CN111780933B - Method and system for diagnosing leakage fault of high-pressure heater based on neural network and thermodynamic modeling - Google Patents
Method and system for diagnosing leakage fault of high-pressure heater based on neural network and thermodynamic modeling Download PDFInfo
- Publication number
- CN111780933B CN111780933B CN202010623173.4A CN202010623173A CN111780933B CN 111780933 B CN111780933 B CN 111780933B CN 202010623173 A CN202010623173 A CN 202010623173A CN 111780933 B CN111780933 B CN 111780933B
- Authority
- CN
- China
- Prior art keywords
- pressure heater
- regulating valve
- leakage
- opening
- pressure
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01M—TESTING STATIC OR DYNAMIC BALANCE OF MACHINES OR STRUCTURES; TESTING OF STRUCTURES OR APPARATUS, NOT OTHERWISE PROVIDED FOR
- G01M3/00—Investigating fluid-tightness of structures
- G01M3/02—Investigating fluid-tightness of structures by using fluid or vacuum
- G01M3/26—Investigating fluid-tightness of structures by using fluid or vacuum by measuring rate of loss or gain of fluid, e.g. by pressure-responsive devices, by flow detectors
- G01M3/28—Investigating fluid-tightness of structures by using fluid or vacuum by measuring rate of loss or gain of fluid, e.g. by pressure-responsive devices, by flow detectors for pipes, cables or tubes; for pipe joints or seals; for valves ; for welds
- G01M3/2876—Investigating fluid-tightness of structures by using fluid or vacuum by measuring rate of loss or gain of fluid, e.g. by pressure-responsive devices, by flow detectors for pipes, cables or tubes; for pipe joints or seals; for valves ; for welds for valves
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Evolutionary Computation (AREA)
- General Health & Medical Sciences (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Data Mining & Analysis (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Health & Medical Sciences (AREA)
- Testing Of Devices, Machine Parts, Or Other Structures Thereof (AREA)
- Examining Or Testing Airtightness (AREA)
Abstract
The invention provides a method and a system for diagnosing leakage faults of a high-pressure heater based on a neural network and thermodynamic modeling. The invention comprises the following steps: establishing a mathematical model about the opening of each drain regulating valve of each high-pressure heater and the flow of the fluid flowing through the drain regulating valve based on the collected historical data; extracting characteristic vector parameters when the high-pressure heater system normally operates, and constructing a training set and a test set; performing model training on training set data through a neural network algorithm, and recording the maximum error range of a comparison graph of the opening of the drain regulating valve of the heater and the opening of an actual drain regulating valve; calculating the difference value between each high pressure steam extraction flow and the drainage flow in real time to complete the first safety diagnosis; and comparing the difference value of the opening degree of the drain regulating valve of the high-pressure heater calculated by the neural network model with the actual opening degree to finish the second safety diagnosis, and automatically switching the first safety diagnosis and the second safety diagnosis. The method has higher accuracy, and can greatly improve the real-time performance of leakage diagnosis of the high-pressure heater.
Description
Technical Field
The invention relates to the field of power system fault diagnosis, in particular to a method and a system for diagnosing leakage faults of a high-pressure heater based on a neural network and thermodynamic modeling.
Background
The high-pressure heater is a device for heating the feed water by using partial extraction of the steam turbine. As a heat conversion device, the heat conversion device is mainly applied to a heat regeneration system of a large thermal power generating unit. The high-pressure heater is composed of a shell and a pipe system, a steam condensation section is arranged at the upper part of an inner cavity of the shell, a drainage cooling section is arranged at the lower part of the inner cavity of the shell, and a water supply inlet and a water supply outlet are arranged at the top ends of a water inlet pipe and a water outlet pipe. After the superheated steam enters the shell from the inlet, the feed water in the upper main spiral pipe can be heated, after the steam is condensed into water, the condensed hot water can heat part of the feed water in the lower cooling spiral pipe, and the condensed water after being utilized flows out of the body through the drainage outlet.
The traditional leakage diagnosis of the high-pressure heater needs to be judged through the operation experience of centralized control personnel, and the unit does not have the alarm of the leakage of the high-pressure heater, so that the leakage diagnosis precision and the real-time performance of the high-pressure heater are poor. In the prior art, some flow measurement points are additionally arranged to assist judgment, the cost for additionally arranging the measurement points is higher, the maintenance cost of the measurement points is correspondingly increased in the later period, after the pipeline is punched and additionally arranged with the flow measurement points, the original structure of the pipeline is damaged, the throttling loss is increased, and the economical efficiency and the safety are deteriorated.
The method and the system for judging the leakage of the high-pressure heater system, which are disclosed as 109459195A, disclose a method and a system for calculating a plurality of historical drainage total deviation values of an outlet of the high-pressure heater system according to a heat balance principle, a material conservation principle and a plurality of historical measuring point data during the normal operation of the high-pressure heater system, and then determining a leakage criterion according to a preset algorithm. In the face of faults with high leakage and large leakage, diagnosis information cannot be given in time, and the fault diagnosis response speed of the method is low.
Disclosure of Invention
In light of the above-mentioned technical problems, a method and system for diagnosing leakage fault of high pressure heater based on neural network and thermodynamic modeling are provided. The technical means adopted by the invention are as follows:
a method for diagnosing leakage faults of a high-pressure heater based on a neural network and thermodynamic modeling comprises the following steps:
s1, constructing mathematical models of the opening of each drain control valve of each high-pressure heater and the flow of the fluid flowing through the drain control valve based on the collected historical data;
s2, extracting characteristic vector parameters corresponding to the normal opening of the drainage regulating valve of each high-pressure heater when the high-pressure heater system normally operates, and dividing the extracted characteristic vector data into a training set and a test set;
s3, performing model training on training set data through a neural network algorithm to obtain a model of the opening degree of the drain regulating valve of the high-pressure heater, calculating a comparison graph of the opening degree of the drain regulating valve of each heater and the opening degree of an actual drain regulating valve based on the model, and recording a maximum error range D1;
s4, calculating expected values of normal drainage valve opening of the high-pressure heaters in real time based on a heat balance principle and a material conservation principle, and calculating the steam extraction flow rate of each high-pressure heater and the average value of the drainage flow rate in the previous time period; comparing the difference value of each high pressure steam extraction flow and the drainage flow to finish the first safety diagnosis;
comparing the difference D2 between the opening of the drain regulating valve of the high-pressure heater calculated by the neural network model and the actual opening, and comparing the values D2 and D1 to finish second safety diagnosis;
and when the characteristic vector data of the neural network is beyond the range of the sample data, automatically switching to the first safety diagnosis.
Further, the mathematical model is specifically: the input of the model is the opening of the regulating valve, the pressure in front of the regulating valve, the pressure behind the regulating valve and the working medium temperature, the intermediate parameter is the admittance of the regulating valve and the nonlinear characteristic of the regulating valve, the output of the model is the fluid flow passing through the regulating valve, and the specific formula is as follows:
wherein: cond is the conduction coefficient of the valve and the communication pipeline, c is the opening degree of the regulating valve, f (c) is a broken line function, p is obtained by debuggingIntoIndicating the valve inlet pressure, pGo outDenotes the valve outlet pressure, t denotes the hydrophobic temperature, f (t) denotes the temperature correction.
Further, in step S1, the accuracy of the constructed mathematical model is also determined by the following method: by adjusting the conduction coefficient of the drainage regulating valve of each high-pressure heater and the nonlinear characteristic of the regulating valve at different opening degrees, the time domain average value of the extraction flow of each high-pressure heater is equal to the time domain average value obtained by subtracting the drainage flow of the previous-stage heater from the drainage flow of the corresponding high-pressure heater.
Further, in step S3, the accuracy of the opening model of the drain regulator valve of the high-pressure heater is verified by the following method: and (3) transferring fault data of the leakage of the high-pressure heater, substituting the fault data of the leakage of the high-pressure heater into the model, and if the minimum difference between the opening degree value of the normal drainage regulating valve calculated by the model and the opening degree of the actual regulating valve is far larger than D1, considering the model as a qualified model.
Further, the method also comprises the following steps: and S5, substituting D2 into the mathematical model, and calculating the expected leakage amount of the high-pressure heater.
Further, the method also comprises the following steps: s6, according to the similarity principle: wdqi*Wdneti/(‖Wdq876‖*‖Wdnet876|) calculating a cosine included angle of the two vectors as an identification degree n% of the two vectors, wherein i ═ 1.. n represents the number of the high-pressure heater, and the smaller i is, the larger the extraction pressure is;
setting the first safety diagnostic maximum leak amount to WmaxThe second safety diagnosis maximum leakage amount is wmax，n％*(max(Wdqi)/Wmax*(max(Wdneti)/wmax) As confidence pre of the high pressure heater leak diagnosis;
when the confidence coefficient of the leakage of the high-pressure heater exceeds a set threshold value, the leakage of the high-pressure heater is judged, and the maximum value pre max (W) of the leakage of the high-pressure heater is measureddqi,Wdneti)。
The invention also provides a system for diagnosing the leakage fault of the high-pressure heater based on the neural network and the thermodynamic modeling, which comprises the following components:
the data integration unit comprises a characteristic vector acquisition module for acquiring real-time data of each high-pressure heater and a storage module for storing the acquired data;
the mathematical model construction unit is used for constructing a drainage regulating valve mathematical model based on the collected historical data;
the neural network model building unit is used for obtaining a drain regulating valve opening model of the high-pressure heater through neural network training based on the collected historical data;
the first safety judgment unit is used for comparing the difference value of the flow of the steam flowing through the drainage regulating valve of the heater at the current stage minus the flow of the steam flowing through the drainage regulating valve of the heater at the previous stage with the deviation of the steam extraction flow of the high-pressure heater calculated by real-time acquired information and finishing first safety diagnosis;
the second safety judgment unit is used for inputting the characteristic vector of the actual opening of the hydrophobic regulating valve collected in real time into the neural network model to calculate to obtain the predicted opening of the hydrophobic regulating valve, and the deviation between the actual opening and the predicted opening is compared with D1 to complete second safety diagnosis;
and the alarm unit is used for sending out a visual alarm signal when the first safety diagnosis belongs to the leakage fault or the second safety diagnosis belongs to the leakage fault.
Further, still include:
and the high-pressure heater leakage confirmation unit is used for comparing the leakage confidence coefficient of the high-pressure heater with a set threshold value, judging the leakage of the high-pressure heater if the leakage confidence coefficient of the high-pressure heater exceeds the set threshold value, and calculating the leakage confidence coefficient of the high-pressure heater through the similarity between a vector formed by the deviation of each steam extraction and drainage flow of the first safety judgment unit and a vector formed by converting the deviation of each actual valve opening value and an expected valve opening value of the second safety judgment unit into a flow deviation.
The invention has the following advantages:
1. the real-time value of the drain regulating valve of the high-pressure heater is calculated through the neural network model, and when the error between the actual value and the expected value exceeds a set threshold value, early warning is immediately sent out, so that the real-time performance of leakage diagnosis of the high-pressure heater can be greatly improved.
2. The extraction steam quantity of each high-pressure heater is calculated according to the heat balance principle, and the specific high-pressure heater leakage can be accurately positioned by analyzing and comparing the difference value between the extraction steam quantity and the hydrophobic quantity, so that the accuracy is higher.
3. The difference value of the opening degree of the drain valve calculated by the neural network model and the actual opening degree can be converted into the high pressure leakage amount by utilizing the established mathematical model of the drain valve, and the two different types of variables are converted into the same type of variables, so that the high pressure leakage diagnosis has data reference, and the reference of the diagnosis result is more visual and clear.
4. The two methods are combined according to the similarity principle, so that the accuracy is improved, and the confidence coefficient of high-pressure leakage fault diagnosis is increased.
5. The flow of fluid through each hydrophobic control valve can be calculated by establishing a mathematical model for the hydrophobic control valves of the high-pressure heater. The method can be applied to the field of centralized control operation of thermal power plants, and can save a large amount of investment for additionally installing the measuring points and the later maintenance cost. The damage to the pipeline structure caused by additionally installing the flow measuring point is avoided, the throttling loss after the flow measuring point is additionally installed is reduced, and the safety and the economical efficiency of the unit are improved.
According to the scheme, no equipment is required to be additionally installed, high voltage leakage fault diagnosis can be carried out in real time by only applying the method, and leakage high voltage leakage can be accurately positioned. The high-pressure heater system is suitable for most of the existing high-pressure heater systems of power plants, and has strong applicability and popularization value.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings needed to be used in the description of the embodiments or the prior art will be briefly introduced below, and it is obvious that the drawings in the following description are some embodiments of the present invention, and for those skilled in the art, other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a general flow diagram of the present invention.
FIG. 2 is a flow chart of the embodiment of the present invention, which takes the high pressure heater number 678 as an example, wherein (a) is the preliminary processing stage and (b) is the result determination stage.
FIG. 3 is a comparison graph of neural network predicted opening of the trap in an embodiment of the present invention;
FIG. 4 is a graph comparing the calculated drain flow rate and the actual extraction flow rate of the mathematical model in the embodiment of the present invention;
FIG. 5 is a confidence comparison graph in an embodiment of the invention.
FIG. 6 is a calculated leakage trend graph in accordance with an embodiment of the present invention.
Detailed Description
In order to make the objects, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are some, but not all, embodiments of the present invention. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, the present embodiment discloses a method for diagnosing a leakage fault of a high pressure heater based on a neural network and thermodynamic modeling, which includes the following steps:
s1, constructing mathematical models of the opening of each drain control valve of each high-pressure heater and the flow of the fluid flowing through the drain control valve based on the collected historical data; the mathematical model is specifically as follows: the input of the model is the opening of the regulating valve, the pressure in front of the regulating valve, the pressure behind the regulating valve and the working medium temperature, the intermediate parameter is the admittance of the regulating valve and the nonlinear characteristic of the regulating valve, the output of the model is the fluid flow passing through the regulating valve, and the specific formula is as follows:
wherein: cond is the conduction coefficient of the valve and the communication pipeline, c is the opening degree of the regulating valve, f (c) is a broken line function, p is obtained by debuggingIntoIndicating the valve inlet pressure, pGo outDenotes the valve outlet pressure, t denotes the hydrophobic temperature, f (t) denotes the temperature correction.
Because the energy conservation and the mass conservation are calculated on the premise of mass flow, and the modeling can only calculate the volume flow of the fluid, the density correction rho is required to be added, and the rho can change regularly along with the temperature, and the variation is shown in table 1. As can be seen from the water density meter, when the water is in a liquid state, the temperature has a large influence on the density, and the pressure has a small influence on the density, so that only the temperature correction f (t) needs to be added.
TABLE 1
Density of water kg/m3Watch (A)
The step also judges the accuracy of the constructed mathematical model by the following method: by adjusting the conduction coefficient of the drainage regulating valve of each high-pressure heater and the nonlinear characteristic of the regulating valve at different opening degrees, the time domain average value of the extraction flow of each high-pressure heater is equal to the time domain average value obtained by subtracting the drainage flow of the previous-stage heater from the drainage flow of the corresponding high-pressure heater.
S2, extracting characteristic vector parameters corresponding to the normal opening of the drainage regulating valve of each high-pressure heater when the high-pressure heater system normally operates, and dividing the extracted characteristic vector data into a training set and a test set;
s3, performing model training on training set data through a neural network algorithm to obtain a model of the opening degree of the drain regulating valve of the high-pressure heater, calculating a comparison graph of the opening degree of the drain regulating valve of each heater and the opening degree of an actual drain regulating valve based on the model, and recording a maximum error range D1;
the accuracy of the opening model of the drain regulating valve of the high-pressure heater is verified by the following method:
and (3) transferring fault data of the leakage of the high-pressure heater, substituting the fault data of the leakage of the high-pressure heater into the model, and if the minimum difference between the opening degree value of the normal drainage regulating valve calculated by the model and the opening degree of the actual regulating valve is far larger than D1, considering the model as a qualified model.
S4, calculating expected values of normal drainage valve opening of the high-pressure heaters in real time based on a heat balance principle and a material conservation principle, and calculating the steam extraction flow rate of each high-pressure heater and the average value of the drainage flow rate in the previous time period; comparing the difference value of each high pressure steam extraction flow and the drainage flow to finish the first safety diagnosis;
comparing the difference D2 between the opening of the drain regulating valve of the high-pressure heater calculated by the neural network model and the actual opening, and comparing the values D2 and D1 to finish second safety diagnosis;
and when the characteristic vector data of the neural network is beyond the range of the sample data, automatically switching to the first safety diagnosis.
And S5, substituting D2 into the mathematical model, and calculating the expected leakage amount of the high-pressure heater.
S6, according to the similarity principle: wdqi*Wdneti/(‖Wdqi‖*‖Wdneti|) calculating a cosine included angle of the two vectors as an identification degree n% of the two vectors, wherein i ═ 1.. n represents the number of the high-pressure heater, and the smaller i is, the larger the extraction pressure is;
setting the first safety diagnostic maximum leak amount to WmaxThe second safety diagnosis maximum leakage amount is wmax，n％*(max(Wdqi)/Wmax*(max(Wdneti)/wmax) As confidence pre of the high pressure heater leak diagnosis;
when the confidence coefficient of the leakage of the high-pressure heater exceeds a set threshold value, the leakage of the high-pressure heater is judged, and the maximum value pre max (W) of the leakage of the high-pressure heater is measureddqi,Wdneti)。
Example 1
As shown in fig. 2(a) (b), taking a steam turbine set of a huaneng large power plant as an example, the high pressure heater is 8/7/6 sections, and the high pressure heater number and the number of different sets are different, but the method provided by the embodiment can be adopted.
And 2, utilizing the fluid mechanics principle, wherein the volume flow passing through the pipeline under the turbulent flow state is in a linear relation with the differential pressure opening signals on the two sides of the pipeline. The Reynolds coefficient of the round smooth pipeline is about 2500, the Reynolds coefficient of the high-pressure drainage pipeline is obviously increased after throttling and perennial corrosion of the drainage valve, relevant parameters are substituted into a Reynolds coefficient calculation formula Re ═ ρ vd/η, the Reynolds coefficient of the high-pressure drainage pipeline passing through the drainage valve is about 4500, and therefore the high-pressure drainage is judged to be in a turbulent flow state. A mathematical model is established for each high pressure and high pressure drain control valve, the input of the model is the opening degree of the control valve, the pressure before the control valve, the pressure after the control valve and the working medium temperature, the intermediate parameters are the conduction coefficient (admittance) of the control valve and the nonlinear characteristic of the control valve, and the output of the model is the fluid flow passing through the control valve.
And 3, under the condition that the water level is measured by high steam addition, each high steam extraction amount is equal to the corresponding high steam extraction amount, namely the hydrophobic amount of the previous-stage heater, and according to the mass conservation law, because the steam and the hydrophobic density are different, a density correction coefficient rho is required to be added during calculation, and the rho can regularly change along with the difference of the temperature. And (3) extracting historical data of the high pressure steam extraction system to verify the accuracy of the mathematical model established in the step (2), and enabling the time domain average value of each high pressure steam extraction flow to be equal to the time domain average value of the corresponding high pressure steam extraction flow by adjusting the conduction coefficient of each high pressure steam extraction trap and the nonlinear characteristics of the trap at different opening degrees. The time domain interval is selected to be 300 s. The error can be controlled within 5%.
And 4, extracting characteristic vector parameters (data contain all working conditions of the high pressure system in normal operation as much as possible) corresponding to the opening of each high pressure normal hydrophobic valve in normal operation of the high pressure system, and dividing the extracted characteristic vector data into a training set and a test set (split according to the quantity proportion of 3/1).
And 5, carrying out model training on training set data by utilizing an nn.sequential neural network algorithm in the pytorch, predicting a test set by utilizing the trained model to obtain a comparison graph of the opening degree of the drain trap of each heater and the actual opening degree of the drain trap calculated by the neural network and a maximum error value, recording a maximum error range D1, and storing the model (net).
And 6, calling high plus leakage fault data, substituting the high plus leakage fault data into a trained model (net) to verify the accuracy of the model, and if the minimum difference between the normal hydrophobic regulating valve opening degree value calculated by the model and the actual regulating valve opening degree is far larger than D1, considering the model (net) as a qualified model.
And 7, realizing two models: and calculating expected values of the opening of each high-pressure steam extraction valve and the average value of each high-pressure steam extraction flow and the drainage flow in the previous time period (300s) in real time by using an intelligent operation management platform developed by the Huaneng large continuous power plant.
And 9, comparing the difference D2 between the opening of the high-pressure and high-pressure drain regulating valve calculated by the neural network model and the actual opening, and giving out an early warning when D2 is larger than D1.
Step 11, according to the similarity principle: wdq876*Wdnet876/(‖Wdq876‖*‖Wdnet876II) solving a cosine included angle of the two vectors as the identification degree n% of the two vectors;
setting the first safety diagnostic maximum leak amount to WmaxThe second safety diagnosis maximum leakage amount is wmax，n％*(max(Wdq876)/Wmax*(max(Wdnet876)/wmax) As confidence pre of the high pressure heater leak diagnosis;
and step 12, judging high plus leakage when the confidence coefficient of the high plus leakage exceeds a set threshold value. The high plus leakage is taken to be maximum max (W)dq876,Wdnet87)。
The data results obtained in the above steps are shown in fig. 3-6, in fig. 3, the line tending to be gentle is the predicted opening, and the line with larger fluctuation is the actual opening; in fig. 4, the air extraction amount of the rear section is larger than the hydrophobic amount; in fig. 5, the data near the lower side is the confidence of normal data, and the data near the upper side is the confidence of failure data; in fig. 6, data near the lower side is normal data, and data near the upper side is failure data; according to the method, the high-voltage leakage fault can be diagnosed in real time, the leakage high-voltage leakage can be accurately positioned, and the safety and the economy of the unit can be improved.
The embodiment of the invention also discloses a system for diagnosing the leakage fault of the high-pressure heater based on the neural network and the thermodynamic modeling, which comprises the following steps: the data integration unit comprises a characteristic vector acquisition module for acquiring real-time data of each high-pressure heater and a storage module for storing the acquired data;
the mathematical model construction unit is used for constructing a drainage regulating valve mathematical model based on the collected historical data;
the neural network model building unit is used for obtaining a drain regulating valve opening model of the high-pressure heater through neural network training based on the collected historical data;
the first safety judgment unit is used for comparing the difference value of the flow of the steam flowing through the drainage regulating valve of the heater at the current stage minus the flow of the steam flowing through the drainage regulating valve of the heater at the previous stage with the deviation of the steam extraction flow of the high-pressure heater calculated by real-time acquired information and finishing first safety diagnosis;
the second safety judgment unit is used for inputting the characteristic vector of the actual opening of the hydrophobic regulating valve collected in real time into the neural network model to calculate to obtain the predicted opening of the hydrophobic regulating valve, and the deviation between the actual opening and the predicted opening is compared with D1 to complete second safety diagnosis;
and the alarm unit is used for sending out a visual alarm signal when the first safety diagnosis belongs to the leakage fault or the second safety diagnosis belongs to the leakage fault.
Further comprising:
and the high-pressure heater leakage confirmation unit is used for comparing the leakage confidence coefficient of the high-pressure heater with a set threshold value, judging the leakage of the high-pressure heater if the leakage confidence coefficient of the high-pressure heater exceeds the set threshold value, and calculating the leakage confidence coefficient of the high-pressure heater through the similarity between a vector formed by the deviation of each steam extraction and drainage flow of the first safety judgment unit and a vector formed by converting the deviation of each actual valve opening value and an expected valve opening value of the second safety judgment unit into a flow deviation.
Finally, it should be noted that: the above embodiments are only used to illustrate the technical solution of the present invention, and not to limit the same; while the invention has been described in detail and with reference to the foregoing embodiments, it will be understood by those skilled in the art that: the technical solutions described in the foregoing embodiments may still be modified, or some or all of the technical features may be equivalently replaced; and the modifications or the substitutions do not make the essence of the corresponding technical solutions depart from the scope of the technical solutions of the embodiments of the present invention.
Claims (7)
1. A method for diagnosing leakage faults of a high-pressure heater based on a neural network and thermodynamic modeling is characterized by comprising the following steps:
s1, constructing mathematical models of the opening of each drain control valve of each high-pressure heater and the flow of the fluid flowing through the drain control valve based on the collected historical data;
s2, extracting characteristic vector parameters corresponding to the normal opening of the drainage regulating valve of each high-pressure heater when the high-pressure heater system normally operates, and dividing the extracted characteristic vector data into a training set and a test set;
s3, performing model training on training set data through a neural network algorithm to obtain a model of the opening degree of the drain regulating valve of the high-pressure heater, calculating a comparison graph of the opening degree of the drain regulating valve of each heater and the opening degree of an actual drain regulating valve based on the model, and recording a maximum error range D1;
s4, calculating expected values of normal drainage valve opening of the high-pressure heaters in real time based on a heat balance principle and a material conservation principle, and calculating the steam extraction flow rate of each high-pressure heater and the average value of the drainage flow rate in the previous time period; comparing the difference value of each high pressure steam extraction flow and the drainage flow to finish the first safety diagnosis;
comparing the difference D2 between the opening of the drain regulating valve of the high-pressure heater calculated by the neural network model and the actual opening, and comparing the values D2 and D1 to finish second safety diagnosis;
when the characteristic vector of the neural network is in the sample data range, selecting a second safety diagnosis, and when the characteristic vector data of the neural network exceeds the sample data range, automatically switching to the first safety diagnosis;
also comprises the following steps: s5, substituting D2 into the mathematical model, and calculating the expected leakage amount of the high-pressure heater;
also comprises the following steps: s6, according to the similarity principle: wdqi*Wdneti/(‖Wdqi‖*‖Wdneti|) calculating a cosine included angle of the two vectors as an identification degree n% of the two vectors, wherein i ═ 1.. n represents the number of the high-pressure heater, and the smaller i is, the larger the extraction pressure is;
setting the first safety diagnostic maximum leak amount to WmaxThe second safety diagnosis maximum leakage amount is wmax，n％*(max(Wdqi)/Wmax*(max(Wdneti)/wmax) As confidence pre of the high pressure heater leak diagnosis;
when the confidence coefficient of the leakage of the high-pressure heater exceeds a set threshold value, the leakage of the high-pressure heater is judged, and the maximum value pre max (W) of the leakage of the high-pressure heater is measureddqi,Wdneti)。
2. The method for diagnosing the leakage fault of the high-pressure heater based on the neural network and the thermodynamic modeling as claimed in claim 1, wherein the mathematical model is specifically as follows: the input of the model is the opening of the regulating valve, the pressure in front of the regulating valve, the pressure behind the regulating valve and the working medium temperature, the intermediate parameter is the admittance of the regulating valve and the nonlinear characteristic of the regulating valve, the output of the model is the fluid flow passing through the regulating valve, and the specific formula is as follows:
wherein: cond is the conduction coefficient of the valve and the communication pipeline, c is the opening degree of the regulating valve, f (c) is a broken line function, p is obtained by debuggingIntoIndicating the valve inlet pressure, pGo outDenotes the valve outlet pressure, t denotes the hydrophobic temperature, f (t) denotes the temperature correction.
3. The method for diagnosing the leakage fault of the high-pressure heater based on the neural network and the thermodynamic modeling as claimed in claim 1, wherein in step S1, the accuracy of the constructed mathematical model is further determined by: by adjusting the conduction coefficient of the drainage regulating valve of each high-pressure heater and the nonlinear characteristic of the regulating valve at different opening degrees, the time domain average value of the extraction flow of each high-pressure heater is equal to the time domain average value obtained by subtracting the drainage flow of the previous-stage heater from the drainage flow of the corresponding high-pressure heater.
4. The method for diagnosing the leakage fault of the high-pressure heater based on the neural network and the thermodynamic modeling as claimed in claim 1, wherein in the step S3, the accuracy of the opening model of the drain regulating valve of the high-pressure heater is verified by the following method: and (3) transferring fault data of the leakage of the high-pressure heater, substituting the fault data of the leakage of the high-pressure heater into the model, and if the minimum difference between the opening degree value of the normal drainage regulating valve calculated by the model and the opening degree of the actual regulating valve is far larger than D1, considering the model as a qualified model.
5. A system of any one of claims 1 to 4 for a method of diagnosing a leakage fault of a high pressure heater based on a neural network and thermodynamic modeling, comprising: the data integration unit comprises a characteristic vector acquisition module for acquiring real-time data of each high-pressure heater and a storage module for storing the acquired data;
the mathematical model construction unit is used for constructing a drainage regulating valve mathematical model based on the collected historical data;
the neural network model building unit is used for obtaining a drain regulating valve opening model of the high-pressure heater through neural network training based on the collected historical data;
the first safety judgment unit is used for comparing the difference value of the flow of the steam flowing through the drainage regulating valve of the heater at the current stage minus the flow of the steam flowing through the drainage regulating valve of the heater at the previous stage with the deviation of the steam extraction flow of the high-pressure heater calculated by real-time acquired information and finishing first safety diagnosis;
the second safety judgment unit is used for inputting the characteristic vector of the actual opening of the hydrophobic regulating valve collected in real time into the neural network model to calculate to obtain the predicted opening of the hydrophobic regulating valve, and the deviation between the actual opening and the predicted opening is compared with D1 to complete second safety diagnosis;
and the alarm unit is used for sending out a visual alarm signal when the first safety diagnosis belongs to the leakage fault or the second safety diagnosis belongs to the leakage fault.
6. The system of claim 5, further comprising:
and the high-pressure heater leakage confirmation unit is used for comparing the leakage confidence coefficient of the high-pressure heater with a set threshold value, judging the leakage of the high-pressure heater if the leakage confidence coefficient of the high-pressure heater exceeds the set threshold value, and calculating the leakage confidence coefficient of the high-pressure heater through the similarity between a vector formed by the deviation of each steam extraction and drainage flow of the first safety judgment unit and a vector formed by converting the deviation of each actual valve opening value and an expected valve opening value of the second safety judgment unit into a flow deviation.
7. A computer-readable storage medium, in which a computer program is stored, wherein the computer program is configured to perform the method of any one of claims 1 to 4 when the computer program runs.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010623173.4A CN111780933B (en) | 2020-07-01 | 2020-07-01 | Method and system for diagnosing leakage fault of high-pressure heater based on neural network and thermodynamic modeling |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010623173.4A CN111780933B (en) | 2020-07-01 | 2020-07-01 | Method and system for diagnosing leakage fault of high-pressure heater based on neural network and thermodynamic modeling |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111780933A CN111780933A (en) | 2020-10-16 |
| CN111780933B true CN111780933B (en) | 2022-04-15 |
Family
ID=72761619
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010623173.4A Active CN111780933B (en) | 2020-07-01 | 2020-07-01 | Method and system for diagnosing leakage fault of high-pressure heater based on neural network and thermodynamic modeling |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111780933B (en) |
Families Citing this family (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112924115A (en) * | 2021-03-16 | 2021-06-08 | 中电华创(苏州)电力技术研究有限公司 | Device and method for monitoring internal leakage of high-temperature and high-pressure pipeline valve |
| CN112880919A (en) * | 2021-03-25 | 2021-06-01 | 中电华创电力技术研究有限公司 | Method and device for detecting pipe side leakage of high-pressure heater of steam turbine system |
| CN113218596B (en) * | 2021-04-29 | 2023-09-05 | 中电华创电力技术研究有限公司 | Leakage detection method and device for #4 low-pressure heater of coal-fired power generation unit |
| CN114186485B (en) * | 2021-11-30 | 2024-09-13 | 西安热工研究院有限公司 | Low-voltage heater leakage fault detection method, system, equipment and medium |
| CN114923287B (en) * | 2022-04-14 | 2023-10-13 | 西安热工研究院有限公司 | High-voltage heater fault early warning method, system and medium |
| CN116717734B (en) * | 2023-08-10 | 2023-10-20 | 牧马人(山东)勘察测绘集团有限公司 | Data monitoring method for steady-state operation of natural gas pipeline |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2291970C1 (en) * | 2005-08-09 | 2007-01-20 | Государственное образовательное учреждение высшего профессионального образования "Ульяновский государственный технический университет" | Method for operation of thermal power station |
| CN101865710A (en) * | 2010-05-26 | 2010-10-20 | 北京航空航天大学 | Method for measuring flow of negative-pressure gas |
| US20120185728A1 (en) * | 2010-12-24 | 2012-07-19 | Commonwealth Scientific And Industrial Research Organisation | System and method for detecting and/or diagnosing faults in multi-variable systems |
| CN104568344A (en) * | 2014-10-09 | 2015-04-29 | 芜湖扬宇机电技术开发有限公司 | EGR (Exhaust Gas Recirculation) pipeline fault detection diagnosis method |
| WO2017031996A1 (en) * | 2015-08-27 | 2017-03-02 | 广州神马移动信息科技有限公司 | Method and device for calculating similarity of search terms, searching method and device using search terms |
| CN109459195A (en) * | 2018-09-05 | 2019-03-12 | 中国神华能源股份有限公司 | For judging the method and system of high-pressure heater system leak |
| CN109670265A (en) * | 2018-12-29 | 2019-04-23 | 华润电力技术研究院有限公司 | High-pressure heater modeling method, device and computer equipment |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110288135B (en) * | 2019-06-10 | 2022-10-18 | 华北电力大学 | Drainage water level energy-saving optimization method for high-pressure heating system |
-
2020
- 2020-07-01 CN CN202010623173.4A patent/CN111780933B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| RU2291970C1 (en) * | 2005-08-09 | 2007-01-20 | Государственное образовательное учреждение высшего профессионального образования "Ульяновский государственный технический университет" | Method for operation of thermal power station |
| CN101865710A (en) * | 2010-05-26 | 2010-10-20 | 北京航空航天大学 | Method for measuring flow of negative-pressure gas |
| US20120185728A1 (en) * | 2010-12-24 | 2012-07-19 | Commonwealth Scientific And Industrial Research Organisation | System and method for detecting and/or diagnosing faults in multi-variable systems |
| CN104568344A (en) * | 2014-10-09 | 2015-04-29 | 芜湖扬宇机电技术开发有限公司 | EGR (Exhaust Gas Recirculation) pipeline fault detection diagnosis method |
| WO2017031996A1 (en) * | 2015-08-27 | 2017-03-02 | 广州神马移动信息科技有限公司 | Method and device for calculating similarity of search terms, searching method and device using search terms |
| CN109459195A (en) * | 2018-09-05 | 2019-03-12 | 中国神华能源股份有限公司 | For judging the method and system of high-pressure heater system leak |
| CN109670265A (en) * | 2018-12-29 | 2019-04-23 | 华润电力技术研究院有限公司 | High-pressure heater modeling method, device and computer equipment |
Non-Patent Citations (4)
| Title |
|---|
| Revisiting heater size sensitive pool boiling critical heat flux using neural network modeling_ Heater length of the half of the Rayleigh-Taylor Instability Wavelength maximizes CHF;Mingfu He等;《Thermal Science and Engineering Progress》;20191231;1-7 * |
| 冲突证据的相似性度量方法及其在信息融合故障诊断中的应用;王玉成;《中国优秀硕士学位论文全文数据库 信息科技辑》;20110915(第9期);I140-120 * |
| 提高火电厂高压加热器经济性能方法研究;康莉;《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》;20150215(第2期);C042-672 * |
| 柴油机燃油系统控制锥阀空化流动特性试验研究;冯祥;《中国优秀博硕士学位论文全文数据库(硕士) 工程科技Ⅱ辑》;20170315(第3期);C039-349 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111780933A (en) | 2020-10-16 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN111780933B (en) | Method and system for diagnosing leakage fault of high-pressure heater based on neural network and thermodynamic modeling | |
| CN105955069B (en) | A kind of nuclear power plant system grade state monitoring method based on in-circuit emulation | |
| CN106773666B (en) | Model parameter obtaining method for pressurized water reactor primary circuit system | |
| CN106682376B (en) | Whole-process steam turbine modeling and identification method for actual characteristics of parameters changing along with working conditions | |
| Fast et al. | Development and multi-utility of an ANN model for an industrial gas turbine | |
| CN109471420B (en) | CVA-SFA-based method for monitoring control performance of air preheater of large coal-fired power generator set of intelligent power plant | |
| CN109491358B (en) | Control performance monitoring method for boiler dynamic information of million-kilowatt ultra-supercritical unit | |
| CN105527113B (en) | A kind of nuclear power station heat energy efficiency monitoring and fault diagnosis system and method | |
| CN104500979A (en) | Online diagnosis and elimination system for steam flowing retention section in heating pipe network and working method | |
| CN111626003B (en) | Heating system heat load layered prediction method, storage medium and prediction equipment | |
| CN111898816A (en) | Dynamic state estimation method for comprehensive energy system | |
| CN110795845A (en) | Heat supply system margin quantitative calculation method and system | |
| Wang et al. | Deep-learning modeling and control optimization framework for intelligent thermal power plants: A practice on superheated steam temperature | |
| CN105865220B (en) | A kind of operation method of double pressure condenser optimized operating device | |
| CN117704349B (en) | Low-load hydrodynamic self-adaptive regulation and control method and device for coal-fired power plant boiler | |
| CN108388922A (en) | A kind of recognition methods and system for the major-minor equipment state exception of fired power generating unit | |
| CN117167721A (en) | Method and system for controlling overtemperature of early warning boiler superheater based on improved GRNN algorithm | |
| CN111695300A (en) | Steam turbine valve flow fitting method and system | |
| CN102967393A (en) | Supercritical unit boiler separator outlet temperature on-line calibration method | |
| CN113756886A (en) | Fault judgment method and system for steam turbine heater | |
| CN114048667A (en) | Fault early warning method | |
| CN118035929B (en) | Power plant safety monitoring and early warning method and system based on thermodynamics and data mining | |
| CN117054794B (en) | Water heater detection method and device | |
| Guo et al. | Monitoring of Operation Efficiency Module of Coal Fired Power Plant | |
| Tang et al. | Computer Prediction Model of Heat Consumption in Thermal System of Coal-Fired Power Station Based on Big Data Analysis and Information Sorting |
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 |