1. A method for quantitatively evaluating a degree of a failure of an electrical device, the method comprising:

constructing an equipment fault severity evaluation model;

calculating a multivariate estimation value representing the health state of the equipment by using the severity evaluation model;

acquiring a measuring point actual value, and calculating the severity of the measuring point fault through the measuring point actual value and the measuring point estimated value;

and calculating the feature fault severity, the fault mode severity, the component fault severity and the equipment fault severity step by step on the basis of the measured point fault severity.

2. The method for quantitatively evaluating the degree of fault of the power equipment according to claim 1, wherein the constructing of the model for evaluating the severity of the fault of the power equipment specifically comprises:

acquiring a historical data sample of a measuring point of the power equipment, cleaning the sample, removing the sample with bad points and null values, and removing abnormal points by using a Lauda rule;

screening the equipment measuring point variables by using a PCA dimension reduction method, acquiring historical data of the equipment measuring point variables as samples, removing highly similar samples by using a Pearson correlation analysis method, reducing the number of samples, and improving the calculation efficiency on the premise of ensuring the integrity of data;

and carrying out interval mapping on the sample data values to obtain a full-working-condition matrix model representing the health state of the equipment.

3. The method according to claim 1, wherein the calculating the multivariate estimation value representing the health status of the equipment comprises:

calculating an estimated value of a measuring point in a healthy state by using an observation vector consisting of data of the measuring point of equipment at any moment and the similarity between the observation vector and a full-working-condition healthy state matrix model, wherein a calculation formula of the estimated value is as follows:

wherein, X_estIs the estimated value of the health state of the equipment, D is the matrix of the health state of the equipment under all working conditions, W is the weight vector, X_obsIs an observation vector of the device, D^TIs the transpose of matrix D.

4. The method for quantitatively evaluating the fault degree of the power equipment as claimed in claim 1, wherein the method for calculating the severity of the station fault comprises the following steps:

calculating the mapping point abscissa of a basic point of a point to be measured of the equipment on an inverse function coordinate curve of a Sigmoid function, wherein the basic point comprises a low report point, a low extreme point, an estimated value point, a high extreme point and a high report point;

defining variable vectors of a plurality of thresholds of a point to be measured of the equipment, and mapping the variable vectors to the abscissa by using a linear interpolation method to obtain an interpolation result, wherein the plurality of thresholds comprise a negative protection point, a negative value point, an estimated value point, a positive value point and a positive protection point;

and calculating an interpolation function value of the actual measurement value of the point to be measured, calculating a Sigmoid inverse function value of the interpolation function value, and mapping to a coordinate system to obtain a measuring point severity curve.

5. The method for quantitatively evaluating the degree of failure of an electric power equipment according to claim 1, wherein the method for calculating the severity of the characteristic failure comprises:

calculating a difference value between an actual value and an estimated value of a point to be measured, comparing the difference value with an early warning threshold value, and if the difference value is smaller than the early warning threshold value, the severity of the characteristic fault is 0;

if the difference is larger than the early warning threshold, the characteristic fault severity is the maximum value of the fault severity of the measuring point in the direction corresponding to the point to be measured.

6. The method for quantitatively evaluating the degree of failure of an electric power equipment according to claim 1, wherein the method for calculating the severity of the failure mode comprises:

acquiring all characteristic fault severity combinations of points to be measured, arranging the combinations according to the magnitude sequence, and then utilizing a formula: failure mode severity i (max) +1/2i (sec) +1/4i (thi)

And calculating the severity of the fault pattern, wherein I (max) is the maximum value of the severity of the characteristic fault of the point to be measured, I (sec) is the second value of the severity of the characteristic fault of the point to be measured, and I (thi) is the third value of the severity of the characteristic fault of the point to be measured.

7. The method for quantitatively evaluating the degree of failure of an electric power equipment according to claim 1, wherein the method for calculating the severity of failure of the component comprises:

acquiring all fault mode severity contained in the component, calculating the severity of each fault mode, sequencing all fault mode severity, and combining the first three items to form the fault severity of the component;

component failure severity n (max) +1/2n (sec) +1/4n (thi)

Where n (max) is the maximum failure mode failure severity, n (sec) is the second failure mode failure severity, and n (thi) is the third failure mode failure severity.

8. The method for quantitatively evaluating the degree of the fault of the power equipment as claimed in claim 1, wherein the method for calculating the severity of the fault of the equipment comprises the following steps:

acquiring all components under the power plant equipment to be tested, and reading the component fault severity of all the components;

sorting the component fault severity according to size, and taking the first three items, then:

equipment failure severity m (max) +1/2m (sec) +1/4m (thi)

Where M (max) is the maximum component failure severity, M (sec) is the second component failure severity, and M (thi) is the third component failure severity.

Background

As an advanced productivity and basic industry, the power industry plays an important role in promoting the development of national economy and social progress, and meanwhile, a power plant equipment system for transmitting power is a large and complex system and has the characteristics of high failure rate and high failure hazard. The forced outage rate and the accident rate of many electric generating sets are high due to the aspects of design, manufacture, installation, maintenance, management, operation and the like.

The traditional power plant equipment state evaluation mainly depends on the subjective experience of equipment managers, analysis and diagnosis cannot be carried out according to massive monitoring data of equipment, and unit hidden dangers cannot be found and processed in time; or the operation parameters of a certain part are monitored and paid attention to by depending on the operation of the power plant equipment, parameter alarm is carried out according to parameter values, then fault qualitative analysis is carried out on the equipment, the fault severity or the deterioration trend of the whole equipment is less concerned, the fault rate of the diagnosis result of the equipment is high, and the operation state of the power generation equipment cannot be mastered carefully.

Disclosure of Invention

The embodiment of the invention aims to provide a quantitative evaluation method for the fault degree of power equipment, which is used for solving the problem that the fault early warning of the existing power industry equipment cannot be quantitatively evaluated, so that the early warning efficiency is low.

In order to achieve the above object, the embodiments of the present invention mainly provide the following technical solutions:

in a first aspect, embodiments of the present invention provide a method for quantitatively evaluating a fault level of an electrical device,

the method comprises the following steps: constructing an equipment fault severity evaluation model; calculating a multivariate estimation value representing the health state of the equipment by using the severity evaluation model according to a similarity principle; acquiring a measuring point actual value, and calculating the severity of the measuring point fault through the measuring point actual value and the measuring point estimated value; and calculating the feature fault severity, the fault mode severity, the component fault severity and the equipment fault severity step by step on the basis of the measured point fault severity.

Further, the constructing of the equipment fault severity assessment model specifically includes: acquiring a historical data sample of a measuring point of the power equipment, cleaning the sample, removing the sample with bad points and null values, and removing abnormal points by using a Lauda rule; screening the equipment measuring point variables by using a PCA dimension reduction method, acquiring historical data of the equipment measuring point variables as samples, removing highly similar samples by using a Pearson correlation analysis method, reducing the number of samples, and improving the calculation efficiency on the premise of ensuring the integrity of data; and carrying out interval mapping on the sample data values to obtain a full-working-condition matrix model representing the health state of the equipment.

Further, the calculating the multivariate estimation value representing the health state of the equipment specifically comprises: calculating an estimated value of a measuring point in a healthy state by using an observation vector consisting of data of the measuring point of equipment at any moment and the similarity between the observation vector and a full-working-condition healthy state matrix model, wherein a calculation formula of the estimated value is as follows:

wherein, X_estIs the estimated value of the health state of the equipment, D is the matrix of the health state of the equipment under all working conditions, W is the weight vector, X_obsIs an observation vector of the device, D^TIs the transpose of matrix D.

Further, the method for measuring the point fault severity comprises the following steps: calculating the mapping point abscissa of a basic point of a point to be measured of the equipment on an inverse function coordinate curve of a Sigmoid function, wherein the basic point comprises a low report point, a low extreme point, an estimated value point, a high extreme point and a high report point; defining variable vectors of a plurality of thresholds of a point to be measured of the equipment, and mapping the variable vectors to the abscissa by using a linear interpolation method to obtain an interpolation result, wherein the plurality of thresholds comprise a negative protection point, a negative value point, an estimated value point, a positive value point and a positive protection point; and calculating an interpolation function value of the actual measurement value of the point to be measured, calculating a Sigmoid inverse function value of the interpolation function value, and mapping to a coordinate system to obtain a measuring point severity curve.

Further, the method for calculating the severity of the feature fault comprises the following steps: calculating a difference value between an actual value and an estimated value of a point to be measured, comparing the difference value with an early warning threshold value, and if the difference value is smaller than the early warning threshold value, the severity of the characteristic fault is 0; if the difference is larger than the early warning threshold, the characteristic fault severity is the maximum value of the fault severity of the measuring point in the direction corresponding to the point to be measured.

Further, the method for calculating the severity of the failure mode comprises the following steps: acquiring all characteristic fault severity combinations of points to be measured, arranging the combinations according to the magnitude sequence, and then utilizing a formula: and calculating the severity of the fault pattern, wherein I (max) is the maximum value of the severity of the characteristic fault of the point to be measured, I (sec) is the second value of the severity of the characteristic fault of the point to be measured, and I (thi) is the third value of the severity of the characteristic fault of the point to be measured.

Further, the method for calculating the severity of the component fault comprises the following steps: acquiring all fault mode severity contained in the component, calculating the severity of each fault mode, sequencing all fault mode severity, and combining the first three items to form the fault severity of the component;

component failure severity n (max) +1/2n (sec) +1/4n (thi)

Where n (max) is the maximum failure mode failure severity, n (sec) is the second failure mode failure severity, and n (thi) is the third failure mode failure severity.

Further, the method for calculating the severity of the equipment fault comprises the following steps: acquiring all components under the power plant equipment to be tested, and reading the component fault severity of all the components; sorting the component fault severity according to size, and taking the first three items, then:

equipment failure severity m (max) +1/2m (sec) +1/4m (thi)

Where M (max) is the maximum component failure severity, M (sec) is the second component failure severity, and M (thi) is the third component failure severity.

The technical scheme provided by the embodiment of the invention at least has the following advantages:

the embodiment of the invention combines a mathematical prediction model with power equipment data and services, calculates multivariable predicted values representing equipment states based on a similarity principle to obtain the severity of measured point faults, the severity of characteristic faults, the severity of fault modes, the severity of component faults and the severity of equipment faults, provides a set of complete quantitative evaluation system through multi-layer step-by-step calculation, can realize the conversion of the health degree of equipment in the power industry from qualitative evaluation to quantitative evaluation, enables power plant practitioners to more carefully master the running state of power generation equipment, and can discover problems and solve problems more quickly. The fault degree of all power equipment of the power plant can be evaluated on line in real time, faults can be found conveniently and timely, the fault is located accurately, maintenance plans are arranged reasonably according to the fault degree, and fine management of equipment maintenance is achieved.

Drawings

Fig. 1 is a flowchart of a quantitative evaluation method for a fault degree of an electrical device according to an embodiment of the present invention.

Fig. 2 is a severity curve corresponding to an actual value of a temperature measurement point of a bearing of a blower provided by an embodiment of the present invention.

Detailed Description

The following description of the embodiments of the present invention is provided for illustrative purposes, and other advantages and effects of the present invention will become apparent to those skilled in the art from the present disclosure.

In the following description, for purposes of explanation and not limitation, specific details are set forth such as particular system structures, interfaces, techniques, etc. in order to provide a thorough understanding of the present invention. It will be apparent, however, to one skilled in the art that the present invention may be practiced in other embodiments that depart from these specific details. In other instances, detailed descriptions of well-known systems, circuits, and methods are omitted so as not to obscure the description of the present invention with unnecessary detail.

The embodiment of the invention provides a quantitative evaluation method for the fault degree of power equipment, and with reference to fig. 1, the method mainly comprises the following steps:

s1, constructing an equipment fault severity evaluation model;

the method specifically comprises the steps of obtaining a historical data sample of a measuring point of the power equipment, cleaning the sample, removing the sample with bad points and null values, and removing abnormal points by using a Lauda rule; screening the equipment measuring point variables by using a PCA dimension reduction method, acquiring historical data of the equipment measuring point variables as samples, removing highly similar samples by using a Pearson correlation analysis method, reducing the number of samples, and improving the calculation efficiency on the premise of ensuring the integrity of data; and carrying out interval mapping on the sample data values to obtain a full-working-condition matrix model representing the health state of the equipment.

The method is implemented by taking a blower of a power plant as an implementation object, specifically, measuring point historical data associated with the blower is obtained from a database of the power plant, and a measuring point can be any part of the blower or the measured value and the working state of a part, and the like. Sample data is screened out through data preprocessing and feature selection, in the embodiment, the sample data selects 27 features of a blower and 265 pieces of multi-time data, the features comprise unit load, total air supply volume of the blower, bearing temperature of the blower and the like, the value of each feature is mapped in a (0, 1) interval, and a full working condition matrix model D (27 x 265) representing the health state of equipment is obtained, and the data structure of D is as follows:

s2, calculating a multivariate estimation value representing the health state of the equipment by using the severity evaluation model according to a similarity principle;

the method specifically comprises the following steps: the method comprises the following steps of calculating an estimated value of a measuring point under a healthy state by using an observation vector of the measuring point of equipment at any moment and the similarity between the observation vector and a full-working-condition matrix model, wherein a calculation formula of the estimated value is as follows:

wherein, X_estIs the estimated value of the health state of the equipment, D is the full-working-condition matrix of the equipment, W is the weight vector, X_obsIs an observation vector of the device, D^TIs the transpose of matrix D.

The model is used for calculating the deviation between the real-time value of the fan bearing temperature measuring point and the corresponding estimated value on line, the real-time data of 27 measuring points of the blower are sequentially input into the model to obtain the corresponding estimated value of each measuring point, the bearing temperature measuring point value X _ brg _ obs and the estimated value X _ brg _ est are mainly concerned, and a plurality of observed values of the bearing temperature measuring point of the blower are as follows:

X_brg_obs＝[27.327，27.9107，28.4944，29.1786，40.021，46.1017，43.9362，40.5169，44.005，49.3169，55.697，55.259，54.3452，53.4313，52.5175，52.2685，53.8086，53.7587，53.7926，53.8462，53.7859，53.5803，53.3752，53.17，52.92，52.74，52.6376，56.817，57.1728，57.5979，58.8376，58.4172，57.2601，56.1029，55.9594，54.9942，53.9121，54.502，53.3387，51.4298，48.4798，44.4475，42.7492，41.1446，36.1345，37.4038，35.3828，34.6307]。

the temperature estimation value of the blower bearing calculated by the model is processed by inverse normalization as follows:

X_brg_est＝[33.175742，32.07001，33.826606，34.19302，35.8048，36.620419，35.473246，37.933709，38.099515，37.541385，37.053043，36.682271，36.512462，36.300227，35.924912，36.231769，36.776485，36.710903，36.723298，36.830576，36.821398，36.833325，36.379439，36.514388，35.771201，38.584022，39.234339，39.359326，40.311301，40.171115，40.183486，39.799297，39.392914，39.324046，38.183132，38.7319，38.438631，38.035186，36.677627，36.320003，36.517968，35.841866，35.35513，35.483625，35.247917，36.400716]。

the deviation of the actual value from the estimated value is:

the deviation bias [ -5.8487, -4.1593, -5.3322, -5.0144, 4.2162, 9.4813, 8.4629, 3.9332, 6.4503, 11.3832, 17.5975, 17.7177, 17.2922, 16.7491, 16.005, 15.9683, 17.8837, 17.527, 17.0162, 17.1353, 17.0626, 16.7497, 16.5538, 16.3367, 16.5408, 16.2266, 16.8664, 18.233, 17.9384, 18.2386, 18.5263, 18.2461, 17.0766, 16.3036, 16.5665, 15.6701, 15.7289, 15.7701, 14.9001, 13.3946, 11.8022, 8.1275, 6.2312, 5.3027, 0.7794, 1.9201, 0.1349, -1.77 ].

This deviation bias can be used as a basis for the severity of the failure of the device, see in particular the calculation of the severity of the characteristic failure below.

S3, acquiring the actual value of the measuring point, and calculating the severity of the fault of the measuring point according to the actual value of the measuring point and the estimated value of the measuring point;

specifically, calculating the mapping point abscissa of a basic point of a point to be measured of the equipment on an inverse function coordinate curve of a Sigmoid function, wherein the basic point comprises a low report point, a low extreme point, an estimated value point, a high extreme point and a high report point;

defining variable vectors of a plurality of thresholds of a point to be measured of the equipment, and mapping the variable vectors to the abscissa by using a linear interpolation method to obtain an interpolation result, wherein the plurality of thresholds comprise a negative protection point, a negative value point, an estimated value point, a positive value point and a positive protection point;

and calculating an interpolation function value of the actual measurement value of the point to be measured, calculating a Sigmoid inverse function value of the interpolation function value, and mapping to a coordinate system to obtain a measuring point severity curve.

In a practical example, it may be: 1) defining the mapping points SY of the Y coordinate of the Sigmoid inverse function of the actual fan bearing temperature measuring points, namely { -10, -7, 0, 7, 10} of the low report, the low extreme value, the estimated value point, the high extreme value and the high report, calculating a value SX of an X coordinate corresponding to the Y coordinate point SY of the Sigmoid inverse function,

wherein, Length [ SY ] represents the Length of a longitudinal axis of a SY coordinate point, i represents a subscript value corresponding to a threshold value in a SY list, and x represents an abscissa value of a Sigmoid inverse function corresponding to each threshold value in the SY.

Corresponding SX results were obtained: SX ═ 0.0000453979, 0.000911051, 0.5, 0.999089, 0.999955 }.

Defining a negative protection point and a negative protection point of an actual measuring point, and a variable vector X of an estimated value point, a positive protection point and a positive protection point, wherein X is { X1, X2, X3, X4, X5}, mapping X into SX by using a linear interpolation method, wherein the interpolation formula is as follows:

MM＝Table[X[[i]，SX[i，1，1，2]]，{i，1，Length[X]}]

where i represents the index value of the measured point in vector X and X represents the actual measured point value.

Obtaining an interpolation result: { { x1, 0.0000453979}, { x2, 0.000911051}, { x3, 0.5}, { x4, 0.999089}, and { x5, 0.999955} }.

Substituting the above fan bearing observed value X _ brg _ obs, i.e. the actual value into the interpolation function to obtain: f [ X _ brg _ obs ] ═ 0.002875230.002949890.004810150.006597660.997275330.999938030.999885840.995956090.999618170.99998020.999999520.999999560.999999430.999999210.999998760.999998740.99999960.99999950.999999330.999999370.999999340.999999210.999999110.999998990.99999910.999998920.999999260.999999680.999999610.999999680.999999730.999999680.999999350.999998970.999999120.999998490.999998540.999998580.99999760.999994080.99998460.999860390.999564570.999240160.748592190.936329430.547069260.07741333;

substituting F [ X _ brg _ obs ] into the inverse function of the Sigmoid function, and calculating a Y coordinate value SY corresponding to an X coordinate point SX of the Sigmoid inverse function:

SY＝[-5.848741999999998，-5.823034000000007，-5.332205999999999，-5.014419999999998，5.90268，9.688768600000001，9.077772399999997，5.5064912，7.8701476，10.829914599999999，14.558491000000004，14.630569000000001，14.375294199999997，14.0494174，13.6030228，13.580963800000003，14.7302128，14.516158599999997，14.209668999999998，14.2811782，14.237561199999998，14.0498344，13.932281199999998，13.802005，13.924336600000002，13.735367200000002，14.1198394，14.9397868，14.763076600000002，14.9431444，15.115779400000001，14.947651，14.245968399999999，13.782161799999997，13.939891600000003，13.402092399999997，13.437380800000001，13.46206，12.940041400000002，12.036768399999998，11.081303799999997，8.876498199999999，7.7387391999999995，7.1816403999999965，1.091118，2.6882449999999904，0.1888362000000029，-2.4780224000000075]。

referring to fig. 2, the result is mapped to a coordinate system to obtain a severity curve corresponding to the actual value of the temperature measuring point of the bearing of the blower, wherein the abscissa in the graph is time and the ordinate is severity.

And S4, calculating the feature fault severity, the fault mode severity, the component fault severity and the equipment fault severity step by step on the basis of the measured point fault severity.

Specifically, the calculation method of the severity of the characteristic fault comprises the following steps:

and calculating the difference value between the actual value and the estimated value of the point to be measured, and comparing the difference value with an early warning threshold value, wherein the difference value between the actual value and the estimated value is the deviation bias in the above, and the early warning threshold value is determined by the working range of the measured point data in a normal state and can be the acceptable error range of the measured point data.

If the difference is smaller than the early warning threshold, the severity of the characteristic fault is 0;

if the difference is larger than the early warning threshold, the characteristic fault severity is the maximum value of the fault severity of the measuring point in the direction corresponding to the point to be measured.

Observing the change trend of the severity of the bearing temperature of the blower in the graph 2, wherein the deviation of the bearing temperature measuring point exceeds an alarm threshold value between 0 and 10 moments, a high extreme point appears, the high extreme point is changed into a high report point along with the rise of the temperature, the severity coefficient is larger than 10, after monitoring the alarm, an operator adjusts operation parameters or turns to overhaul, the overhaul operator receives a work order to overhaul, the temperature of the bearing of the blower is gradually reduced to be normal, and the severity coefficient is also reduced to be close to 0. Through online calculation of the severity of the measuring points, the degradation degree of the equipment operation parameters is quantized, and the operation is favorably scheduled and distributed according to the degradation degree of the equipment, and the operation is adjusted or the maintenance or the modification is carried out to improve the performance of the equipment.

Further, the method for calculating the severity of the failure mode comprises the following steps:

acquiring all characteristic fault severity combinations of points to be measured, arranging the combinations according to the magnitude sequence, and then utilizing a formula: failure mode severity i (max) +1/2i (sec) +1/4i (thi)

And calculating the severity of the fault pattern, wherein I (max) is the maximum value of the severity of the characteristic fault of the point to be measured, I (sec) is the second value of the severity of the characteristic fault of the point to be measured, and I (thi) is the third value of the severity of the characteristic fault of the point to be measured, namely calculating the severity of the fault pattern on the basis of the severity of the characteristic fault.

Still further, a component failure mode severity method comprises:

acquiring all the severity of the fault modes contained in the component, calculating the severity of each fault mode, sequencing all the severity of the fault modes, and taking the first three items, then:

component failure severity n (max) +1/2n (sec) +1/4n (thi)

Where n (max) is the maximum failure mode severity, n (sec) is the second failure mode severity, and n (thi) is the third failure mode severity.

Similarly, the method for calculating the severity of the equipment fault comprises the following steps:

acquiring all components under the power plant equipment to be tested, and reading the component fault severity of all the components;

sorting the component fault severity according to size, and taking the first three items, then:

equipment failure severity m (max) +1/2m (sec) +1/4m (thi)

Where m (max) is the maximum component failure severity, m (sec) is the second component failure severity, and m (thi) is the third component failure severity, i.e., the equipment failure severity is calculated based on the component failure severity.

The embodiment of the invention combines a mathematical prediction model with power equipment data and services, calculates the obtained predicted value based on the similarity principle to obtain the severity of the fault of the measuring point, the severity of the characteristic fault, the severity of the fault mode, the severity of the fault of the component and the severity of the fault of the equipment, provides a set of complete quantitative evaluation system through multi-layer step-by-step calculation, can realize the conversion of the health degree of the equipment in the power industry from qualitative evaluation to quantitative evaluation, enables the power plant practitioner to more carefully master the running state of the power generation equipment, and can discover the problems and solve the problems more early. The fault degree of all power equipment of the power plant can be evaluated on line in real time, faults can be found conveniently and timely, the fault is located accurately, maintenance plans are arranged reasonably according to the fault degree, and fine management of equipment maintenance is achieved.

The disclosed embodiments of the present invention provide a computer-readable storage medium having stored therein computer program instructions which, when run on a computer, cause the computer to perform the above-described method.

Those skilled in the art will appreciate that the functionality described in the present invention may be implemented in a combination of hardware and software in one or more of the examples described above. When software is applied, the corresponding functionality may be stored on or transmitted over as one or more instructions or code on a computer-readable medium. Computer-readable media includes both computer storage media and communication media including any medium that facilitates transfer of a computer program from one place to another. A storage media may be any available media that can be accessed by a general purpose or special purpose computer.

The above-mentioned embodiments, objects, technical solutions and advantages of the present invention are further described in detail, it should be understood that the above-mentioned embodiments are only exemplary embodiments of the present invention, and are not intended to limit the scope of the present invention, and any modifications, equivalent substitutions, improvements and the like made on the basis of the technical solutions of the present invention should be included in the scope of the present invention.