to the monitoring device; the data analysis and diagnostic instrument is used to obtain the current time waveform and the normal current time waveform pre-stored in the monitoring device and compare and analyze to obtain a status determination result. This can accurately determine the opening and closing status of the GIS disconnector, thereby ensuring the safety of electrical equipment.

The above patent is a method of sensing the current state of a driving motor. Although the description indicates that it monitors the on/off state of the GIS disconnector, the detection method based on the motor current can theoretically reflect all mechanical states of the GIS disconnector, because the change in mechanical state will definitely be reflected on the motor shaft, and this change in mechanical state will definitely bring about a change in the motor current. However, there is a premise, which is also the unified defect of all motor current detection methods including the above patent: the driving voltage of the driving motor needs to remain unchanged. When the voltage changes, if the voltage is not measured, all conclusions about the change in current caused by the change in mechanical state are invalid. Even if the detection method based on motor current takes the current into consideration, there is no defect that if the phase between voltage and current changes, the conclusion that the current reflects the mechanical state is invalid.

2. Vibration signal perception:

Chinese patent CN115201673A provides an online monitoring system for abnormal vibration of disconnectors of GIS equipment, which relates to the field of substation equipment monitoring, and includes a vehicle body and a data processing module, a data output module and an image acquisition module all mounted on the vehicle body; the image processing module is used to acquire an image of the disconnector and send the image of the disconnector to the data processing module, the data processing module analyzes and determines the position of the disconnector and the switching action trend of the disconnector, and converts the results of the analysis and judgment into switch position and status signals, and marks the switch position and the status signals; the data output module sends the mark to a mobile terminal. The present invention has the advantages of ensuring that the basis for the operation of the disconnector is accurate, reducing manual workload, improving work efficiency, and reducing the occurrence of misoperation and accidents.

The above patent is a detection method for the vibration signal of an isolating switch. In fact, there are two methods for monitoring the isolating switch through the vibration signal (for example, references [1] to [4]):

(1) Detect the vibration signal during the opening and closing process of the disconnector, and use this vibration signal to reflect the mechanical state of the disconnector. The underlying logic is that the change in the mechanical state of the disconnector will definitely be reflected in the vibration signal during the opening and closing process. Because vibration is generated by machinery, the vibration signal contains all the information of the mechanical structure.

(2) The contact state of the disconnector is analyzed by monitoring the vibration signal of the disconnector in the closed and current-carrying state. The theoretical basis is that, through deduction, when the disconnector is in the current-carrying state, its cavity will be subjected to an electromagnetic force of twice the power frequency. Once the contact state changes, the proportion of this double frequency in the vibration signal will also change. By analyzing the double frequency content in the vibration signal, the mechanical state of the GIS disconnector can be deduced.

[1] Wu Xutao, Zhao Jinfei, Ma Yunlong, He Ninghui, Ma Bo, Li Junhao. GIS mechanical defect diagnosis method based on vibration response under multi-frequency excitation[J]. Power Capacitors and Reactive Power Compensation, 2022, 43(04): 108-115. DOI: 10.14044/j.1674-1757.pcrpc.2022.04.015.

[2] Chen Fuguo, Cai Jie, Li Zhongqi. Research on fault diagnosis of high-voltage disconnector based on long short-term memory network[J]. China Testing, 2022, 48(07): 114-119.

[3] Wang Xupeng. Research on vibration status detection of GIS equipment and identification method of poor contact defects of disconnectors[D]. Chongqing University, 2021. DOI: 10.27670/d.cnki.gcqdu.2021.002650.

[4] Zhao Tinggang, Liu Hao, Su Xuhui, Liu Pei, Jiang Huan, Zhao Lihua. Experimental study on contact defect of GIS disconnector based on vibration signal[J]. Hydropower Energy Science, 2020, 38(04): 158-161.

However, all detection methods based on vibration signals have the following disadvantages and shortcomings:

1) Insufficient denoising methods. All current GIS disconnector mechanical state perception based on vibration signals has an unavoidable problem, which is denoising. Because the on-site environment is very complex, the sources of vibration signals are not just the two points mentioned above. Environmental noise (environmental noise includes vibration of other energized equipment, wind, wire dancing, etc.) can also produce vibration. Existing denoising methods mostly use highly complex intelligent algorithms or signal processing algorithms, which require too much calculation and are not suitable for on-site applications. More importantly, current denoising methods are all aimed at a single scenario, such as a specific set of measured signals or experimental environmental noise. At the substation site, the environmental noise is different. Even in the same substation, the noise is different at different locations. Therefore, the denoising methods of the existing technology are not industrially popular.

2) The lack of small sample size. Whether through vibration simulation or through experiments or actual measurements, the number of samples obtained is too small to support the training of intelligent algorithms. In other words, the algorithm model trained in this way can only diagnose the fault types within the range of the collected samples and cannot exceed the quantitative range of the fault degree of its samples in terms of quantitative ability. Once the sample range is exceeded, the diagnostic algorithm is prone to missed reports and false alarms.

Therefore, how to overcome the shortcomings of the prior art is a problem that urgently needs to be solved in the current technical field.

Summary of the invention

The purpose of the present invention is to solve the shortcomings of the prior art and provide a GIS disconnector mechanical state recognition method based on curve similarity. The method is based on multi-window slope meshing point extraction and Fréchet distance similarity analysis, thereby realizing the recognition of the disconnector mechanical state.

To achieve the above purpose, the technical solution adopted by the present invention is as follows:

A method for identifying the mechanical state of a GIS disconnector based on curve similarity comprises the following steps:

Step 1: Acquisition of a template of the power curve of the driving motor during the opening and closing process of the GIS disconnector in multiple mechanical states: Detection of the output power of the driving motor during the opening and closing process of the disconnector in different mechanical states is performed to obtain the power curve in each state as a template; the mechanical state includes a normal state and a defective state;

Step 2: Identification of meshing points and translation points based on multi-window slopes:

Collect the power curve of the driving motor of the GIS disconnector to be identified, and identify its meshing point or disengagement point, as well as the translation point;

Step 3: Calculate the curve similarity based on Fréchet distance:

The power curve of the drive motor to be identified and the closing curve in the power curve under each state are divided into a starting stage, a translation stage and a meshing stage according to their respective translation points and meshing points;

The opening curve is divided into the starting stage, the separation stage and the engagement stage according to the respective translation points and separation points;

Calculate the Fleche distances between the power curve of the drive motor to be identified and the corresponding stages of the power curve of the drive motor in each state in step 1. If the Fleche distance between a certain stage of the power curve of the drive motor to be identified and the power curve of the drive motor in a certain state is the smallest among all the Fleche distances calculated in the stage, then it is determined that the stage is similar to the state, that is, the stage is determined to be the state;

Step 4: Identification of the mechanical status of the GIS disconnector:

If the power curve of the driving motor to be identified is a closing curve, the meshing stage is determined first, and then the translation stage is determined; that is, the corresponding category is entered according to the state determined in the meshing stage, and then the final mechanical state of the GIS disconnector is obtained according to the state determined in the translation stage;

If the power curve of the driving motor to be identified is a tripping curve, first determine the separation stage, and then determine the translation stage; that is, enter the corresponding category according to the state determined in the separation stage, and then obtain the final GIS disconnector mechanical state according to the state determined in the translation stage.

Furthermore, preferably, during the step one detection, a rubber strip is filled between the gear and the conductor to simulate the main shaft jamming defect; the static contact is removed to simulate the phase loss defect; the self-holding circuit power supply is disconnected to simulate the stroke inadequate defect; and the contact finger spring is removed to simulate the contact finger spring deformation defect.

Further, preferably, in step 2, the specific method for identifying the meshing point is:

Obtain multi-window slope features of the standard curve and the curve to be identified;

Find the point in the curve to be identified whose slope characteristics are most similar to the slope characteristics of the meshing point in the standard curve as the suspected meshing point;

Whether the rotation angle of the meshing point is within ±5% of the standard rotation angle, if so, and it is a local minimum, then this point is the meshing point finally identified.

Further, preferably, the multi-window slope feature is obtained in the following manner:

By extracting the slope of the line segment between a point _xi on the curve and points xi _±aj with different spacing on both sides, the multi-window slope feature K = [k _l1 k _r1 k _l2 k _r2 …k _ln k _rn ] of the point is obtained; among them, k _l1 is the slope from point _xi to point xi _-a1 ; k _{r1 is} the slope from point xi to point _xi+a1 ; k _l2 is the slope from point _xi to point xi _-a2 ; k _r2 is the slope from point _xi to point _xi+a2 ; k _ln is the slope from point _xi to point _xi-an ; k _rn is the slope from point _xi to point xi _+an .

Further, preferably, when finding a point in the curve to be identified whose slope feature is most similar to the slope feature of the meshing point in the standard curve as a suspected meshing point, the criterion for similarity of the slope feature is:

(1-σ)*k _{i_bz} ≤k _i ≤(1+σ)*k _{i_bz} (1)

Where, k _{i_bz} is the i-th element in the multi-window slope feature of the meshing point of the standard curve, σ is the allowable fluctuation range of the slope, and k _i is the i-th element in the multi-window slope feature of the to-be-discriminated point of the to-be-identified curve;

Assume that the number of multi-window slope features of a certain point that satisfies formula (1) is β, and δ is the preset minimum number that satisfies formula (1). Then the criterion for the suspected meshing point is:

Further, preferably, in step 3, the specific calculation method of the Fréchet distance is:

(1) The template curve at a certain stage is expressed as

S＝[S ₁ , S ₂ ,…, S _m ,…, S _M ]

Wherein, S _m =(x _m ,y _m ), x, y represent the coordinates of discrete sampling points, m=1, 2, ..., M, represents the sequence number of the sampling point in the sequence S;

(2) The curve to be identified corresponds to a certain stage:

T＝[T ₁ , T ₂ ,…, T _n ,…, T _N ]

Wherein, T _n =(x _n , _yn ), x, y represent the coordinates of discrete sampling points, n = 1, 2, ..., N, represents the sequence number of the sampling point in the sequence T;

(3) Calculate the distance between each discrete point on the template sequence S and the test sequence T:

Finally, we get the distance matrix between sequences:

Wherein, d ₁₁ is the value of d(S ₁ ,T ₁ ); d _MN is the value of d(S _M ,T _N );

(4) Find the shortest distance dmin and the longest distance dmax in the distance matrix. The target Fréchet distance is initialized to DF = dmin, and the loop step size is set to:

Convert the original distance matrix into a 0-1 binary matrix D _0-1:

Where:

(5) Search for 8-connected elements in the binary matrix. When the elements have 8-connectivity, such elements are classified into one category and numbered to obtain the label matrix L. When L(1, 1) = L(M, N) ≠ 0, it means that there is a path connecting the upper left corner and the lower right corner in the matrix;

(6) If there is no such path in (5), then DF = DF + ST, and repeat (4) (5). If there is a connected path or DF = dmax, DF at this time is the final Fréchet distance.

In the present invention, the allowable fluctuation range of the slope can be set according to actual conditions, and the present invention does not impose any limitation thereto, for example, ±1%.

The present invention obtains the power curve of the driving motor in each typical state. Firstly, the segmentation of the stroke curve is realized based on the meshing point recognition method of multi-window slope. Secondly, the Fréchet distance is used to describe the curve shape. A GIS disconnector mechanical state recognition process based on curve shape matching is designed. Finally, the practicability of the method is verified by combining experimental and measured data.

Compared with the prior art, the present invention has the following beneficial effects:

The present invention proposes a meshing point recognition method based on multi-window slope, which extracts the multi-window slope features of the meshing points of the standard curve, finds the points with the most similar slope features in the curve to be identified as suspected meshing points, and then identifies the meshing points in combination with physical features.

The present invention proposes a mechanical state identification method for disconnectors of the same model based on stroke curve morphology identification, which segments the curve to be diagnosed according to the meshing points and determines whether the curve morphology matches the typical state templates of different stages, thereby realizing the identification of the mechanical state.

BRIEF DESCRIPTION OF THE DRAWINGS

Figure 1 is a schematic diagram of the meshing point and the translation point;

Figure 2 is a schematic diagram of multi-window slopes;

Figure 3 is a method for identifying meshing points;

Figure 4 shows the calculation principle of the Fréchet distance;

FIG5 is a typical power curve under an ideal normal state; wherein (a) is an ideal closing power curve; (b) is an ideal opening power curve;

Figure 6 is the calculation process of the Fréchet distance of the opening and closing motor power curves;

Figure 7 shows the typical closing curve of the GIS disconnector, where (a) is the normal state; (b) is the stuck state; (c) is the misaligned state; and (d) is the abnormal stuck state.

DETAILED DESCRIPTION

The present invention is further described in detail below in conjunction with embodiments.

Those skilled in the art will appreciate that the following examples are only used to illustrate the present invention and should not be considered to limit the scope of the present invention. If no specific techniques or conditions are specified in the examples, the techniques or conditions described in the literature in the art or the product specifications are used. If the manufacturer of the materials or equipment used is not specified, they are all conventional products that can be purchased.

A method for identifying the mechanical state of a GIS disconnector based on curve similarity comprises the following steps:

Step 1: Obtain the template of the driving motor power curve during the opening and closing of the GIS disconnector in multiple mechanical states:

The output power of the driving motor of the disconnecting and closing process of the isolating switch in different mechanical states is detected to obtain the power curve in each state; the mechanical state includes the normal state and various defective states;

in,

Jam defect: Fill the gap between the gear and the conductor with a rubber strip to increase the resistance of the gear during rotation, to simulate the spindle jam defect;

Phase-out defect: The static contact is removed to simulate the phase-out defect approximately;

Incomplete stroke defect: The incomplete stroke defect is simulated by disconnecting the power supply of the self-holding circuit;

Contact finger spring deformation defect: The contact finger spring deformation defect is simulated by removing the contact finger spring.

Step 2: Identification of meshing points and translation points based on multi-window slopes:

The meshing point refers to a point on the power curve, which corresponds to the time point when the moving and static contacts of the disconnector touch each other, and when the moving and static contacts touch each other, it will cause the power of the drive motor to change. As shown in Figure 1, the translation point in Figure 1 is the point where the disconnector starts to move:

By extracting the slope of the line segment between a certain point _xi on the curve and the points xi _±aj at different intervals on both sides, the multi-window slope feature K=[k _l1 k _r1 k _l2 k _r2 ...k _ln k _rn ] of the point is obtained, so as to more perfectly describe the slope features on both sides of the point, as shown in Figure 2. Among them, k _l1 is the slope from point _xi to point xi _-a1 ; k _r1 is the slope from point _xi to point _xi+a1 ; k _ln is the slope from point _xi to point xi _-an ; k _rn is the slope from point _xi to point xi _+an . The present invention does not limit the value of n. In Figure 2, n is 3, that is, j is 1 to 3; preferably, the difference between the x-axis values of all two adjacent points is equal.

Due to the differences in intrinsic curves, the slope characteristics on both sides of the meshing point of the travel curve of the same model but different disconnectors must be different. In the identification process, the multi-window slope characteristics of the meshing point are first extracted according to the standard curve (the opening and closing power curve of the drive motor under normal and fault-free conditions, as shown in Figure 5), and then the point with the most similar slope characteristics is found in the curve to be identified as the suspected meshing point. The criteria for similar slope characteristics are:

(1-σ)*k _{i_bz} ≤k _i ≤(1+σ)*k _{i_bz} (1)

Where, k _{i_bz} is the reference slope, which is the i-th element in the multi-window slope feature of the meshing point of the standard curve, σ is the allowable fluctuation range of the slope, and k _i is the i-th element in the multi-window slope feature of the to-be-identified point of the to-be-identified curve. The multi-window slope features of the standard curve and the to-be-identified curve are obtained in the same way.

Assume that the number of multi-window slope features of a certain point that satisfies formula (1) is β, and δ is the preset minimum number that satisfies formula (1). Then the criterion for the suspected meshing point is:

δ can be set according to actual conditions, for example, 3.

Finally, considering special cases, due to abnormal jamming and other deviations, the curve may have pseudo meshing points, so it is necessary to screen according to the physical characteristics of the meshing points. For common linear or angle GIS disconnectors, the meshing point should be close to the standard angle (preferably ±5% of the standard angle) and be a local minimum, so that the meshing point can be finally identified. The meshing point identification method flow of GIS disconnector is shown in Figure 3. The separation point is identified in the same way.

Step 3: Calculate the curve similarity based on Fréchet distance:

After identifying the meshing point information of the power curve of the GIS disconnector drive motor, the power curve of the GIS disconnector opening and closing process can be divided into the starting stage, the translation stage and the meshing stage. The curve similarity analysis is performed in each stage, and the dynamic process reflected in the GIS disconnector opening and closing process is interpreted according to the similarity to complete the mechanical state diagnosis of the GIS disconnector. The curve similarity analysis method adopted by the present invention is the Fréchet distance analysis method. Its principle is shown in Figure 4.

The specific method is:

There is a quantifiable spatial distance between two curves. Suppose discrete sequence P = [p ₁ , p ₂ , ... p _m ], discrete sequence Q = [q ₁ , q ₂ , ... q _n ], there are the following sequence point pairs between the two sequences:

(p _a1 , q _b1 ), (p _a2 , q _b2 ),…, (p _at , q _at )

Where a ₁ = 1, b ₁ = 1, a _t = m, b _t = n, for any i = 1, 2, ..., n, there exists a _i+1 = a _i or a _i+1 = a _i + 1 and b _i+1 = b _i . The length of the point pair between P and Q is defined as:

Where d is the Euclidean distance.

Then the Fréchet distance between discrete sequences P and Q is defined as:

DF(P,Q)=min||D||

In the curve similarity calculation method based on Fréchet distance, the specific implementation method is as follows:

(1) The template curve is expressed as

S＝[S ₁ , S ₂ ,…, S _m ,…, S _M ]

Wherein, S _m =(x _m ,y _m ), x, y represent the coordinates of the discrete sampling point, m=1, 2, ..., M, represents the sequence number of the sampling point in the template sequence S;

(2) The sample curve to be identified is

T＝[T ₁ , T ₂ ,…, T _n ,…, T _N ]

Wherein, T _n =(x _n , _yn ), x, y represent the coordinates of discrete sampling points, n = 1, 2, ..., N, represents the sequence number of the sampling point in the test sequence T;

(3) Calculate the distance between each discrete point on the template sequence S and the test sequence T

The distance is calculated using the Euclidean distance method:

Finally, the distance matrix between sequences is obtained

Wherein, d ₁₁ is the value of d(S ₁ ,T ₁ ); d _MN is the value of d(S _M ,T _N );

(4) Find the shortest distance dmin and the longest distance dmax in the distance matrix, initialize the target Fréchet distance to DF = dmin, and set the loop step size to

Convert the original distance matrix into a 0-1 binary matrix D _0-1

Where:

(5) At this time, the Fréchet distance calculation problem is transformed into the problem of whether there is a path connecting the upper left corner element and the lower right corner element in the binary matrix. Search for 8-connected elements in the binary matrix. When the elements have 8-connectivity, such elements are classified into one category and numbered to obtain the label matrix L. When L(1, 1) = L(M, N) ≠ 0, it means that there is a path connecting the upper left corner and the lower right corner in the matrix;

L(1,1) is the distance between the first points of two curves, the first point of the first line and the first point of the second line; L(M,N) is the distance between the last points of two curves, the mth point of the first line and the nth point of the second line.

(6) If there is no such path in (5), then DF = DF + ST, and repeat (4) (5). If there is a connected path or DF = dmax, the Fréchet distance between the discrete sequences P and Q is DF.

The smaller the Fréchet distance is, the more similar the two curves are; conversely, the larger the Fréchet distance is, the less similar the two curves are.

During the closing and opening process, the GIS disconnector motor power curve is significantly different. Under ideal normal conditions, the typical closing and opening motor power curves are as follows:

For the curve shape, there is only one Fréchet distance for the complete sequence, which cannot accurately describe the similarity of the curve shape. It is necessary to further separate the discrete sequences of the motor power curve at different stages from the ideal normal sequence to calculate the Fréchet distance to judge the similarity. Based on the multi-window slope recognition results, the power curve is divided as follows:

Table 1 Sequence classification types

Complete sequence Sequence 1 Sequence 2 Sequence 3 Closing sequence Start-up phase Translation phase Meshing stage Opening sequence Start-up phase Separation phase Translation phase

After the complete sequence is divided, the test sequence and template sequence at different stages are matched to eliminate the influence between different stages of the curve and retain the curve information that is beneficial to state recognition to the greatest extent. The matching process design is shown in Figure 6.

Step 4: Identification of the mechanical status of the GIS disconnector

Whether the equipment is operating normally can be determined based on the driving motor power curve during the opening and closing process of the GIS disconnector. Defects such as jamming and misalignment can be determined based on the translation and engagement stages of the closing curve and the separation and translation stages of the opening curve. Defects such as abnormal jamming can be determined based on the translation stages of the closing curve and the opening curve.

During the closing process, the GIS disconnector may have stuck, abnormal stuck, and misaligned defects. The test curves under the three defective states are shown in Figure 7. The normal state curve enters the translation stage after a rapid decline, and there is an obvious peak in the meshing stage; the translation stage and meshing stage of the stuck state curve are obviously elevated; the misaligned state curve is similar to the stuck state curve, but there are two peaks in the meshing stage, and the elevation in the translation stage is slightly lower; the abnormal stuck state curve fluctuates significantly, and there is an obvious elevation in the translation stage. According to the curve shape, the meshing stage of the stuck state curve and the misaligned state curve are similar, and the meshing stage of the abnormal stuck state curve and the normal state curve are similar. Using the curve of this stage alone will result in a small difference in the Freche distance, resulting in misjudgment. In the identification process, the meshing stage should be identified first, and divided into normal and abnormal stuck and stuck, misaligned; then the translation stage should be identified to separate the two types of faults again.

After a large number of tests and actual measurement data verification, the method of the present invention has good practicability, high accuracy and is easy to promote and apply.

The above shows and describes the basic principles, main features and advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims (6)

1. A method for identifying the mechanical state of a GIS disconnector based on curve similarity, characterized in that it comprises the following steps: Step 1: Acquisition of a template of the power curve of the driving motor during the opening and closing process of the GIS disconnector in multiple mechanical states: Detection of the output power of the driving motor during the opening and closing process of the disconnector in different mechanical states is performed to obtain the power curve in each state as a template; the mechanical state includes a normal state and a defective state; Step 2: Identification of meshing points and translation points based on multi-window slopes: Collect the power curve of the driving motor of the GIS disconnector to be identified, and identify its engagement point or disengagement point, as well as the translation point; Step 3: Calculate the curve similarity based on Fréchet distance: The power curve of the drive motor to be identified and the closing curve in the power curve under each state are divided into a starting stage, a translation stage and a meshing stage according to their respective translation points and meshing points; The opening curve is divided into the starting stage, the separation stage and the engagement stage according to the respective translation points and separation points; Calculate the Fleche distances between the power curve of the drive motor to be identified and the corresponding stages of the power curve of the drive motor in each state in step 1. If the Fleche distance between a certain stage of the power curve of the drive motor to be identified and the power curve of the drive motor in a certain state is the smallest among all the Fleche distances calculated in the stage, then it is determined that the stage is similar to the state, that is, the stage is determined to be the state; Step 4: Identification of the mechanical status of the GIS disconnector: If the power curve of the driving motor to be identified is a closing curve, the meshing stage is determined first, and then the translation stage is determined; that is, the corresponding category is entered according to the state determined in the meshing stage, and then the final mechanical state of the GIS disconnector is obtained according to the state determined in the translation stage; If the power curve of the driving motor to be identified is a tripping curve, first determine the separation stage, and then determine the translation stage; that is, enter the corresponding category according to the state determined in the separation stage, and then obtain the final GIS disconnector mechanical state according to the state determined in the translation stage. 2. The GIS disconnector mechanical state identification method based on curve similarity according to claim 1 is characterized in that: during the detection in step 1, a rubber strip is filled between the gear and the conductor to simulate the main shaft jam defect; the static contact is removed to simulate the phase loss defect; the self-holding loop power supply is disconnected to simulate the stroke inadequate defect; the contact finger spring is removed to simulate the contact finger spring deformation defect. 3. The GIS disconnector mechanical state identification method based on curve similarity according to claim 1 is characterized in that: in step 2, the specific method of identifying the meshing point is: Obtain multi-window slope features of the standard curve and the curve to be identified; Find the point in the curve to be identified whose slope characteristics are most similar to the slope characteristics of the meshing point in the standard curve as the suspected meshing point; Whether the rotation angle of the meshing point is within ±5% of the standard rotation angle, if so, and it is a local minimum, then this point is the meshing point finally identified. 4. The GIS disconnector mechanical state identification method based on curve similarity according to claim 3 is characterized in that: the multi-window slope feature is obtained in the following manner: By extracting the slope of the line segment between a point _xi on the curve and points xi _±aj with different spacing on both sides, the multi-window slope feature K = [k _l1 k _r1 k _l2 k _r2 …k _ln k _rn ] of the point is obtained; among them, k _l1 is the slope from point _xi to point xi _-a1 ; k _{r1 is} the slope from point xi to point _xi+a1 ; k _l2 is the slope from point _xi to point xi _-a2 ; k _r2 is the slope from point _xi to point _xi+a2 ; k _ln is the slope from point _xi to point _xi-an ; k _rn is the slope from point _xi to point xi _+an . 5. The GIS disconnector mechanical state identification method based on curve similarity according to claim 3 is characterized in that: when finding a point in the curve to be identified whose slope characteristic is most similar to the slope characteristic of the meshing point in the standard curve as a suspected meshing point, the criterion for the similarity of the slope characteristics is: (1-σ)*k _{i_bz} ≤k _i ≤(1+σ)*k _{i_bz} (1) Where, k _{i_bz} is the i-th element in the multi-window slope feature of the meshing point of the standard curve, σ is the allowable fluctuation range of the slope, and k _i is the i-th element in the multi-window slope feature of the to-be-discriminated point of the to-be-identified curve; Assume that the number of multi-window slope features of a certain point that satisfies formula (1) is β, and δ is the preset minimum number that satisfies formula (1). Then the criterion for the suspected meshing point is:

6. The GIS disconnector mechanical state identification method based on curve similarity according to claim 1 is characterized in that: in step 3, the specific calculation method of the Fréchet distance is: (1) The template curve at a certain stage is expressed as S＝[S ₁ , S ₂ ,…, S _m ,…, S _M ] Wherein, S _m =(x _m ,y _m ), x, y represent the coordinates of discrete sampling points, m=1, 2, ..., M, represents the sequence number of the sampling point in the sequence S; (2) The curve to be identified corresponds to a certain stage: T＝[T ₁ , T ₂ ,…, T _n ,…, T _N ] Wherein, T _n =(x _n , _yn ), x, y represent the coordinates of discrete sampling points, n = 1, 2, ..., N, represents the sequence number of the sampling point in the sequence T; (3) Calculate the distance between each discrete point on the template sequence S and the test sequence T:

Finally, we get the distance matrix between sequences:

Wherein, d ₁₁ is the value of d(S ₁ ,T ₁ ); d _MN is the value of d(S _M ,T _N ); (4) Find the shortest distance dmin and the longest distance dmax in the distance matrix. The target Fréchet distance is initialized to DF = dmin, and the loop step size is set to:

Convert the original distance matrix into a 0-1 binary matrix D _0-1 :

Where:

(5) Search for 8-connected elements in the binary matrix. When the elements have 8-connectivity, such elements are classified into one category and numbered to obtain the label matrix L. When L(1, 1) = L(M, N) ≠ 0, it means that there is a path connecting the upper left corner and the lower right corner in the matrix; (6) If there is no such path in (5), then DF = DF + ST, and repeat (4) (5). If there is a connected path or DF = dmax, DF at this time is the final Fréchet distance.

Images