Steam turbine through-flow part performance diagnosis method and system based on knowledge graph

Abstract

The invention discloses a steam turbine through-flow part performance diagnosis method and system based on a knowledge graph, and relates to the technical field of thermal power generation. The key point of current research is to find the performance degradation phenomenon of the steam turbine as soon as possible and complete reason tracing. According to the invention, the preliminary performance diagnosis of the through-flow part of the steam turbine is carried out by calculating the performance diagnosis parameter and the reference value of the through-flow part of the steam turbine in real time and judging whether the deviation exceeds the set threshold, which is simple and rapid and has strong mechanism property; through a knowledge graph retrieval method, the rapid and accurate positioning of the performance degradation position of the through-flow part of the steam turbine and the tracing of degradation reasons are realized, the degradation position can be accurately positioned, and the degradation reasons can be found out.

Description

Technical Field

[0001] The present invention relates to the field of thermal power generation technology, and in particular to a method and system for diagnosing the performance of a flow-through part of a steam turbine based on a knowledge graph. Background Art

[0002] When the steam turbine is running, the flow section is in a high temperature and high pressure environment for a long time, which is prone to performance degradation. Its performance directly affects the power generation efficiency and operation safety of the steam turbine. In this context, it is of great significance to carry out performance monitoring and diagnosis of the flow section of the steam turbine.

[0003] At present, most of the work on the flow section of steam turbines is mainly focused on operational fault diagnosis. Since the steam turbine itself is a complex system, there is a strong coupling between the thermal parameters of each stage. The parameter changes of some stages may be caused by faults in other stages. There is also a phenomenon that a single data seems to be normal, but the performance of the flow section of the steam turbine has deteriorated. In addition, some fault diagnosis methods have high requirements on the professionalism and engineering experience of the crew in practical applications, and there are problems such as difficulty in tracing the source of the fault, strong subjectivity and poor systematization.

[0004] In response to this complex problem, patent document CN111365083B determines the fault location based on the results of stage group efficiency and characteristic flow area, and determines whether the stage group has actually failed based on the corresponding relationship between valve changes and stage group efficiency and characteristic flow area, and then forms a diagnosis report through fault database comparison and expert analysis. Patent document CN116089877A uses the KNN model to determine and output the fault result by calculating the distance and distance weight between the characteristic flow area of ​​the current period and the characteristic flow area of ​​the preset fault sample.

[0005] Although there are a large number of open literatures proposing various feasible methods for diagnosing faults in the flow section of steam turbines, early detection of turbine performance degradation and tracing of the causes remain the current research focus. Summary of the invention

[0006] In order to solve the above-mentioned problems existing in the prior art, the present invention provides a method and system for diagnosing the performance of the flow part of a turbine based on a knowledge graph, so as to realize the rapid and accurate positioning of the performance degradation position of the flow part of the turbine and the tracing of the cause.

[0007] In order to achieve the above-mentioned object of the invention, the present invention adopts the following technical scheme.

[0008] In the first aspect, the present invention provides a method for diagnosing the performance of the flow part of a turbine based on a knowledge graph, which calculates the performance diagnostic parameters and reference values ​​of the flow part of the turbine in real time, and performs preliminary performance diagnosis of the flow part of the turbine when the deviation between the two exceeds a set threshold; through a knowledge graph retrieval method, the performance degradation position of the flow part of the turbine can be quickly and accurately located and the cause of the degradation can be traced.

[0009] Furthermore, the steam turbine flow part performance diagnosis method comprises the following specific steps:

[0010] Step 1: Establish a thermal system mechanism model of the unit according to the heat balance diagram data of the unit's design operating conditions;

[0011] Step 2: real-time collection of unit operation data under current working conditions;

[0012] Step 3, judging the steady-state operating condition of the unit operation data collected in step 2, and obtaining the steady-state operating condition data of the unit through the judgment;

[0013] Step 4, calculating the enthalpy and entropy of water and steam at each measuring point under the steady-state condition, calculating the regenerative steam extraction amount by the heat balance of the regenerative heater, calculating the inlet steam flow of each section of the turbine flow section by the flow balance, and completing the calculation of the performance diagnosis parameters of the turbine flow section;

[0014] Step 5, according to the main steam temperature and pressure of the steam turbine, the reheat steam temperature, the feed water flow rate and the power generation data of the unit under the steady-state condition, a simulation calculation of the variable condition is performed on the basis of the thermal system mechanism model of the unit, and the performance diagnostic parameters of the unit under the corresponding load condition are calculated as the reference value;

[0015] Step 6, calculating the deviation between the turbine flow-through performance diagnostic parameter obtained in step 4 and the reference value;

[0016] Step 7: Determine whether any deviation exceeds the set threshold. If so, perform knowledge graph retrieval method diagnosis; if not, return to step 3.

[0017] Step eight, locate the performance degradation position of the flow passage part of the turbine through the knowledge graph retrieval method and diagnose the cause of the performance degradation.

[0018] Furthermore, in the step 2, the unit operation data includes the temperature and pressure of the main steam and reheated steam, the temperature and pressure of each stage of extraction steam, the temperature and pressure of each cylinder exhaust steam, the inlet and outlet water temperature and drain temperature of each regenerative heater, the feed water flow rate, the condensate flow rate and the unit power generation capacity.

[0019] Furthermore, in the step three, the main steam temperature and pressure of the turbine, the reheat steam temperature, the feed water flow rate and the unit power generation capacity are used as the basis for judging the steady-state operating conditions, and the sliding window method is used for screening. The real-time collected unit operation data are divided in chronological order. If the normalized standard deviation of the samples in the window is within the preset threshold range, it is regarded as a steady-state operating condition.

[0020] Furthermore, in step 3, the calculation formula of the normalized standard deviation is as follows:

[0021]

[0022] In the formula, δ is the normalized standard deviation; is the average value of N sample data in the sliding window; ε is the preset threshold of normalized standard deviation; x i Represents the i-th sample data.

[0023] Furthermore, in the step 4, the performance diagnostic parameters of the turbine flow section include the stage efficiency of each stage of the turbine flow section, the characteristic flow area of ​​each stage, the cylinder efficiency and the heat rate of the single reheat unit.

[0024] Furthermore, in the step 4, the flow section of the steam turbine is divided into sections according to the steam extraction port and the steam exhaust port.

[0025] Furthermore, in the step seven, the knowledge graph involves knowledge related to the diagnosis of performance degradation of the flow part of the turbine, specifically including performance degradation characteristics, turbine stages, performance degradation cause entities and their associations; based on the deviation obtained in step six, it is mapped to the corresponding entity in the knowledge graph through preset rules, and the path search is started with these entities as the starting point to infer and locate the cause of performance degradation.

[0026] Furthermore, in the step eight, the performance degradation of the flow part of the turbine includes scaling degradation, clogging degradation and corrosion and wear degradation. When scaling degradation occurs, the flow area is reduced, and the optimal enthalpy drop distribution of the turbine will be changed, resulting in reduced stage efficiency and changes in the critical pressure ratio; when clogging degradation occurs, the flow area will be reduced, the stage flow rate will drop significantly, and the stage efficiency will be significantly reduced; when corrosion and wear degradation occurs, the stage flow area will increase, and the loss within the stage will also increase.

[0027] In a second aspect, the present invention provides a steam turbine flow part performance diagnosis system based on a knowledge graph, which is used to implement the above-mentioned steam turbine flow part performance diagnosis method.

[0028] The present invention has the following beneficial effects:

[0029] 1. The present invention obtains the performance diagnostic parameters of the flow-through part of the unit by processing the real-time steady-state data, and makes a preliminary judgment on the performance degradation of the flow-through part of the unit by comparing with the benchmark value. Compared with the complicated method of predicting the future performance development trend of the unit based on real-time data, the present invention is simple, fast and has strong mechanism.

[0030] 2. The present invention uses knowledge graph technology to perform performance diagnosis, which can realize the diagnosis and early warning of the early performance degradation stage of the turbine flow section. Compared with case rule base query and judgment methods, the use of knowledge graph technology, a new type of knowledge relationship retrieval method, is more conducive to accurate analysis from the system level in the complex correlation scenario of turbine flow section degradation diagnosis, precise positioning of degradation locations, and identification of degradation causes. BRIEF DESCRIPTION OF THE DRAWINGS

[0031] Figure 1 It is a flow chart of the method for diagnosing the performance of the flow-through part of a steam turbine according to the present invention;

[0032] Figure 2 It is a schematic diagram of visualizing the knowledge graph of the present invention;

[0033] Figure 3 Schematic diagram of the basis for dividing the stages of the unit in an embodiment of the present invention. DETAILED DESCRIPTION

[0034] The present invention will be further described below in conjunction with the accompanying drawings and specific implementation methods.

[0035] See also Figure 1 This embodiment provides a method for diagnosing the performance of a flow-through part of a steam turbine based on a knowledge graph, and the steps are as follows:

[0036] Step 1: Establish a thermal system mechanism model of the unit according to the heat balance diagram data of the unit's design operating conditions;

[0037] Step 2: Collect the unit operation data under the current working conditions in real time, including the temperature and pressure of the main steam and reheat steam, the temperature and pressure of each stage of extraction steam, the temperature and pressure of each cylinder exhaust steam, the inlet and outlet water temperature and drain temperature of each regenerative heater, the feed water flow, the condensate flow and the unit power generation, etc.;

[0038] Step 3: The steady-state operating condition of the unit operation data collected in step 2 is judged, and the parameters such as the main steam temperature and pressure of the steam turbine, the reheat steam temperature, the feed water flow rate and the unit power generation are used as the basis for judging the steady-state operating condition. If it is judged to be steady-state operating condition data, go to step 4; if it is judged to be non-steady-state data, re-judge the next set of data;

[0039] Step 4, calculating the enthalpy and entropy of water and steam at each measuring point under the steady-state condition, calculating the regenerative steam extraction amount by the heat balance of the regenerative heater, calculating the inlet steam amount of each section of the steam turbine flow section by the flow balance, and completing the calculation of the performance diagnosis parameters of the steam turbine flow section;

[0040] Step 5, according to the main steam temperature and pressure of the steam turbine, the reheat steam temperature, the feed water flow rate and the power generation data of the unit under the steady-state condition, a simulation calculation of the variable condition is performed on the basis of the thermal system mechanism model of the unit, and the performance diagnostic parameters of the unit under the corresponding load condition are calculated as the reference value;

[0041] Step 6, calculating the deviation between the turbine flow-through performance diagnostic parameter obtained in step 4 and the reference value;

[0042] Step 7: Determine whether any deviation exceeds the set threshold. If so, perform knowledge graph retrieval method diagnosis; if not, return to step 3.

[0043] Step 8, see Figure 2 , through the knowledge graph, search, locate the performance degradation position of the flow part of the turbine, and diagnose the cause of the performance degradation.

[0044] Specifically, in step 3, the present invention uses a sliding window method to perform screening, and divides the real-time collected unit operation data in time order. If the normalized standard deviation of the samples in the window is within the preset threshold range, it is considered to be a steady-state condition. The normalized standard deviation formula is as follows:

[0045]

[0046] In the formula, δ is the normalized standard deviation; is the average value of N sample data in the sliding window; ε is the preset threshold of normalized standard deviation; x i Represents the i-th sample data. When the above formula is satisfied, it indicates that the N-th operating condition meets the data steady-state screening requirements.

[0047] In this embodiment, the steam turbine of a power plant is a 1000MW unit, an ultra-supercritical, single-intermediate reheat, four-cylinder, four-exhaust, single-shaft, indirect air-cooled condensing steam turbine of model NJK1000-28 / 600 / 620.

[0048] In step 4, in order to accurately locate the degradation location in the subsequent steps, see Figure 3, the turbine flow section is divided into flow sections by extraction and exhaust ports, and the flow section of the unit is divided into 12 sections, namely, main steam to the first extraction section, the first extraction section to the high-pressure cylinder exhaust section, the medium-pressure cylinder steam inlet to the third extraction section, the medium-pressure cylinder steam inlet to the fourth extraction section, the third extraction section to the medium-pressure cylinder exhaust section, the fourth extraction section to the medium-pressure cylinder exhaust section, the low-pressure cylinder steam inlet to the seventh extraction section, the seventh extraction section to the ninth extraction section, the ninth extraction section to the low-pressure cylinder exhaust, the low-pressure cylinder steam inlet to the sixth extraction section, the sixth extraction section to the eighth extraction section, and the eighth extraction section to the low-pressure cylinder exhaust. This segmentation method can more accurately locate the location where the unit flow section degradation occurs by refining the segment division in the high-pressure cylinder, the medium-pressure cylinder, and the low-pressure cylinder.

[0049] In step 4, the performance diagnostic parameters of the steam turbine flow section involved include the stage efficiency of each stage of the steam turbine flow section, the characteristic flow area of ​​each stage, the cylinder efficiency and the unit heat rate.

[0050] Among them, the stage efficiency is calculated using the isentropic enthalpy drop method, and the calculation formula is:

[0051]

[0052] Where η ri is the stage efficiency, %; H ri is the steam inlet enthalpy of the stage, kJ / kg; H ro is the steam enthalpy value of the stage, kJ / kg; ΔH r0 is the ideal enthalpy drop of the stage, kJ / kg.

[0053] The calculation formula of the characteristic flow area of ​​the stage is:

[0054]

[0055] In the formula, is the characteristic flow area of ​​the stage, m 2 ; G is the steam flow rate of the stage, kg / h; v0 is the steam specific volume of the stage, m 3 / kg; p0 is the absolute pressure of steam before the stage, MPa; π is the pressure ratio of the inlet and outlet pressures of the stage.

[0056] The cylinder efficiency calculation formula is:

[0057]

[0058] Where η i is the cylinder efficiency, %; H i is the cylinder steam inlet enthalpy, kJ / kg; H o is the cylinder exhaust enthalpy, kJ / kg; ΔH o is the ideal enthalpy drop of the cylinder, kJ / kg.

[0059] The calculation formula for the heat rate of a single reheat unit is:

[0060]

[0061] Where HR is the heat rate, kJ / kWh; P is the generator output power, kW; G MS , H MS are the main steam flow rate and enthalpy respectively; G HRH , H HRH are the reheat steam flow rate and enthalpy respectively; G FW , H FW are water flow rate and enthalpy respectively; G CRH , H CRH are the cold reheat steam flow rate and enthalpy respectively; G RH , H RH They are the reheater desuperheating water flow rate and enthalpy value respectively; the flow rate unit is kg / h, and the enthalpy value unit is kJ / kg.

[0062] In the step 7, the knowledge graph retrieval method is to represent the actual document-type knowledge and extract it into a triple structure consisting of a head entity, a relationship, and a tail entity, which can be more conveniently used for tasks such as knowledge reasoning and knowledge retrieval. The knowledge graph involves knowledge related to the diagnosis of performance degradation of the flow-through part of the turbine, specifically including performance degradation characteristics, turbine stages, performance degradation cause entities and their associations; based on the deviation obtained in step 6, it is mapped to the corresponding entity in the knowledge graph through preset rules, and the path search is started with these entities as the starting point to reason and locate the cause of performance degradation.

[0063] In the step eight, it can be known from the analysis of the thermal system mechanism that the performance degradation of the flow part of the turbine includes scaling degradation, clogging degradation and corrosion and wear degradation. When scaling degradation occurs, the flow area is reduced, and the optimal enthalpy drop distribution of the turbine will be changed, resulting in reduced stage efficiency and changes in the critical pressure ratio; when clogging degradation occurs, the flow area will be reduced, the stage flow rate will drop significantly, and the stage efficiency will be significantly reduced; when corrosion and wear degradation occurs, the stage flow area will increase, and the loss within the stage will also increase.

[0064] This embodiment also provides a knowledge graph-based steam turbine flow part performance diagnosis system for implementing the above-mentioned steam turbine flow part performance diagnosis method.

[0065] The above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application is described in detail with reference to the above embodiments, a person skilled in the art should understand that the technical solutions described in the above embodiments can still be modified, or some of the technical features can be replaced by equivalents, but these modifications or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A method for diagnosing the performance of a steam turbine flow section based on a knowledge graph, characterized in that: By calculating the performance diagnostic parameters and benchmark values ​​of the steam turbine flow section in real time, a preliminary performance diagnosis of the steam turbine flow section is performed when the deviation between the two exceeds the set threshold; through the knowledge graph retrieval method, the performance degradation position of the steam turbine flow section can be quickly and accurately located and the cause of the degradation can be traced.

2. The method for diagnosing the performance of the flow section of a steam turbine according to claim 1, characterized in that: The specific steps include: Step 1: Establish a thermal system mechanism model of the unit according to the heat balance diagram data of the unit's design operating conditions; Step 2: real-time collection of unit operation data under current working conditions; Step 3, judging the steady-state operating condition of the unit operation data collected in step 2, and obtaining the steady-state operating condition data of the unit through the judgment; Step 4, calculating the enthalpy and entropy of water and steam at each measuring point under the steady-state condition, calculating the regenerative steam extraction amount by the heat balance of the regenerative heater, calculating the inlet steam amount of each section of the steam turbine flow section by the flow balance, and completing the calculation of the performance diagnosis parameters of the steam turbine flow section; Step 5, according to the main steam temperature and pressure of the steam turbine, the reheat steam temperature, the feed water flow rate and the power generation data of the unit under the steady-state condition, a simulation calculation of the variable condition is performed on the basis of the thermal system mechanism model of the unit, and the performance diagnostic parameters of the unit under the corresponding load condition are calculated as the reference value; Step 6, calculating the deviation between the turbine flow-through performance diagnostic parameter obtained in step 4 and the reference value; Step 7: determine whether any deviation exceeds a set threshold. If so, perform a knowledge graph retrieval method diagnosis. If it does not exceed the set threshold, return to step 3; Step eight, locate the performance degradation position of the flow passage part of the turbine through the knowledge graph retrieval method and diagnose the cause of the performance degradation.

3. The method for diagnosing the performance of the flow section of a steam turbine according to claim 2, characterized in that: In the step 2, the unit operation data includes the temperature and pressure of the main steam and reheat steam, the temperature and pressure of each stage of extraction steam, the temperature and pressure of each cylinder exhaust steam, the inlet and outlet water temperature and drain temperature of each regenerative heater, the feed water flow rate, the condensate flow rate and the unit power generation capacity.

4. The method for diagnosing the performance of the flow section of a steam turbine according to claim 3, characterized in that: In the step three, the main steam temperature and pressure of the turbine, the reheat steam temperature, the feed water flow rate and the unit power generation capacity are used as the basis for judging the steady-state operating conditions. The sliding window method is used for screening, and the real-time collected unit operation data are divided in chronological order. If the normalized standard deviation of the samples in the window is within the preset threshold range, it is regarded as a steady-state operating condition.

5. The method for diagnosing the performance of the flow section of a steam turbine according to claim 4, characterized in that: In the step 3, the calculation formula of the normalized standard deviation is as follows: Where δ is the normalized standard deviation; is the average value of N sample data in the sliding window; ε is the preset threshold of normalized standard deviation; x i Represents the i-th sample data.

6. The method for diagnosing the performance of the flow section of a steam turbine according to claim 2, characterized in that: In the step 4, the performance diagnostic parameters of the turbine flow section include the stage efficiency of each stage of the turbine flow section, the characteristic flow area of ​​each stage, the cylinder efficiency and the heat rate of the single reheat unit.

7. The method for diagnosing the performance of a steam turbine flow passage according to claim 2, characterized in that: In the step 4, the flow section of the steam turbine is divided into sections according to the steam extraction port and the steam exhaust port.

8. The method for diagnosing the performance of a steam turbine flow section according to claim 2, characterized in that: In the step seven, the knowledge graph includes performance degradation characteristics, turbine stages, performance degradation cause entities and their associations; based on the deviations obtained in step six, they are mapped to corresponding entities in the knowledge graph through preset rules, and path search is started with these entities as starting points to infer and locate the causes of performance degradation.

9. The method for diagnosing the performance of a steam turbine flow passage according to claim 2, characterized in that: In the step eight, the performance degradation of the flow part of the turbine includes scaling degradation, clogging degradation and corrosion and wear degradation. When scaling degradation occurs, the flow area is reduced, and the optimal enthalpy drop distribution of the turbine will be changed, resulting in reduced stage efficiency and changes in the critical pressure ratio; when clogging degradation occurs, the flow area will be reduced, the stage flow rate will drop significantly, and the stage efficiency will be significantly reduced; when corrosion and wear degradation occurs, the stage flow area will increase, and the loss within the stage will also increase.

10. A steam turbine flow performance diagnosis system based on knowledge graph, characterized in that: Used to implement the turbine flow part performance diagnosis method described in any one of claims 1-9.

Images