CN110006552B - Method for detecting abnormal temperature of equipment - Google Patents

Method for detecting abnormal temperature of equipment Download PDF

Info

Publication number
CN110006552B
CN110006552B CN201910266851.3A CN201910266851A CN110006552B CN 110006552 B CN110006552 B CN 110006552B CN 201910266851 A CN201910266851 A CN 201910266851A CN 110006552 B CN110006552 B CN 110006552B
Authority
CN
China
Prior art keywords
temperature
river
unit equipment
model
stream
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
Application number
CN201910266851.3A
Other languages
Chinese (zh)
Other versions
CN110006552A (en
Inventor
安学利
付婧
郭曦龙
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Tianjin Shuike electromechanical Co.,Ltd.
China Institute of Water Resources and Hydropower Research
Original Assignee
China Institute of Water Resources and Hydropower Research
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by China Institute of Water Resources and Hydropower Research filed Critical China Institute of Water Resources and Hydropower Research
Priority to CN201910266851.3A priority Critical patent/CN110006552B/en
Publication of CN110006552A publication Critical patent/CN110006552A/en
Application granted granted Critical
Publication of CN110006552B publication Critical patent/CN110006552B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01K—MEASURING TEMPERATURE; MEASURING QUANTITY OF HEAT; THERMALLY-SENSITIVE ELEMENTS NOT OTHERWISE PROVIDED FOR
    • G01K1/00—Details of thermometers not specially adapted for particular types of thermometer
    • G01K1/02—Means for indicating or recording specially adapted for thermometers
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01K—MEASURING TEMPERATURE; MEASURING QUANTITY OF HEAT; THERMALLY-SENSITIVE ELEMENTS NOT OTHERWISE PROVIDED FOR
    • G01K13/00—Thermometers specially adapted for specific purposes
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F30/00—Computer-aided design [CAD]
    • G06F30/20—Design optimisation, verification or simulation
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2119/00—Details relating to the type or aim of the analysis or the optimisation
    • G06F2119/08—Thermal analysis or thermal optimisation

Abstract

The invention relates to a method for detecting abnormal temperature of equipment, which comprises the following steps: (1) establishing a unit equipment temperature abnormal state identification model based on WCA and Shepard interpolation based on analysis of unit equipment historical data, wherein the unit equipment historical data comprises temperature state monitoring data of a unit under different environmental factors and operating conditions; (2) acquiring environmental parameters, operating condition parameters and corresponding measured temperature values of the unit equipment under the current working condition, and inputting the environmental parameters and the operating condition parameters of the unit equipment under the current working condition into the abnormal temperature state identification model of the unit equipment to obtain a health standard value output by the model; (3) and comparing the temperature measured value of the unit equipment under the current working condition with the health standard value, and determining the temperature running state of the unit equipment under the current working condition according to the comparison result. The invention can be widely applied to the field of temperature anomaly detection of unit equipment.

Description

Method for detecting abnormal temperature of equipment
Technical Field
The invention relates to a method for detecting temperature anomaly of unit equipment, in particular to a method for detecting temperature anomaly of unit equipment based on Shepard curved surface and water circulation algorithm.
Background
In the operation process of the unit equipment, due to the reasons of surface friction, heat radiation, electricity and the like, the change rule of the temperature of the unit component is complex, and a mathematical model is difficult to establish. In order to obtain the operating temperature state and the development trend of the unit equipment in time and prevent the occurrence of accidents, the influence factors of the temperature parameters need to be fully researched, and a reasonable and reliable unit equipment temperature anomaly detection model is established.
The traditional temperature monitoring alarm value adopts a unified static alarm threshold value, and has two defects: (1) the temperature value of the equipment is alarmed, but the overhaul finds that the temperature is caused by the environmental temperature or the high working rotating speed of the equipment. When the ambient temperature or the rotating speed is reduced, the temperature does not alarm any more. (2) The temperature monitoring value of the equipment has no alarm all the time, but the equipment is found to be seriously damaged during overhaul. This is sufficient to illustrate that conventional static alarms do not reflect well the operational health of the device.
According to the Shepard interpolation model (Shepard interpolation model), one or a combination of several most similar to a prediction object is found out from a historical sample set as a prediction result according to the principle that similar results are generated due to similar reasons, the model has high calculation speed and calculation accuracy, but when the method is applied, model parameters α are often determined manually, and the optimal fitting accuracy cannot be obtained.
Disclosure of Invention
In view of the above problems, an object of the present invention is to provide a method for detecting abnormal temperature of a unit device, which fully considers environmental factors and operating conditions of the unit device, and can effectively perform online state evaluation on the abnormal temperature of the unit device, thereby achieving early warning of the abnormal temperature of the device.
In order to achieve the purpose, the invention adopts the following technical scheme: a method for detecting abnormal temperature of equipment comprises the following steps:
(1) establishing a unit equipment temperature abnormal state identification model based on WCA and Shepard interpolation based on analysis of unit equipment historical data, wherein the unit equipment historical data comprises temperature state monitoring data of a unit under different environmental factors and operating conditions;
(2) acquiring environmental parameters and operating condition parameters of the unit equipment under the current working condition and corresponding temperature measured values r (t), and inputting the environmental parameters and the operating condition parameters of the unit equipment under the current working condition into the unit equipment temperature abnormal state identification model to obtain a health standard value output by the model;
(3) and comparing the temperature measured value r (t) of the unit equipment under the current working condition with the health standard value p (t), and determining the temperature running state of the unit equipment under the current working condition according to the comparison result.
Further, in the step (1), based on analysis of historical data of the unit equipment, the method for establishing the unit equipment temperature abnormal state identification model based on WCA and Shepard interpolation comprises the following steps:
(1.1) analyzing temperature state monitoring data of the unit equipment under different environmental factors and operating conditions, determining the standard health state of the unit equipment, and selecting characteristic parameters capable of reflecting the standard health state of the unit equipment;
(1.2) inputting characteristic parameters of the unit equipment in a standard healthy state under various environmental factors and operating conditions into a Shepard model, and establishing a three-dimensional curved surface model P (f (V, U));
and (1.3) optimizing the built three-dimensional curved surface model P ═ f (V, U) by adopting a WCA algorithm to obtain the optimal result of the parameter α of the Shepard model, and taking the optimized Shepard model as a unit equipment temperature abnormal state identification model.
Further, in the step (1.2), the method for establishing the three-dimensional curved surface model P ═ f (V, U) by inputting the characteristic parameters of the unit equipment in the standard health state under various environmental factors and operating conditions into the Shepard model includes the following steps:
(1.2.1) randomly picking m samples (P) from the characteristic parameterst,Vt,Ut) In (V)t,Ut) As a test sample point, the operating temperature P of the apparatus is measuredtAs response values corresponding to the test sample points, a response value is established from m samples (P)t,Vt,Ut) Forming a m × (2+1) -dimensional matrix, wherein VtIs the ambient temperature of the cabin, UtIs the generator speed, t is 1,2, …, m;
(1.2.2) taking the residual data in the selected characteristic parameters as new sample points (v, u), and calculating response values p (v, u) corresponding to the new sample points (v, u) by adopting a Shepard surface interpolation method based on the m × (2+1) dimensional matrix established in the step (1.2.1);
(1.2.3) based on the new sample points (V, u, p (V, u)) and the m test sample points (Vt,Ut) Of the relation between, calculated to be EuropeEstimation of the response value p (v, u) for the minimum distance E
Figure BDA0002017116900000021
Wherein the response value p (v, u) is compared with the estimated value
Figure BDA0002017116900000022
The mapping relation is the established Shepard model.
Further, in the step (1.3), the WCA algorithm is adopted to optimize the established Shepard model to obtain an optimal result of the parameter α of the Shepard model, and the optimized Shepard model is used as a unit equipment temperature abnormal state identification model, which includes the following steps:
(1.3.1) determining a cost function for calculating a cost function value of each population;
(1.3.2) setting water circulation algorithm control parameters including the number N of the rainfall layerspopTotal number of rivers and oceans NsrMinimum value dmaxAnd the maximum iteration times T of the water circulation algorithm;
(1.3.3) randomly generating an initial population to form an initial stream, a river and an ocean;
(1.3.4) dividing the stream population into a plurality of stream layers as model parameters to be respectively input into the Shepard model, and calculating the cost function value J of each stream layer according to the step (1.3.1)i；
(1.3.5) comparing the value of cost function of each stream layer, and selecting the smallest value JiSelecting N as sea in corresponding stream layerRiver with water-collecting deviceSecond smallest JiThe corresponding stream layer is used as a river, and the number of streams flowing to the specified river and ocean is determined;
(1.3.6) respectively updating the position of the stream flowing to the river, the position of the stream flowing to the ocean and the position of the river flowing to the ocean, and performing position conversion according to the updated cost function value of each stream, each river and each ocean;
(1.3.7) judging whether the evaporation condition is satisfied: if yes, entering step (1.3.8), otherwise entering step (1.3.9);
(1.3.8) if the rivers and streams are close enough to the ocean, adopting different modes to carry out the rainfall process to form new rainfall;
(1.3.9) updating the minimum value of the current iteration;
(1.3.10) judging whether the maximum iteration number is reached, if so, ending the iteration, and outputting the optimal result of the parameter α of the Shepard model, otherwise, returning to the step (1.3.6) until the iteration is ended.
Further, in the step (1.3.6), the method for updating the position of the stream flowing to the river, the position of the stream flowing to the ocean and the position of the river flowing to the ocean respectively and performing position swapping according to the updated cost function value of each stream, river and ocean comprises the following steps:
(1.3.6.1) respectively updating the position of the stream flowing to the river and the position of the stream flowing to the sea;
Figure BDA0002017116900000031
Figure BDA0002017116900000032
in the formula: rand is a random number uniformly distributed between 0 and 1;
Figure BDA0002017116900000033
respectively representing the current positions of the stream, the river and the ocean in the ith iteration process; c is the coefficient of the position update;
(1.3.6.2) inputting the new position of the stream into a Shepard model, calculating a cost function value corresponding to the stream at the moment, and if the cost function of the stream is smaller than that of a river, exchanging the positions of the river and the stream; if the cost function of the stream is smaller than that of the ocean, the ocean and the stream are exchanged;
(1.3.6.3) updating the position of the river towards the ocean;
Figure BDA0002017116900000034
in the formula: rand is a random number uniformly distributed between 0 and 1;
Figure BDA0002017116900000035
respectively representing the current positions of rivers and oceans in the ith iteration process; c is the coefficient of the position update;
(1.3.6.4) inputting the new position of the river into Shepard model, calculating the cost function value corresponding to the river, and if the cost function value of the river is smaller than that of the ocean, the positions of the ocean and the river are exchanged.
Further, in the step (1.3.7), the evaporation conditions are as follows:
Figure BDA0002017116900000036
in the formula (I), the compound is shown in the specification,
Figure BDA0002017116900000037
and
Figure BDA0002017116900000038
the positions of the ocean and the river in the ith iteration process respectively; 1,2, …, Nsr-1；dmaxIs a constant.
Further, in the step (1.3.8), the rainfall process is performed in different modes according to whether the river and stream are close enough to the sea, and the method for forming new rainfall comprises the following steps:
if it is not
Figure BDA0002017116900000041
Or rand<0.1，i＝1，2，…，Nsr-1, the rainfall process is carried out using the following formula, forming new precipitation:
Figure BDA0002017116900000042
in the formula:
Figure BDA0002017116900000043
UB and L B are the upper and lower bounds of the variables, respectively, for the latest position of the newly formed stream;
if it is not
Figure BDA0002017116900000044
i＝1，2，…，NS1The rainfall process is carried out using the following formula, forming new precipitation:
Figure BDA0002017116900000045
in the formula: randn is a normally distributed random number; μ denotes a coefficient of a search area range near the ocean.
Further, in the step (1.3.9), the minimum value is
Figure BDA0002017116900000046
The calculation formula of (2) is as follows:
Figure BDA0002017116900000047
in the formula: t is the maximum iteration number of the water circulation algorithm,
Figure BDA0002017116900000048
the minimum value during the ith iteration,
Figure BDA0002017116900000049
is the minimum value in the (i +1) th iteration.
Further, in the step (3), a calculation formula for comparing the temperature measured value r (t) of the unit equipment under the current working condition with the health standard value p (t) is as follows:
Figure BDA00020171169000000410
in the formula: t represents the operation time of the unit equipment; and w is an early warning threshold value.
Due to the adoption of the technical scheme, the invention has the following advantages: 1. the invention comprehensively considers the influence of the ambient temperature of the unit equipment and the rotating speed of the generator on the equipment temperature, fully utilizes the advantages of the two methods by establishing a state identification model of the abnormal equipment temperature based on WCA and Shepard interpolation, and can effectively carry out online state evaluation on the abnormal unit equipment temperature and realize early warning of the abnormal equipment compared with the defects of no alarm and false alarm caused by adopting a uniform static alarm threshold value for the conventional temperature monitoring alarm value. 2. The equipment temperature abnormity identification model can track the change trend of the state parameters in the running process of the unit equipment in real time, truly and objectively identify the health state of the equipment in a self-adaptive manner, early warn the abnormal state of the equipment in advance, and has good practicability. Therefore, the invention can be widely applied to the field of temperature anomaly detection of unit equipment.
Drawings
FIG. 1 is a diagram illustrating a comparison between an actual measured value and a standard value of an operating temperature of an apparatus according to an embodiment of the present invention;
FIG. 2 is a diagram illustrating an apparatus temperature deviation identification result according to an embodiment of the present invention.
Detailed Description
The invention is described in detail below with reference to the figures and examples.
The invention provides a method for detecting abnormal temperature of unit equipment, which comprises the following steps of firstly selecting a WCA optimization algorithm to adjust Shepard curved surface operation parameters, further obtaining a high-precision fitting model, then calculating a health standard value corresponding to the unit equipment under various environmental factors and operating conditions in real time through the fitting model, and finally obtaining the real operating state of the unit equipment in real time according to the difference between the actual measurement value and the health standard value of the current environmental factor and the operating conditions, wherein the method specifically comprises the following steps:
(1) establishing a unit equipment temperature abnormal state identification model based on WCA and Shepard interpolation based on analysis of unit equipment historical data, wherein the unit equipment historical data comprises temperature state monitoring data of a unit under different environmental factors and operating conditions;
(2) acquiring environmental parameters and operating condition parameters of the unit equipment under the current working condition and corresponding temperature measured values r (t), and inputting the environmental parameters and the operating condition parameters of the unit equipment under the current working condition into the unit equipment temperature abnormal state identification model to obtain a health standard value p (t) output by the model;
(3) and comparing the temperature measured value r (t) of the unit equipment under the current working condition with the health standard value p (t), and determining the temperature running state of the unit equipment under the current working condition according to the comparison result.
The calculation formula when the measured value of the current temperature of the unit equipment is compared with the health standard value is as follows:
Figure BDA0002017116900000051
in the formula: t represents the operation time of the unit equipment; w is an early warning threshold value which is preset to be 50% (which can be properly adjusted according to different unit equipment and different parameters), namely when the measured value exceeds 50% of the health standard value, fault early warning is carried out, so that the abnormal state and early fault of the equipment can be timely discovered.
In the step (1), the method for establishing the unit equipment temperature abnormal state identification model based on the WCA and Shepard interpolation based on the analysis of the historical data of the unit equipment comprises the following steps:
and (1.1) deeply analyzing mass temperature state monitoring data of the unit equipment under different environmental factors and operating conditions, determining the standard health state of the unit equipment, and selecting characteristic parameters capable of reflecting the standard health state of the unit equipment.
And (1.2) inputting characteristic parameters of the unit equipment in a standard health state under various environmental factors and operating conditions into the Shepard model to obtain the Shepard model P ═ f (V, U).
And (1.3) optimizing the established Shepard model by adopting a WCA algorithm to obtain the optimal result of the parameter α of the Shepard model, and taking the optimized Shepard model as a unit equipment temperature abnormal state identification model.
In the step (1.2), the unit equipment is in a standard health state, characteristic parameters of the unit equipment under various environmental factors and operating conditions are input into the Shepard model, and a three-dimensional curved surface model P is established as f (V, U). On the basis of a Shepard curved surface interpolation method, establishing a unit equipment operation temperature standard model P (f, V) according to a standard sample, wherein the temperature standard model comprehensively considers multi-source information such as cabin environment temperature, generator rotating speed and the like, wherein P is equipment temperature, V is cabin environment temperature, and U is generator rotating speed; the modeling method can reflect the environmental factors and the working condition factors (the cabin environmental temperature and the generator rotating speed) influencing the temperature state of the equipment more practically, thereby effectively utilizing the existing normal operation data of the equipment. Specifically, the method for establishing the temperature standard model P ═ f (V, U) of the unit equipment comprises the following steps:
(1.2.1) randomly selecting m samples (P) from the characteristic parameters selected in the step 1.2)t,Vt,Ut) In (V)t,Ut) As a test sample point, the operating temperature P of the apparatus is measuredtAs response values corresponding to the test sample points, a response value is established from m samples (P)t,Vt,Ut) Forming a m × (2+1) -dimensional matrix, wherein VtIs the ambient temperature of the cabin, UtIs the generator speed, t is 1,2, …, m:
Figure BDA0002017116900000061
(3.2) taking the residual data in the selected characteristic parameters as new sample points (v, u), and calculating response values p (v, u) corresponding to the new samples (v, u), namely the temperatures of the corresponding devices by adopting a Shepard curved surface interpolation method based on the m × (2+1) dimensional matrix established in the step (1.2.1).
(3.3) based on the new sample points (V, u, p (V, u)) and the m test sample points (V)t,Ut) The estimated value of the response value p (v, u) is calculated
Figure BDA0002017116900000062
Minimizing Euclidean distance E, response value p (v, u) and estimation value
Figure BDA0002017116900000063
The mapping relationship of (a) to (b),namely the temperature standard model of the equipment.
Wherein the Euclidean distance E is as follows:
Figure BDA0002017116900000064
in the formula, ωtNew sample points (V, u) and test sample points (V) are represented as weightst,Ut) Estimation of response values
Figure BDA0002017116900000065
The size of the contribution of (a) to (b),
Figure BDA0002017116900000066
the parameter α is optimized by WCA method to make Shepard model reach the best fitting accuracy, gammat＝[(v-Vt)2+(u-Ut)2]0.5Indicates the new sample point (V, u) and the test sample point (V)t,Ut) The distance between them.
Calculating an estimated value of a response value p (v, u) of a new sample point (v, u) that minimizes the Euclidean distance E
Figure BDA0002017116900000067
Namely:
to satisfy
Figure BDA0002017116900000068
To Euclidean distance
Figure BDA0002017116900000069
Taking the derivative and let it be 0, i.e.:
Figure BDA00020171169000000610
calculating the formula (5) to obtain a new sample point (v, u) and an estimated value
Figure BDA00020171169000000611
A mapping relation between them, i.e. devicesThe temperature standard model is:
Figure BDA0002017116900000071
in the step (1.3), the WCA algorithm is adopted to optimize the established three-dimensional curved surface model to obtain the optimal parameters α of the Shepard curved surface model, so that the Shepard model achieves the optimal regression performance, and the method comprises the following steps:
(1.3.1) determining a cost function for calculating a cost function value for each population.
Figure BDA0002017116900000072
In the formula: n is the number of the acquired data; y isiThe ith real value is actually output;
Figure BDA0002017116900000073
the ith regression value output for the regression model.
(1.3.2) setting water circulation algorithm control parameters including the number N of the rainfall layerspopTotal number of rivers and oceans NsrInitial minimum value dmaxMaximum iteration times T of water circulation algorithm and optimization parameter number Nvar＝1。
(1.3.3) determining the identification range of the parameter to be identified α, and randomly generating an initial population to form an initial stream (raindrops), river and sea, wherein the identification range of the parameter to be identified α is determined empirically.
Wherein, the calculation formula of the initial stream, river and ocean is as follows:
Figure BDA0002017116900000074
Nsr＝Nriver with water-collecting device+1 (9)
NStream＝Npop-Nsr(10)
In the formula: n is a radical ofvarFor searching the dimension of the space, i.e. one NvarDimension optimization questionsTitle to be obtained; n is a radical ofRiver with water-collecting deviceThe number of rivers; n is a radical ofsrRiver and ocean totals; n is a radical ofStreamThe number of streams.
(1.3.4) dividing the stream population into a plurality of raindrop layers as model parameters to be input into the Shepard model, taking various environmental factors and operating conditions of the unit equipment in a standard health state as input signals of the Shepard model to obtain a simulation output quantity result of the Shepard model, and calculating a cost function value J of each raindrop layer according to the step (1.3.1) based on the obtained simulation output quantity result and actual temperature data of the unit equipmenti. Wherein, the expression of each raindrop layer is as follows:
Figure BDA0002017116900000075
(1.3.5) comparing the magnitude of the cost function value of each raindrop layer, and selecting the smallest cost function value JiSelecting N as sea according to raindrop layerRiver with water-collecting deviceSecond smallest JiAnd the corresponding raindrop layer is used as a river, and the number of streams flowing to the specified river and the ocean is determined.
Figure BDA0002017116900000081
In the formula: NS (server)nFor the number of streams flowing to a particular river or ocean, round { f } is rounded to an integer value of f, N is 1,2, …, Nsr。
(1.3.6) respectively updating the position of the stream flowing to the river, the position of the stream flowing to the ocean and the position of the river flowing to the ocean, and carrying out position conversion according to the updated cost function value of each stream, each river and each ocean.
Specifically, the method comprises the following steps:
(1.3.6.1) updating the position of stream flow to river and the position of stream flow to sea respectively.
Figure BDA0002017116900000082
Figure BDA0002017116900000083
In the formula: rand is a random number uniformly distributed between 0 and 1;
Figure BDA0002017116900000084
respectively representing the current positions of the stream, the river and the ocean in the ith iteration process; c is the coefficient of the location update, taking the empirical value of 2.
(1.3.6.2) inputting the new position of the stream into the Shepard model, taking various environmental factors and operating conditions of the unit equipment in a standard health state as input signals of the Shepard model to obtain a simulation output result of the Shepard model, and calculating a cost function value corresponding to the stream at the moment according to the step (1.3.1) based on the obtained simulation output result and actual temperature data of the unit equipment. If the cost function of the stream is smaller than that of the river, the positions of the river and the stream are exchanged; and if the cost function of the stream is smaller than that of the ocean, the ocean and the stream are exchanged.
(1.3.6.3) updating the position of the river flow to the ocean.
Figure BDA0002017116900000085
In the formula: rand is a random number uniformly distributed between 0 and 1;
Figure BDA0002017116900000086
respectively representing the current positions of rivers and oceans in the ith iteration process; c is the coefficient of the location update, taking the empirical value of 2.
(1.3.6.4) inputting the new position of the river into the Shepard model, taking various environmental factors and operating conditions of the unit equipment in a standard health state as input signals of the Shepard model to obtain a simulation output quantity result of the Shepard model, calculating a cost function value corresponding to the river at the moment by using a formula (7) based on the obtained simulation output quantity result and actual temperature data of the unit equipment, and exchanging the positions of the sea and the river if the cost function value of the river is smaller than that of the sea.
(1.3.7) judging whether the evaporation condition is satisfied: if yes, go to step (1.3.8), otherwise go to step (1.3.9).
Wherein, the evaporation conditions are as follows:
Figure BDA0002017116900000091
in the formula (I), the compound is shown in the specification,
Figure BDA0002017116900000092
and
Figure BDA0002017116900000093
the positions of the ocean and the river in the ith iteration process respectively; 1,2, …, Nsr-1；dmaxIs a very small constant.
(1.3.8) depending on whether the river and stream are close enough to the ocean, the rainfall process is carried out in different ways, and the method for forming new rainfall is as follows:
if it is not
Figure BDA0002017116900000094
Or rand<0.1，i＝1，2，…，Nsr-1, performing a rainfall event using equation (16), forming new precipitation.
Figure BDA0002017116900000095
In the formula:
Figure BDA0002017116900000096
UB and L B are the upper and lower bounds of the variables, respectively, for the latest position of the newly formed stream, and rand is as defined above.
If it is not
Figure BDA0002017116900000097
i＝1，2，…，NS1And (5) carrying out a rainfall process by using the formula (17) to form new rainfall.
Figure BDA0002017116900000098
In the formula: randn is a normally distributed random number; mu represents the coefficient of the search area range near the ocean, and the smaller mu is, the closer the search range is to the ocean (optimal solution), and generally mu is 0.1.
(1.3.9) updating the minimum value of the current iteration by adopting a formula (18), wherein the updating formula is as follows:
Figure BDA0002017116900000099
in the formula: t is the maximum iteration number of the water circulation algorithm,
Figure BDA00020171169000000910
the minimum value during the ith iteration,
Figure BDA00020171169000000911
is the minimum value in the (i +1) th iteration.
(1.3.10) judging whether the maximum iteration number is reached, if so, ending the iteration, and outputting the optimal result of the parameter α of the Shepard model, otherwise, returning to the step (1.3.6) until the iteration is ended.
Example one
In the embodiment, the effectiveness of the equipment temperature anomaly detection model based on WCA and Shepard interpolation is verified by using the online monitoring data of a certain wind turbine generator set (rated power: 4MW, rated rotating speed of a generator: 1600r/min) of a certain wind field as a sample. The operating temperature of the unit equipment in the engine room is determined by the ambient temperature of the engine room in which the wind turbine generator operates and the rotating speed of the generator, and the change and the complexity of the operating temperature of the equipment are caused by the continuous conversion of the working condition parameters, so that the real health state and the evolution trend of the equipment cannot be directly obtained from the temperature data of the equipment. Therefore, the self characteristics of the unit equipment are needed, and a fine equipment temperature anomaly detection model capable of adapting to the working condition change of the unit is established.
(1) Selecting standard sample data
And selecting the online monitoring data of the unit in the initial operation stage and in good operation state and under different seasons and different operation conditions as the standard state data of the operation temperature of the unit equipment in health.
(2) Establishing equipment operating temperature standard model
In 1700 groups of health standard data of the equipment, 900 groups of health standard data are extracted to establish an operating temperature standard model (at this time, a WCA method is adopted to optimize parameters α of the Shepard model so that the model achieves the optimal fitting accuracy, the optimization result is α ═ 5.68.), an accurate mapping relation between input parameters (cabin environment temperature and generator rotating speed) and output parameters (equipment operating temperature) is obtained when the equipment operates normally, the remaining 800 groups of data are used as test samples to carry out model verification, and in order to enable the Shepard interpolation model to have good abnormality identification capability, the 1700 groups of health standard data are selected to cover possible cabin environment temperature and generator rotating speed change intervals of the unit equipment as much as possible.
The cabin environment temperature and the generator speed in the 800 test samples are input into the operating temperature standard model, the model outputs the operating temperature standard value p (i), and the comparison between p (i) and the measured data r (i) is shown in table 1, and due to space relationship, table 1 only gives 20 groups of test samples. As can be seen from the table, the standard value and the measured value of the operating temperature of the equipment are basically consistent, and the average relative error of the established model is 1.97 percent, so that the accuracy is very high.
TABLE 1 calculation results of Shepard interpolation based device temperature model
Figure BDA0002017116900000101
Figure BDA0002017116900000111
In the embodiment, the calculation accuracy of the Shepard interpolation model and the Support Vector Machine (SVM) model is compared at the same time, and 800 groups of test samples are respectively substituted into 2 models for calculation, and the accuracy is shown in table 2. As can be seen from the table, the Shepard model has high calculation accuracy and is very suitable for online calculation of temperature anomaly identification of wind turbine equipment.
TABLE 2 comparison table of calculation accuracy of two models
Model (model) Calculation accuracy (average relative error,%)
Shepard interpolation 1.97
SVM 5.40
(3) Establishing an equipment operating temperature abnormity identification model
Substituting real-time acquired cabin environment temperature and generator rotating speed data in a unit equipment state monitoring system into an equipment operating temperature standard model, calculating to obtain a standard value of the equipment operating temperature under the current environment temperature and the unit operating condition, and calculating the deviation H of the equipment operating temperature under the current environment temperature and the unit operating condition by adopting formula (1)dAnd early warning is carried out at the moment when the state mutation and the temperature deviation are higher than the preset value, and a final equipment operation temperature abnormity identification model is established.
As shown in fig. 1 and 2, 200 sets of measured state monitoring data are selected for identifying abnormal operating temperature in a certain operating period of the equipment. The cabin environment temperature and the generator rotation speed in the measured data are input into the temperature standard model to obtain a temperature standard value, and the temperature standard values and measured values of 200 samples to be identified are given in fig. 1. Fig. 2 shows the calculated device temperature deviation according to equation (1). As can be seen from the figure, 200 samples to be identified run at normal temperature, the device has no abnormality and can continue to run.
In summary, the method for identifying the abnormal temperature of the equipment provided by the invention realizes the organic coupling of the external environment temperature, the unit operation process and the state parameter, can fully consider the change processes of the external environment temperature and the unit equipment operation condition in real time, improves the precision of the model, enables the equipment operation temperature information obtained based on the model to be closer to reality, can comprehensively describe the change process of the equipment operation temperature, and can find the possible abnormality in the equipment operation process in advance.
The above embodiments are only used for illustrating the present invention, and the structure, connection mode, manufacturing process, etc. of the components may be changed, and all equivalent changes and modifications performed on the basis of the technical solution of the present invention should not be excluded from the protection scope of the present invention.

Claims (5)

1. A method for detecting abnormal temperature of equipment is characterized by comprising the following steps:
(1) establishing a unit equipment temperature abnormal state identification model based on WCA and Shepard interpolation based on analysis of unit equipment historical data, wherein the unit equipment historical data comprises temperature state monitoring data of a unit under different environmental factors and operating conditions;
in the step (1), the method for establishing the unit equipment temperature abnormal state identification model based on WCA and Shepard interpolation based on the analysis of the historical data of the unit equipment comprises the following steps:
(1.1) analyzing temperature state monitoring data of the unit equipment under different environmental factors and operating conditions, determining the standard health state of the unit equipment, and selecting characteristic parameters capable of reflecting the standard health state of the unit equipment;
(1.2) inputting characteristic parameters of the unit equipment in a standard health state under various environmental factors and operating conditions into a Shepard model, and establishing the Shepard model P ═ f (V, U), wherein P is the equipment temperature, V is the cabin environmental temperature, and U is the rotating speed of a generator;
the method comprises the following steps:
(1.2.1) randomly picking m samples (P) from the characteristic parameterst,Vt,Ut) In (V)t,Ut) As a test sample point, the operating temperature P of the apparatus is measuredtAs response values corresponding to the test sample points, a response value is established from m samples (P)t,Vt,Ut) Forming a m × (2+1) -dimensional matrix, wherein PtIs the operating temperature of the plant, VtIs the ambient temperature of the cabin, UtThe rotating speed of the generator is t 1,2, …, and m is the number of samples;
(1.2.2) taking the residual data in the selected characteristic parameters as new sample points (v, u), and calculating response values p (v, u) corresponding to the new sample points (v, u) by adopting a Shepard interpolation method based on the m × (2+1) dimensional matrix established in the step (1.2.1);
(1.2.3) based on the new sample points (V, u) and the m test sample points (V)t,Ut) The estimated value of the response value p (v, u) which minimizes the Euclidean distance E is calculated
Figure FDA0002449946270000011
Wherein the response value p (v, u) is compared with the estimated value
Figure FDA0002449946270000012
The mapping relation is that the Shepard model is established:
Figure FDA0002449946270000013
in the formula (I), the compound is shown in the specification,
Figure FDA0002449946270000014
as an estimate of the new sample point, ωtNew sample points (V, u) and test sample points (V) are represented as weightst,Ut) Estimation of response values
Figure FDA0002449946270000015
The size of the contribution of (a) to (b),
Figure FDA0002449946270000016
parameter α is the parameter of Shepard model, γt＝[(v-Vt)2+(u-Ut)2]0.5New sample point (V, u) and test sample point (V) are shownt,Ut) The distance between them;
(1.3) optimizing the established Shepard model by adopting a WCA algorithm to obtain an optimal result of the parameter α of the Shepard model, and taking the optimized Shepard model as a unit equipment temperature abnormal state identification model;
in the step (1.3), the WCA algorithm is adopted to optimize the established Shepard model to obtain an optimal result of the parameter α of the Shepard model, and the optimized Shepard model is used as a unit equipment temperature abnormal state identification model, which comprises the following steps:
(1.3.1) determining a cost function for calculating a cost function value of each population;
(1.3.2) setting water circulation algorithm control parameters including the number N of the rainfall layerspopTotal number of rivers and oceans NsrMinimum value dmaxAnd the maximum iteration times T of the water circulation algorithm, and determining the identification range of the parameter α to be identified according to experience;
(1.3.3) randomly generating an initial population to form an initial stream, a river and an ocean;
(1.3.4) dividing the stream population into a plurality of stream layers as model parameters to be respectively input into the Shepard model, and calculating the cost function value J of each stream layer according to the step (1.3.1)i；
(1.3.5) comparing the value of the cost function of each stream layer, selecting the stream layer with the minimum cost function value as the sea, and selecting N according to the sequence of the cost function values from small to largeRiver with water-collecting deviceTaking each stream layer as a river, and determining the number of streams flowing to a specified river and ocean, wherein N isRiver with water-collecting deviceThe number of rivers;
(1.3.6) respectively updating the position of the stream flowing to the river, the position of the stream flowing to the ocean and the position of the river flowing to the ocean, and performing position conversion according to the updated cost function value of each stream, each river and each ocean;
in the step (1.3.6), the method for respectively updating the position of the stream flowing to the river, the position of the stream flowing to the ocean and the position of the river flowing to the ocean and performing position swapping according to the updated cost function value of each stream, river and ocean comprises the following steps:
(1.3.6.1) respectively updating the position of the stream flowing to the river and the position of the stream flowing to the sea;
Figure FDA0002449946270000021
Figure FDA0002449946270000022
in the formula: rand is a random number uniformly distributed between 0 and 1;
Figure FDA0002449946270000023
respectively representing the current positions of the stream, the river and the ocean in the ith iteration process; c is the coefficient of the position update;
(1.3.6.2) inputting the new position of the stream into a Shepard model, calculating a cost function value corresponding to the stream at the moment, and if the cost function of the stream is smaller than that of a river, exchanging the positions of the river and the stream; if the cost function of the stream is smaller than that of the ocean, the ocean and the stream are exchanged;
(1.3.6.3) updating the position of the river towards the ocean;
Figure FDA0002449946270000024
in the formula: rand is a random number uniformly distributed between 0 and 1;
Figure FDA0002449946270000025
respectively representing the current positions of rivers and oceans in the ith iteration process; c is the coefficient of the position update;
(1.3.6.4) inputting the new position of the river into the Shepard model, calculating a cost function value corresponding to the river at the moment, and if the cost function value of the river is smaller than that of the ocean, exchanging the positions of the ocean and the river;
(1.3.7) judging whether the evaporation condition is satisfied: if yes, entering step (1.3.8), otherwise entering step (1.3.9);
(1.3.8) carrying out rainfall process in different modes according to whether rivers and streams are close to the ocean enough to form new rainfall;
(1.3.9) updating the minimum value of the current iteration;
(1.3.10) judging whether the maximum iteration number is reached, if so, ending the iteration, and outputting the optimal result of the parameter α of the Shepard model, otherwise, returning to the step (1.3.6) until the iteration is ended;
(2) acquiring environmental parameters and operating condition parameters of the unit equipment under the current working condition and corresponding temperature measured values r (t), and inputting the environmental parameters and the operating condition parameters of the unit equipment under the current working condition into the unit equipment temperature abnormal state identification model to obtain a health standard value output by the model;
(3) and comparing the temperature measured value r (t) of the unit equipment under the current working condition with the health standard value p (t), and determining the temperature running state of the unit equipment under the current working condition according to the comparison result.
2. The method for detecting a temperature abnormality of a device according to claim 1, wherein: in the step (1.3.7), the evaporation conditions are as follows:
Figure FDA0002449946270000031
in the formula (I), the compound is shown in the specification,
Figure FDA0002449946270000032
and
Figure FDA0002449946270000033
respectively, during the ith iteration, seaAnd the location of the river; 1,2, …, Nsr-1。
3. The method for detecting a temperature abnormality of a device according to claim 1, wherein: in the step (1.3.8), according to whether the river and stream are close enough to the sea or not, the method for forming new rainfall by adopting different modes to carry out the rainfall process comprises the following steps:
if it is not
Figure FDA0002449946270000034
Or rand<0.1，i＝1，2，…，Nsr1, rand is a random number evenly distributed between 0 and 1, then the rainfall process is carried out using the following formula:
Figure FDA0002449946270000035
in the formula:
Figure FDA0002449946270000036
UB and L B are the upper and lower bounds of the variables, respectively, for the latest position of the newly formed stream;
if it is not
Figure FDA0002449946270000037
i＝1，2，…，NS1The rainfall process was carried out using the following formula:
Figure FDA0002449946270000038
in the formula: randn is a normally distributed random number; μ represents a coefficient of a search area range near the ocean; n is a radical ofvarIs the dimension of the search space.
4. The method for detecting a temperature abnormality of a device according to claim 1, wherein: in the step (1.3.9), the minimum value
Figure FDA0002449946270000039
The calculation formula of (2) is as follows:
Figure FDA00024499462700000310
in the formula: t is the maximum iteration number of the water circulation algorithm,
Figure FDA00024499462700000311
the minimum value during the ith iteration,
Figure FDA00024499462700000312
is the minimum value in the (i +1) th iteration.
5. The method for detecting a temperature abnormality of a device according to claim 1, wherein: in the step (3), a calculation formula for comparing the temperature measured value r (t) of the unit equipment under the current working condition with the health standard value p (t) is as follows:
Figure FDA0002449946270000041
in the formula: t represents the operation time of the unit equipment; and w is an early warning threshold value.
CN201910266851.3A 2019-04-03 2019-04-03 Method for detecting abnormal temperature of equipment Active CN110006552B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201910266851.3A CN110006552B (en) 2019-04-03 2019-04-03 Method for detecting abnormal temperature of equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201910266851.3A CN110006552B (en) 2019-04-03 2019-04-03 Method for detecting abnormal temperature of equipment

Publications (2)

Publication Number Publication Date
CN110006552A CN110006552A (en) 2019-07-12
CN110006552B true CN110006552B (en) 2020-07-14

Family

ID=67169728

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201910266851.3A Active CN110006552B (en) 2019-04-03 2019-04-03 Method for detecting abnormal temperature of equipment

Country Status (1)

Country Link
CN (1) CN110006552B (en)

Families Citing this family (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN112211794B (en) * 2020-09-02 2023-01-06 五凌电力有限公司新能源分公司 Cabin temperature abnormity early warning method, device, equipment and medium of wind turbine generator
CN112185070B (en) * 2020-09-11 2021-12-17 珠海格力电器股份有限公司 Fault early warning method, storage medium and electronic equipment
CN112504511B (en) * 2020-12-15 2023-08-15 润电能源科学技术有限公司 Generator stator temperature monitoring method, device and medium
CN114415581B (en) * 2021-12-28 2022-11-04 广东德尔智慧工厂科技有限公司 Mechanical equipment operation and maintenance method and system
CN115574963B (en) * 2022-12-08 2023-03-17 南京智联达科技有限公司 Wireless temperature acquisition system and temperature acquisition method based on IOT technology

Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103926079A (en) * 2014-04-09 2014-07-16 中国水利水电科学研究院 Method for detecting output abnormity of mixed-flow hydroelectric generating set
CN103940611A (en) * 2014-04-09 2014-07-23 中国水利水电科学研究院 Self-adaptive anomaly detection method for rolling bearing of wind generator set under variable working conditions
CN107742053A (en) * 2017-11-28 2018-02-27 国华（河北）新能源有限公司 Wind turbines abnormality recognition method and device
CN107844780A (en) * 2017-11-24 2018-03-27 中南大学 A kind of the human health characteristic big data wisdom computational methods and device of fusion ZED visions
CN108376298A (en) * 2018-02-12 2018-08-07 湘潭大学 A kind of Wind turbines generator-temperature detection fault pre-alarming diagnostic method
CN109034478A (en) * 2018-07-27 2018-12-18 中南大学 A kind of line of high-speed railway strong wind iteration competition high-precision forecasting method

Family Cites Families (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US8392869B2 (en) * 2010-06-15 2013-03-05 California Institute Of Technology Systems and methods for circuit lifetime evaluation

Patent Citations (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN103926079A (en) * 2014-04-09 2014-07-16 中国水利水电科学研究院 Method for detecting output abnormity of mixed-flow hydroelectric generating set
CN103940611A (en) * 2014-04-09 2014-07-23 中国水利水电科学研究院 Self-adaptive anomaly detection method for rolling bearing of wind generator set under variable working conditions
CN107844780A (en) * 2017-11-24 2018-03-27 中南大学 A kind of the human health characteristic big data wisdom computational methods and device of fusion ZED visions
CN107742053A (en) * 2017-11-28 2018-02-27 国华（河北）新能源有限公司 Wind turbines abnormality recognition method and device
CN108376298A (en) * 2018-02-12 2018-08-07 湘潭大学 A kind of Wind turbines generator-temperature detection fault pre-alarming diagnostic method
CN109034478A (en) * 2018-07-27 2018-12-18 中南大学 A kind of line of high-speed railway strong wind iteration competition high-precision forecasting method

Non-Patent Citations (3)

* Cited by examiner, † Cited by third party
Title
基于三维曲面的抽水蓄能电站机组故障预警模型;安学利等;《水力发电》;20130131;第39卷(第1期);第71-74页 *
基于实测信息的海堤PHM系统框架及关键技术研究;蓝祝光等;《水土保持通报》;20170630;第37卷(第3期);第307-313页 *
基于水循环算法的梯阶水电站短期优化调度;黄景光等;《水电能源科学》;20190131;第37卷(第1期);第65-69页 *

Also Published As

Publication number Publication date
CN110006552A (en) 2019-07-12

Similar Documents

Publication Publication Date Title
CN110006552B (en) Method for detecting abnormal temperature of equipment
CN106779200A (en) Based on the Wind turbines trend prediction method for carrying out similarity in the historical data
CN107291991B (en) Early defect early warning method for wind turbine generator based on dynamic network sign
Zhan et al. Abnormal vibration detection of wind turbine based on temporal convolution network and multivariate coefficient of variation
CN113107785B (en) Real-time monitoring method and device for power performance abnormity of wind turbine generator
McKay et al. Global sensitivity analysis of wind turbine power output
CN110378042A (en) Gearbox of wind turbine oil temperature method for detecting abnormality and system based on SCADA data
CN113408068A (en) Random forest classification machine pump fault diagnosis method and device
CN115329986A (en) Wind turbine generator anomaly detection and positioning method based on interpretable graph neural network
CN115314235A (en) System and method for network attack detection in wind turbine control systems
CN115438726A (en) Device life and fault type prediction method and system based on digital twin technology
CN114742363A (en) Energy efficiency state evaluation method, system and medium for wind turbine generator
CN110108457B (en) Primary fan shaft temperature diagnosis method based on universal gravitation neural network
CN116771610A (en) Method for adjusting fault evaluation value of variable pitch system of wind turbine
CN112067289A (en) Motor shaft and transmission shaft abnormal vibration early warning algorithm based on neural network
KR102110319B1 (en) System for generating learning data
US11339763B2 (en) Method for windmill farm monitoring
Zhang Comparison of data-driven and model-based methodologies of wind turbine fault detection with SCADA data
CN114936520A (en) Steam turbine fault early warning analysis method based on improved MSET
CN114049014A (en) Method, device and system for evaluating operation state of offshore wind turbine generator
CN114320770A (en) Effective wind speed measuring method and device for wind turbine generator
CN112699598A (en) Intelligent diagnosis method and device for abnormal oil temperature of gear box
Zhang et al. Fault Diagnosis of Wind Turbine Gearbox Based on Multisensor Data Fusion
Wang et al. Wind turbine gearbox forecast using Gaussian process model
Cross et al. Model-based condition monitoring for wind turbines

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
TR01 Transfer of patent right
TR01 Transfer of patent right

Effective date of registration: 20210922

Address after: 20 Chegongzhuang West Road, Haidian District, Beijing 100038

Patentee after: China Institute of Water Resources and Hydropower Research

Patentee after: Tianjin Shuike electromechanical Co.,Ltd.

Address before: 20 Chegongzhuang West Road, Haidian District, Beijing 100038

Patentee before: China Institute of Water Resources and Hydropower Research