Cloud-edge cooperative fault diagnosis method and device for gas generator set, and medium

Abstract

The invention discloses a cloud edge cooperative fault diagnosis method for a gas generator set, and the method comprises the following steps: S1, obtaining fault criteria and symptoms of a gas turbine based on domain knowledge, carrying out the fault diagnosis analysis of the gas turbine based on technology fusion according to multiple criteria and symptoms, and building a logic rule database of different faults; s2, establishing a gas turbine modular simulation model, and performing energy efficiency diagnosis and optimization operation on the simulation model based on the logic rule database; s3, analyzing a fault influence mechanism of the optimized simulation model, and evaluating the service life of the core component of the gas turbine based on the fault influence mechanism of the simulation model; and S4, performing overall health assessment on the gas turbine based on the service life assessment result of each core component. According to the method, the effect of accurately mapping the relationship between the equipment health state and the working condition parameters can be achieved, so that the effect of comprehensively and dynamically evaluating the health level of the gas generator set is achieved, and the effect of efficiently and accurately diagnosing the fault of the gas turbine is achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of gas generator set fault diagnosis, and in particular to a cloud-edge collaborative fault diagnosis method, device and medium for a gas generator set. Background Art

[0002] As a key power equipment, the efficient and stable operation of heavy-duty gas turbine generator sets is directly related to the safety and economy of the power grid. Traditional gas turbine operation and maintenance mainly relies on the on-site DCS control system to implement real-time monitoring. When a parameter is detected to exceed the preset threshold, the system will trigger an alarm, and then the operator will judge and handle the fault based on experience and alarm information. In addition, regular maintenance inspections are also an important part of ensuring the health of the unit. This is usually based on a fixed maintenance plan rather than real-time status monitoring. In this mode, the depth of data analysis is limited, and the fault prediction and health management (PHM) capabilities are weak. Often, a response is not made until a fault occurs, affecting the unit's operating efficiency and reliability.

[0003] Although the operation and maintenance management of gas turbine generator sets has gradually introduced remote monitoring and diagnosis platforms, the following problems still exist: (1) In the traditional mode, fault diagnosis and processing are highly dependent on the subjective judgment of operators, which easily leads to the lack of objectivity and accuracy of fault diagnosis; (2) The existing DCS system can only alarm for single parameter out-of-limit, lacks comprehensive analysis and cross-validation capabilities, and is difficult to give accurate warnings at the early stage of faults, missing the best time for intervention, and easily causing huge waste of data resources; (3) There are limitations in algorithm accuracy and data fusion capabilities, and a large amount of operation data has not been effectively mined and utilized, resulting in a huge waste of data resources and failing to be converted into actual value for improving operation and maintenance efficiency and preventive maintenance. Its actual effectiveness and warning accuracy need to be improved.

[0004] To this end, this application specifically proposes a cloud-edge collaborative fault diagnosis method for gas generator sets to solve the above-mentioned technical problems.

[0005] The above contents are only used to assist in understanding the technical solution of the present application, and do not constitute an admission that the above contents are the closest prior art. Summary of the invention

[0006] The main purpose of the present invention is to provide a cloud-edge collaborative fault diagnosis method for a gas generator set to solve the technical problems raised in the background technology.

[0007] The present invention adopts the following technical solutions to solve the above technical problems:

[0008] A cloud-edge collaborative fault diagnosis method for a gas generator set, wherein a computer device performs the following steps:

[0009] S1. Obtain the fault criteria and symptoms of gas turbines based on domain knowledge, and perform fault diagnosis and analysis on gas turbines based on technology fusion based on multiple criteria and symptoms, and establish a logical rule database for different faults;

[0010] S2. Establish a modular simulation model of gas turbines, and perform energy efficiency diagnosis and optimization operations on the simulation model based on a logical rule database;

[0011] S3. Analyze the failure impact mechanism of the optimized simulation model, and evaluate the life of the core components of the gas turbine based on the failure impact mechanism of the simulation model;

[0012] S4. Conduct an overall health assessment of the gas turbine based on the life assessment results of each core component.

[0013] Preferably, the specific operation flow of performing fault diagnosis and analysis on the gas turbine based on technology fusion according to multiple criteria and symptoms in step S1 and establishing a logic rule database for different faults includes:

[0014] S11. Analyze the signs and criteria of typical failure modes in gas generator sets in combination with expert domain knowledge, and comprehensively analyze the results to form a standard fault diagnosis information library;

[0015] S12. Based on the analysis results of the faults by domain experts, various fault judgment modes and fault symptom criteria are obtained as fault knowledge, and the quantitative indicator criteria and qualitative symptom criteria of each fault are integrated to realize multi-dimensional intelligent diagnosis of each fault;

[0016] S13. Based on the multi-dimensional intelligent diagnosis results of each fault, a preliminary model of logic rules for different faults is established, and the logic rules are centrally managed to establish a logic rule database for different faults.

[0017] Preferably, the specific operation process of analyzing the signs and criteria of typical failure modes in the gas generator set in step S11 includes:

[0018] S111. Summarize the actual faults and classify them;

[0019] S112. Determine the degree of membership between each fault symptom and the fault cause based on expert experience or statistical methods, and establish a fault membership table;

[0020] S113. Establish a fuzzy relationship matrix based on the membership table and calculate the fuzzy relationship to draw a conclusion as the fault diagnosis result.

[0021] Preferably, the specific operation process of integrating the quantitative indicator criteria and the qualitative symptom criteria of each fault in step S12 includes:

[0022] S121. Determine the main fault warning range through the preliminary parameters of the gas generator set, disassemble and analyze the gas generator set equipment to build an intelligent warning model;

[0023] S122 develops diagnostic model monitoring parameters based on the analysis of qualitative signs for typical fault types, builds a parameter prediction model based on the intelligent early warning model, and compares the prediction results with the actual operation results to obtain changes in parameters such as increase, decrease, and fluctuation, and converts the monitoring parameters into fault signs;

[0024] S123. According to the quantitative indicators of typical faults, based on the characteristic simulation of gas turbine equipment and the gas path performance degradation analysis, a gas turbine real-time performance simulation model is established to monitor the degradation of key indicators in real time;

[0025] S124. Establish a logical reasoning model based on the fault diagnosis information database and integrate the above qualitative and quantitative signs.

[0026] Preferably, the specific operation process of step S2 includes:

[0027] S21. For key components of gas generator sets, mathematical models are established respectively, and the mathematical models are modified based on the characteristics of the components;

[0028] S23. Analyze the impact of air intake system degradation on the thermal performance of the gas turbine, and optimize the performance of the key components of the gas turbine and the simulation model simultaneously based on the actual operating data of the gas turbine.

[0029] Preferably, the specific operation process of analyzing the fault impact mechanism of the optimized simulation model in step S3 includes:

[0030] S31. Establish a numerical simulation model for NOx emissions during the startup process of F-class heavy-duty gas turbines;

[0031] S32. Analyze the influence of fuel distribution mode on NOx emission concentration under PPM mode based on numerical simulation model and obtain the influence mechanism;

[0032] S33. Based on the numerical simulation model, the dynamic pressure changes at specific positions of the combustion chamber under high load conditions, low load conditions, and high load conditions with the same nozzle flow in PPM mode are analyzed to optimize the influencing mechanism.

[0033] Preferably, the specific operation process of performing the life assessment of the core components of the gas turbine in step S3 includes:

[0034] L1. Using the standard parameters of the gas turbine and the actual operating condition data, after data filtering, the turbine gas path section of the entire gas turbine is thermodynamically analyzed to calculate some missing information, quantify the heat load and load of the core components of the gas turbine, the appropriate benchmark parameters of the core components of the gas turbine, and create boundary conditions for the CFD analysis of the core components. The boundary conditions include the distribution of the values ​​of gas path temperature, pressure and mass flow;

[0035] L2. Perform CFD analysis on the core components of the gas turbine. Based on the multi-level CFD calculation results, combined with the geometric files and material parameters of the core components, conduct thermal fluid-solid coupling analysis under different working conditions, realize load processing and data conversion of temperature load and aerodynamic load, obtain accurate cooling characteristics and strength characteristics, and then obtain the stress and strain distribution of the blades through finite element analysis;

[0036] L3. By considering the creep, high cycle fatigue and low cycle fatigue loads of core components under basic load conditions, as well as the coupling effects of loads, a life analysis algorithm model of creep, high cycle fatigue and low cycle fatigue loads of core components is constructed based on the influence mechanism, and the life loss of core components is analyzed to obtain its estimated remaining life and reliability status, so as to determine the failure mode, potential failure location and component life of the components;

[0037] L4. Establish a probabilistic analysis model based on the estimated remaining life and reliability status, conduct probabilistic analysis on the parameter variability and uncertainty during the operation of the gas turbine and their impact, and obtain deterministic calculation results;

[0038] L5. Based on the deterministic calculation results in the above steps, the physical field is obtained as the nominal field for probability analysis, and parametric modeling, Monte Carlo-based Latin hypercube sampling, and variance reduction are performed in sequence for probability simulation;

[0039] L6. Based on the crack length, depth, fatigue and creep crack growth rate, and crack instability length of the component, the remaining life of the component is further determined based on the damage tolerance analysis and probabilistic simulation results of the structural integrity, and the optimal maintenance interval is formulated.

[0040] Preferably, the specific operation steps of step S4 include:

[0041] S41. Establish a health assessment system, preset assessment parameters, and construct a health function, where the characteristic numbers in the preset function are a, b, c, d, e, and f;

[0042] S42. Classify and statistically analyze the monitoring values ​​of the evaluation parameters of the gas turbine generator set in normal working state and faulty working state, and select appropriate values ​​as the values ​​of the health function characteristic numbers a, b, c, d, e, and f based on the statistical results;

[0043] S43. Establish the membership function of each system and each evaluation parameter to obtain the membership status;

[0044] S44. Establish a factor set consisting of n influencing factors of the evaluation object:

[0045] U={u 1 ,u 2 ,u 3 ,…,u n}

[0046] Create a review set consisting of m evaluation results as the evaluation level set:

[0047] V={v 1 , v 2 , v 3 , …, v m}

[0048] The comment set is a combination of a series of qualitative evaluations, such as excellent, good, medium, poor, and very poor. The number of elements in the comment set should be appropriate. Too many elements will make the distinction between different levels unclear, and too few elements will make the evaluation unclear.

[0049] Usually 4 to 5;

[0050] S45. The evaluation matrix R representing the fuzzy relationship between the factor set and the evaluation set is obtained as follows:

[0051]

[0052] Among them, r ij (i∈[1,n],j∈[1,m]) represents the sum of the factors u i To see things being rated as level v j degree of affiliation;

[0053] S46. The fuzzy set A on U is used to represent the weight distribution of each factor. The weight distribution of each factor here is realized by using the hierarchical analysis method, which is recorded as:

[0054] A={a 1 , a 2 , a 3 , …, a n}

[0055] Where a i (i∈[1,n]) represents the i-th factor u i The weight of , then the fuzzy comprehensive evaluation membership result is expressed as:

[0056]

[0057] Among them, ° is the fuzzy synthesis operator selected according to the actual situation of the problem;

[0058] S47. For multi-level problems, the calculation starts from the lowest parameter layer, and recursively proceeds layer by layer to finally obtain the evaluation result of the target layer;

[0059] S48. The evaluation results are processed in the membership function to obtain the health of the system.

[0060] On the other hand, the present invention further discloses a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the steps of the above method.

[0061] On the other hand, the present invention further discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the above method.

[0062] It can be seen from the above technical solution that the present invention provides a cloud-edge collaborative fault diagnosis method for a gas generator set. Compared with the prior art, the present invention has the following advantages:

[0063] 1. The present invention combines expert domain knowledge with data analysis technology, sets up a comprehensive analysis process in the early stage of fault diagnosis, and can form a standardized fault diagnosis information library to provide a data basis for intelligent diagnosis, thereby improving the accuracy and efficiency of intelligent diagnosis, and ultimately achieving rapid positioning and processing of gas generator set faults.

[0064] 2. The present invention establishes fault logic rules and centrally manages them, and incorporates a flexible logical reasoning mechanism into the system design, which can increase the adaptability and accuracy of fault diagnosis, thereby achieving the effect of accurately judging the fault type even under complex working conditions, and ultimately realizing intelligent identification and classification management of different fault modes.

[0065] 3. The present invention introduces probability analysis and thermal-fluid-solid coupling methods, considers the random variation and uncertainty of parameters in the life prediction link, and can construct a probability model that is closer to the actual situation, so as to accurately evaluate the remaining life and reliability of core components, facilitate the scientific formulation of preventive maintenance strategies, and reduce the risk of unexpected downtime.

[0066] 4. The present invention uses multi-source information fusion technology and deep data analysis, and utilizes the life assessment results of core components to further establish a health assessment system and calculate the health degree. It can effectively integrate historical and real-time data, and accurately map the relationship between the health status of the equipment and the operating parameters, thereby achieving a comprehensive and dynamic assessment of the health level of the gas generator set, thereby playing a role in efficient and accurate gas turbine fault diagnosis, and ultimately ensuring the long-term stable operation and efficient management of the gas turbine equipment.

[0067] It should be understood that the content described in this section is not intended to identify the key or important features of the embodiments of the present invention, nor is it intended to limit the scope of the present invention. Other features of the present invention will become easy to understand through the following description. Of course, it is not necessary to achieve all of the advantages described above simultaneously for any product implementing the present invention. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] The drawings constituting a part of the present application are used to provide a further understanding of the present invention. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0069] Figure 1 It is a schematic diagram of the overall process of the present invention;

[0070] Figure 2 Establishing a path diagram for the preliminary model of the logic rules of the present invention;

[0071] Figure 3 It is a schematic diagram of the overall model of the gas turbine thermal system of the present invention;

[0072] Figure 4 It is a schematic diagram of the calculation flow of the gas turbine thermal system model of the present invention;

[0073] Figure 5 It is a schematic diagram of the change of the thermal performance of the gas turbine of the present invention as the filter pressure difference deteriorates;

[0074] Figure 6 It is a line segment diagram showing the effect of the fuel distribution method on the NOx emission concentration under the PPM mode of the present invention;

[0075] Figure 7 It is a schematic diagram of the temperature distribution and NOx concentration distribution in the combustion chamber under two different PPM loads of the present invention;

[0076] Figure 8 It is a schematic diagram of the health function line segment form of the present invention. DETAILED DESCRIPTION

[0077] The technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all of the embodiments. In the absence of conflict, the embodiments in this application and the features in the embodiments can be combined with each other. Based on the embodiments in the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0078] In the embodiment, see Figures 1 to 8 .

[0079] like Figure 1 As shown, a cloud-edge collaborative fault diagnosis method for a gas generator set proposed in an embodiment of the present invention mainly includes fault expert knowledge, real-time simulation, machine learning algorithm, etc., and uses logical reasoning technology to integrate multiple algorithms and data to form a multi-algorithm fusion analysis capability that complements data algorithm, mechanism simulation, and logical reasoning. Specifically, the following steps are performed by a computer device:

[0080] S1. Obtain the fault criteria and symptoms of the gas turbine based on domain knowledge, and perform fault diagnosis and analysis on the gas turbine based on multiple criteria and symptoms based on technical integration, and establish a logical rule database for different faults. The specific operation process includes:

[0081] S11. Analyze the signs and criteria of typical fault modes in gas generator sets in combination with expert domain knowledge, and comprehensively analyze the results to form a standard fault diagnosis information database to provide basic support for intelligent fault diagnosis;

[0082] The specific operation process for analyzing the signs and criteria of typical failure modes in gas generator sets includes:

[0083] S111. Summarize the actual faults and classify them;

[0084] S112. Determine the degree of membership between each fault symptom and the fault cause based on expert experience or statistical methods, and establish a fault membership table;

[0085] S113. Establishing a fuzzy relationship matrix based on the membership table and calculating the fuzzy relationship to draw conclusions as fault diagnosis results;

[0086] S12. Based on the analysis results of the faults by domain experts, various fault judgment modes and fault symptom criteria are obtained as fault knowledge, and the quantitative indicator criteria and qualitative symptom criteria of each fault are integrated to realize multi-dimensional intelligent diagnosis of each fault;

[0087] The specific operation process of integrating the quantitative index judgment criteria and qualitative symptom judgment criteria of each fault includes:

[0088] S121. Determine the main fault warning range through the preliminary parameters of the gas generator set, disassemble and analyze the gas generator set equipment to build an intelligent warning model;

[0089] S122 develops diagnostic model monitoring parameters based on the analysis of qualitative signs for typical fault types, builds a parameter prediction model based on the intelligent early warning model, and compares the prediction results with the actual operation results to obtain changes in parameters such as increase, decrease, and fluctuation, and converts the monitoring parameters into fault signs;

[0090] S123. According to the quantitative indicators of typical faults, based on the characteristic simulation of gas turbine equipment and the gas path performance degradation analysis, a gas turbine real-time performance simulation model is established to monitor the degradation of key indicators in real time;

[0091] S124. Establish a logical reasoning model based on the fault diagnosis information database to integrate the above qualitative and quantitative signs;

[0092] The gas turbine performance monitoring and diagnosis method based on fuzzy logic simulates the human uncertainty concept judgment and reasoning thinking mode to monitor and diagnose the performance of the unit. The general method of gas turbine monitoring and diagnosis based on fuzzy logic can be summarized in the above step S11, which is: summarize the actual faults and classify them; obtain the membership between each fault symptom and the fault cause based on expert experience or statistical methods; establish a fault membership table; establish a fuzzy relationship matrix and calculate; draw a conclusion;

[0093] At this time, the logical reasoning and judgment of fault analysis and diagnosis can increase the flexibility of system fault diagnosis and the practicality and accuracy of analysis and diagnosis;

[0094] In summary, by combining expert domain knowledge with data analysis technology and setting up a comprehensive analysis process at the early stage of fault diagnosis, a standardized fault diagnosis information database can be formed to provide a data basis for intelligent diagnosis, thereby improving the accuracy and efficiency of intelligent diagnosis, and ultimately achieving rapid positioning and processing of gas generator set faults;

[0095] S13. Based on the multi-dimensional intelligent diagnosis results of each fault, refer to Figure 2 The path shown can establish a preliminary model of logic rules for different faults, and centrally manage the logic rules to establish a database of logic rules for different faults;

[0096] At this time, by establishing fault logic rules and centrally managing them, and incorporating flexible logical reasoning mechanisms into system design, the adaptability and accuracy of fault diagnosis can be increased, so that the fault type can be accurately judged even under complex working conditions, and ultimately the intelligent identification and classification management of different fault modes can be realized;

[0097] In a specific embodiment, a common fault of a gas turbine system is taken as an example for analysis as follows:

[0098] (1) Compressor blade fouling

[0099] Blade fouling is one of the common faults in the gas turbine gas circuit. During the operation of the compressor, particles will enter the gas turbine with the air and gradually form a layer of fouling on the blades. When the compressor inlet bearing seal fails, the entry of lubricating oil mist will accelerate the formation of fouling. At the same time, during the operation of the entire gas turbine system, tiny substances that fall off from its system and stick to the blades will also form fouling, such as rust or scale falling from the pipeline, cooling water in the cooling system, and lubricating oil mist in the bearing lubrication system.

[0100] On the one hand, blade fouling will cause the blade surface to become rough, increasing flow losses; on the other hand, fouling will change the blade profile, which will lead to changes in the airflow angle and blade throat area. In severe cases, it may even cause gas turbine surge. Fouling will cause the compressor flow and efficiency to decrease, and the gas consumption will increase while keeping the load power unchanged.

[0101] When the compressor blades are fouled, the qualitative signs of related parameters mainly include: decreased compressor efficiency, decreased compressor flow, decreased compressor pressure ratio, decreased gas turbine power, decreased gas turbine efficiency, increased compressor outlet temperature, increased compressor outlet pressure, increased gas turbine exhaust temperature, increased gas turbine exhaust pressure, increased turbine exhaust temperature dispersion, etc.; the relevant quantitative indicator criteria include: a 2% decrease in compressor efficiency and a 7% decrease in equivalent flow;

[0102] (2) Compressor blade wear

[0103] During the operation of the gas turbine, the air flow will inevitably contain solid particles, which will cause scouring and cutting of the components and cause wear. Especially when the air filter is damaged, large particles will enter the gas turbine and continuously scour the blades, increasing the roughness of the blades. When there is more moisture and salt in the air, it will aggravate the corrosion and wear of the blades.

[0104] When the compressor blades are worn, the flow area of ​​the compressor increases, and the related flow rate increases accordingly. On the other hand, the wear of the blades changes the aerodynamic performance of the blades, which is manifested as an increase in the drag coefficient and a decrease in the flow rate. The increase in the drag coefficient reduces the pressure ratio of the low-pressure compressor and the compressor efficiency. If the power remains unchanged, the air consumption will increase. The reduction in the pressure ratio may further cause the compressor to surge. The research results show that the wear of the compressor blades has a greater impact on the later stages because the pressure ratio of the later stages is higher.

[0105] When compressor blades are fouled, the qualitative signs of related parameters mainly include: compressor efficiency decrease, gas consumption increase, compressor pressure ratio decrease, gas turbine power decrease, gas turbine efficiency decrease, etc.; the relevant quantitative indicator criteria include: efficiency decrease of 2%;

[0106] (3) Compressor surge

[0107] During the operation of the gas turbine, due to the incoordination between the airflow parameters and the blade geometric parameters, flow instability phenomena such as rotating stall and surge may occur. Once surge occurs, the unit will vibrate violently, and even cause blade breakage, component damage, etc., causing serious damage to the entire gas turbine and posing a great threat to the gas turbine and auxiliary equipment. When the air flow of the compressor is reduced to a certain extent, so that the operating condition point of the compressor crosses the surge boundary line, the compressor will enter an unstable working state. This is mainly because when the compressor operates under conditions deviating from the design conditions, a positive or negative angle of attack will be formed between the relative velocity parameters of the airflow and the geometric parameters of the blades. In particular, when the compressor generates a positive angle of attack due to the air flow being less than the design value, under the action of inertia, the airflow flowing on the back arc side of the blade has a tendency to separate from the back arc surface of the blade. When the air flow of the compressor continues to decrease to a certain extent, the compressor will enter an unstable working state. When the speed reaches a certain level, airflow separation will occur on the back arc surface of the blade, forming an unstable vortex boundary layer and developing into a local stall. As the compressor rotor continues to rotate, this local stall phenomenon will develop into several large local stall areas and rotate with the compressor rotor at a speed lower than the rotational speed, that is, rotational stall, and eventually develop and expand to the entire compressor channel, causing channel blockage. At this time, the front airflow is blocked and the air flow decreases until the outlet back pressure is low, then the blockage is released, the blocked airflow will rush down, the air flow increases, the outlet back pressure rises sharply, the compressor channel is blocked again, the air flow decreases again, until the outlet back pressure is low, then the blockage is released again, the blocked airflow rushes down again, the airflow is alternately blocked and unblocked, causing axial oscillation of the airflow, that is, surge occurs. It can be seen that when the air flow of the compressor is less than the design value to produce a positive angle of attack, surge failure may occur;

[0108] When the compressor surges, the qualitative signs of related parameters mainly include: engine load drop, air flow fluctuation, large fluctuation of compressor inlet pressure, large fluctuation of compressor exhaust pressure, low-frequency component of vibration value, compressor exhaust temperature fluctuation, compressor inlet temperature fluctuation, etc.

[0109] (4) Combustion chamber failure

[0110] The core of the combustion chamber is the flame tube. The working environment of the flame tube is particularly harsh. It is usually in high temperature, high pressure and violent burning flames. It is subject to high mechanical and thermal loads and is prone to failure. For example, the most common failures are: deformation, distortion and crack damage. According to statistics, the probability of failure of the flame tube components in the combustion chamber is one-third.

[0111] When the flame tube fails, the qualitative signs of related parameters mainly include: abnormal combustion pulsation, increased dispersion of turbine exhaust temperature, decreased efficiency of the engine, decreased efficiency of the combustion chamber, etc.; the relevant quantitative indicator criteria include: the combustion chamber efficiency decreases by 5%;

[0112] (5) Turbine blade scaling

[0113] The reasons for turbine blade fouling are similar to those for compressor blade fouling. In addition, due to incomplete combustion, the residual ash and other particles produced will adhere to the blade surface. The continuous accumulation of surface deposits will lead to fouling. As the unit runs longer, this fouling phenomenon becomes more and more serious, affecting the performance of the gas turbine. After the turbine blades are fouled, the flow resistance becomes larger, the flow area becomes smaller, and the efficiency is greatly reduced.

[0114] When turbine blades are fouled, the qualitative signs of related parameters mainly include: decreased turbine efficiency, decreased turbine flow, decreased engine efficiency, decreased engine power, increased turbine wheel temperature, increased turbine exhaust temperature dispersion, etc.; the relevant quantitative indicators include: a 6% decrease in turbine equivalent flow and a 2% decrease in turbine expansion efficiency;

[0115] (6) The corrosion and wear of gas turbine blades are mainly caused by solid particles entering the turbine. When the turbine is working, the solid particles will continuously impact the blades, causing the blades to wear and reduce the air flow resistance of the channel. The wear and corrosion of the turbine blades will reduce the working performance of the turbine, and also reduce its life and safety reliability;

[0116] When turbine blades are worn, the qualitative signs of related parameters mainly include: turbine efficiency decrease, turbine flow increase, gas turbine power decrease, gas turbine efficiency decrease, turbine temperature increase, gas turbine exhaust pressure increase, turbine exhaust temperature dispersion increase, etc.; the relevant quantitative indicators include: turbine equivalent flow increase of 6%, turbine expansion efficiency decrease of 2%;

[0117] S2. Establish a modular simulation model of the gas turbine and perform energy efficiency diagnosis and optimization operations on the simulation model based on the logic rule database. The specific operation process includes:

[0118] S21. Figure 3 and Figure 4 As shown, for the key components of the gas generator set such as the compressor, combustion chamber, gas turbine, etc., their mathematical models are established respectively through simulation codes, and the mathematical models are modified based on the characteristics of the components;

[0119] The mathematical models include:

[0120] (1) Compressor module: After the air in the environment enters the compressor, it is compressed into high-temperature and high-pressure working gas by the high-speed rotating blades. A compressor generally operates according to its characteristic line under different working conditions. The characteristic line of the compressor mainly describes the change relationship between its reduced speed Nc, reduced flow Gc, pressure ratio πc, and efficiency ηc. As long as any two of the parameters are known, the values ​​of the remaining two parameters can be known;

[0121] At this point, it should be noted that, in the F-class gas engine, a large proportion of air will be extracted from different stages of the compressor to cool the turbine blades, but due to the lack of actual measurement data of the cooling air of the compressor part, it is modeled in the form of equivalent cooling air in this application;

[0122] (2) Combustion chamber module: The combustion chamber module does not consider the influence of volume inertia in steady state and satisfies the mass conservation equation. The conservation here refers not only to the overall mass conservation between air, fuel and gas, but also to the mass conservation of each component before and after these three working fluids. In addition, the combustion chamber module satisfies the energy conservation equation, and the combustion efficiency of the combustion chamber is also considered in this equation. At the same time, considering that there will be resistance to fuel combustion and gas flow in the combustion chamber, there will be a certain pressure loss at the inlet and outlet of the combustion chamber, which is also crucial for the calculation of the combustion chamber.

[0123] (3) Turbine module: The high-temperature and high-pressure gas at the outlet of the combustion chamber expands and works in the turbine, converting heat energy into mechanical energy. Part of the work output by the gas turbine turbine is used to drive the compressor, and part is used to drive the external load to rotate. The turbine is similar to the compressor and also operates according to the characteristic line under different working conditions. The characteristics of the turbine are also described by the changing relationship between the four parameters of reduced speed nt, reduced flow Gt, expansion ratio πt, and efficiency ηt. Only by knowing any two of the parameters, the values ​​of the remaining two parameters can be known. Therefore, the modeling method of the turbine is similar to that of the compressor.

[0124] It should also be noted that accurate component characteristics are the prerequisite for ensuring diagnostic accuracy. Since the compressor module and turbine module are established in the form of characteristic lines, and the characteristic lines are difficult to obtain directly from the manufacturer, the component characteristics are generally calculated by the following methods: primitive blade cascade method, step-by-step superposition method, and characteristic diagram correction method. Taking the compressor as an example, the primitive blade cascade method calculates the characteristics step by step through the geometric shape of each blade cascade, the step-by-step superposition method calculates the overall performance of the compressor step by step through the first-level general characteristic curve, and the characteristic diagram correction method takes into account that when the compressor blade shape and various performance index parameters are similar, its characteristic line also has similar trends and curve shapes;

[0125] Through the analysis of the above three methods, the present application adopts the characteristic diagram correction method. Taking the compressor as an example, based on the characteristic curve of the reference compressor, the reference characteristic line is corrected by scaling, interpolating and translating the operating data of the compressor of the actual unit, so that it can match the characteristics of the actual operating compressor. Taking the compressor as an example, the present application uses the universal characteristic line for translation, interpolation and scaling to obtain the characteristic curve of the actual unit.

[0126] S23. Analyze the impact of air intake system degradation on the thermal performance of the gas turbine, and optimize the performance of the key components of the gas turbine and the simulation model based on the actual operating data of the gas turbine;

[0127] In a specific embodiment, if Figure 5 As shown, it shows the influence of three performance parameters, namely, gas turbine output power, compressor outlet pressure, and turbine outlet temperature, on the changes of three types of pressure difference influencing factors. Among them, relative humidity and atmospheric pollutant concentration are combined into four fixed environmental types, and the filter degradation state is selected as a continuous variable. The performance parameters of the gas turbine under the filter pressure difference corresponding to low humidity and clean air (RH<60%, DC=0) conditions and healthy filter (DI=0) are selected as the benchmark performance parameters. During the degradation of the filter pressure difference, the fuel flow rate and the IGV opening are controlled to remain unchanged. The calculated changes of the thermal performance of the gas turbine with the degradation of the filter pressure difference show that when the filter is close to its service life, the gas turbine performance is more susceptible to the influence of adverse environmental factors.

[0128] Finally, the degradation diagnosis model is fed with the operating data of the gas turbine unit in a certain period of time, and the compressor performance degradation changes are calculated based on the operating data;

[0129] Due to the constant change of operating conditions and the existence of measurement errors, the diagnostic results of compressor efficiency degradation and flow degradation fluctuate greatly. It is difficult to use the original data for subsequent neural network model training and trend prediction. Therefore, it is also necessary to perform noise reduction on the above two flow and efficiency diagnostic results. The noise reduction method used is the Gaussian weighted moving average method.

[0130] S3. Analyze the failure impact mechanism of the optimized simulation model, and evaluate the life of the core components of the gas turbine based on the failure impact mechanism of the simulation model;

[0131] The specific operation process of analyzing the fault impact mechanism of the optimized simulation model includes:

[0132] S31. A numerical simulation model is established for NOx emissions during the startup process of a Class F heavy-duty gas engine. This model is used to study the effect of fuel distribution on NOx emission concentration under PPM mode. Figure 6As shown in the figure, with the increase of D5 fuel allocation ratio, NOx emission shows an upward trend. When D5 increases to about 45%, the increase of NOx emission increases slightly. The effect of PM1 on emission concentration is similar to that of D5. With the increase of PM1 fuel allocation ratio, NOx emission increases, and the increase rate increases slightly. The effect of PM1 allocation ratio on NOx concentration is greater than that of D5.

[0133] S32. Analyze the influence of fuel distribution mode on NOx emission concentration under PPM mode based on numerical simulation model and obtain the influence mechanism;

[0134] like Figure 7 As shown in the figure, since the NOx generation mechanism in the gas turbine combustion chamber is thermal NOx, the NOx concentration and temperature in the combustion chamber show a positive correlation trend. In both combustion modes, the NOx generation concentration peak is located in the PM1 flame or near the PM1 flame. Therefore, the PM1 flame combustion intensity directly affects the NOx emission level.

[0135] S33. Based on the numerical simulation model, the dynamic pressure changes at specific positions of the combustion chamber under high load conditions, low load conditions, and high load conditions with the same nozzle flow rate in the PPM mode are optimized to optimize the influencing mechanism;

[0136] In summary, the 9FA gas turbine uses the DLN2.0+ combustion system, and the combustion control strategy is complex. To develop the combustion adjustment technology for this gas turbine, it is necessary to have a very detailed understanding of the switching load, switching point combustion temperature, and fuel distribution law of each combustion mode. Too low PM1 fuel flow will have an adverse effect on the dynamic characteristics of the combustion chamber.

[0137] In addition, the specific operational procedures for life assessment of core components of gas turbines include:

[0138] L1. Using the standard parameters of the gas turbine (type, reference point performance) and the actual operating condition data, the entire turbine gas path section is thermodynamically analyzed after data filtering to calculate some missing information, quantify the heat load and load of the core components of the gas turbine, and create boundary conditions for the CFD analysis of the core components;

[0139] At this point, once the design calculations for the base load have been performed, appropriate benchmark parameters for the various engine parts (such as the compressor, combustion chamber, and turbine parts) can be obtained. Using the component geometries provided by all units of the engine, multi-stage CFD analysis can be performed using simplified moving blades, guide vanes, and flow path geometries to obtain component boundary conditions, including the distribution of values ​​for gas path temperature, pressure, and mass flow, so as to perform CFD analysis on the components;

[0140] L2. Perform CFD analysis on the core components of the gas turbine. Based on the multi-level CFD calculation results, combined with the geometric files and material parameters (thermophysical parameters, mechanical properties, and microstructure) of the core components, perform thermal fluid-solid coupling analysis under different working conditions, realize load processing and data conversion of temperature loads and aerodynamic loads, obtain accurate cooling characteristics and strength characteristics, and then obtain the stress and strain distribution of the blades through finite element analysis;

[0141] L3. By considering the creep, high cycle fatigue and low cycle fatigue loads of core components under basic load conditions, as well as the coupling effects of loads, a life analysis algorithm model of creep, high cycle fatigue and low cycle fatigue loads of core components is constructed based on the influence mechanism, and the life loss of core components is analyzed to obtain its estimated remaining life and reliability status, so as to determine the failure mode, potential failure location and component life of the components;

[0142] L4. Establish a probabilistic analysis model based on the estimated remaining life and reliability status, conduct probabilistic analysis on the parameter variability and uncertainty during the operation of the gas turbine and their impact, and obtain deterministic calculation results;

[0143] L5. Based on the deterministic calculation results in the above steps, the physical field is obtained as the nominal field for probability analysis, and parametric modeling, Monte Carlo-based Latin hypercube sampling, and variance reduction are performed in sequence for probability simulation;

[0144] L6. Based on the crack length, depth, fatigue and creep crack growth rate, and crack instability length of the component, damage tolerance analysis based on structural integrity and probabilistic simulation results, the remaining life of the component is further obtained and the optimal maintenance interval is formulated;

[0145] In a specific embodiment, based on the three-dimensional steady-state viscous Navier-Stokes equation, considering the coupling of aerodynamic and thermal physical fields, a thermal fluid-solid coupling analysis model, a deterministic life and a probabilistic life analysis model of a high-pressure turbine blade of a gas turbine with a complete internal and external cooling structure are established by using conjugate calculation technology and SST turbulence model, the blade temperature distribution and the corresponding strain stress field are obtained, and the probabilistic model is improved by using the deterministic analysis results as the nominal field, the probabilistic life of the blade is obtained, and the sensitivity analysis of the factors affecting the blade life is carried out. The research results have important reference significance for understanding and mastering advanced gas turbine turbine life assessment technology. The specific research methods include:

[0146] (1) High-fidelity thermal-fluid-structure coupling analysis technology for turbine blades with complex cooling structures

[0147] For turbine blades with complex cooling structures, a gas-thermal coupling analysis model is constructed that takes into account the flow in the fluid domain, heat transfer at the fluid-solid interface, and heat conduction in the solid domain. A thermal-solid coupling characteristic analysis model that takes into account temperature, aerodynamic, and centrifugal loads is constructed. A high-fidelity thermal-fluid-solid coupling analysis model is jointly formed to realize load processing and data conversion of temperature loads and aerodynamic loads, and obtain accurate cooling characteristics and strength characteristics.

[0148] (2) Data-driven high-fidelity integrated intelligent prediction technology for turbine blade aerodynamic strength

[0149] For turbine blades operating at high temperatures, we developed a data-driven, high-fidelity, integrated intelligent prediction technology for aerodynamic strength. We built a data-driven prediction model through a deep convolutional neural network, coupled the aerodynamic and strength analysis processes, and predicted the operating status of turbine blades under different operating conditions.

[0150] (3) High-precision fatigue life analysis algorithm for creep-high-cycle-low-cycle coupling

[0151] Based on the fatigue life analysis model of blade material creep-high cycle-low cycle coupling, the creep, high cycle fatigue and low cycle fatigue loads in turbine blade operation, as well as the coupling effects of these loads, are fully considered to develop a high-precision fatigue life analysis algorithm. By analyzing the life loss of the blade, its remaining life and reliability status are obtained to achieve efficient health management;

[0152] (4) Probabilistic analysis model establishment and sensitivity analysis

[0153] The probabilistic life prediction process emphasizes the impact of random variation and uncertainty of data. The probability analysis method no longer uses fixed values ​​of variables, but probability distribution instead. The analysis of parameter variability and uncertainty and their impact during the operation of the gas turbine requires corresponding probability models and probability analysis. The physical field obtained by deterministic calculation is used as the nominal field of probability analysis. Parametric modeling technology, Latin hypercube sampling based on Monte Carlo method, variance reduction technology and thermal fluid-solid coupling analysis method are used for probability simulation. The numerical characteristics of the discrete distribution of input parameter samples are very close to the indicators of continuous life distribution, which can reflect the random characteristics of the probability model and have the rationality and availability of probability analysis.

[0154] In summary, by introducing probability analysis and thermal-fluid-solid coupling methods, the random variation and uncertainty of parameters are considered in the life prediction process, and a probability model that is closer to the actual situation can be constructed, so as to accurately evaluate the remaining life and reliability of core components, and ultimately facilitate the scientific formulation of preventive maintenance strategies and reduce the risk of unexpected downtime;

[0155] S4. Perform an overall health assessment of the gas turbine based on the life assessment results of each core component. The specific steps include:

[0156] S41. Establish a health assessment system and preset assessment parameters, such as Figure 8 As shown, a health function is constructed, and the number of features in the preset function is a, b, c, d, e, f;

[0157] It should be noted that the criteria for selecting evaluation parameters are:

[0158] (1) Scientificity: The selection of evaluation parameters must be based on science, requiring the parameters to reflect the operating status of the equipment to a certain extent and conform to certain objective facts. If the selection is unscientific, the accuracy of the evaluation results will be reduced, affecting the correct decision-making of the entire equipment management, resulting in unsafe operation and even serious safety accidents;

[0159] (2) Independence: The evaluation parameters reflect the level of equipment health status from a certain aspect. The correlation between parameters is required to be small and have a certain degree of independence to avoid overlap or inclusion, which affects the accuracy of the evaluation.

[0160] (3) Feasibility: Feasibility includes both theoretical and technical aspects. On the one hand, theoretically, the evaluation parameters must be statistically consistent, clearly defined, and achievable with today’s technological level. On the other hand, technical feasibility requires that data be easy to collect, simple to operate or obtainable through relatively simple calculations, and that all evaluation parameters are standardized.

[0161] (4) Consistency: This is proposed for the relationship between various parameters, that is, there should be no contradiction between the various parameters;

[0162] (5) Comprehensiveness: All major parameters that affect the health status of the equipment should be included in the parameter set, and the parameter set should meet the principle of comprehensiveness. The selected evaluation parameters should be able to fully and clearly reflect all possibilities of the system's operating status and more comprehensively reflect the state change trend of the system;

[0163] S42. Classify and statistically analyze the monitoring values ​​of the evaluation parameters of the gas turbine generator set in normal working state and faulty working state, and select appropriate values ​​as the values ​​of the health function characteristic numbers a, b, c, d, e, and f based on the statistical results;

[0164] At this time based on Figure 8 It should be noted that for the smaller the better type of parameters, we only need to consider the distribution of the function on the right side of the dividing line; for the too low bad type of parameters, we only need to consider the distribution of the function on the left side of the dividing line; and for the intermediate type of parameters, we need to consider the overall distribution of the function on both sides of the dividing line. The key to using this health function is how to determine the characteristic numbers a, b, c, d, e, and f in the function.

[0165] By classifying and unifying the monitoring values ​​of the evaluation parameters of the gas turbine generator set in normal working state and fault working state, and then using statistical thinking to analyze, appropriate values ​​are selected as the values ​​of the health function characteristic numbers a, b, c, d, e, and f by appropriate calculation based on the statistical results;

[0166] In the statistical process, probability density statistics are mainly completed based on the monitoring data of known parameters;

[0167] S43. Based on the parameter evaluation, the membership function of each system and each evaluation parameter is established to obtain the membership status;

[0168] S44. The evaluation results of the gas turbine generator system are calculated by using the variable weight analytic hierarchy process combined with fuzzy comprehensive evaluation theory and other methods, and then the health of the system is obtained by performing membership post-processing, which specifically includes:

[0169] Establish a factor set consisting of n influencing factors of the evaluation object:

[0170] U={u 1 ,u 2 ,u 3 ,…,u n}

[0171] Create a review set consisting of m evaluation results as the evaluation level set:

[0172] V={v 1 , v 2 , v 3 , …, v m}

[0173] The comment set is a combination of a series of qualitative evaluations, such as excellent, good, fair, poor, and extremely poor. The number of elements in the comment set should be appropriate. Too many elements will make it unclear to distinguish between the levels, and too few elements will make the evaluation unclear. Generally, there are 4 to 5 elements in the comment set.

[0174] The judgment matrix R representing the fuzzy relationship between the factor set and the judgment set is obtained as follows:

[0175]

[0176] Among them, r ij (i∈[1,n],j∈[1,m]) represents the sum of the factors u i To see things being rated as level v j degree of affiliation;

[0177] The fuzzy set A on U is used to represent the weight distribution of each factor. The weight distribution of each factor here is realized by using the hierarchical analysis method, which is recorded as:

[0178] A={a 1 , a 2 , a 3 , …, a n}

[0179] Where a i (i∈[1,n]) represents the i-th factor u i The weight of , then the fuzzy comprehensive evaluation membership result is expressed as:

[0180]

[0181] Among them, ° is the fuzzy synthesis operator selected according to the actual situation of the problem;

[0182] For multi-level problems, we start from the lowest parameter layer and recursively calculate layer by layer to finally get the evaluation result of the target layer.

[0183] The evaluation results are processed in the membership function to obtain the health of the system;

[0184] In summary, the real-time health evaluation model based on multi-source information fusion, through in-depth mining of the historical operation data of the equipment, explores the mapping relationship between parameter status and health degree under different working conditions and the weight value of the influence of different measurement point parameters on the equipment status, and finally can comprehensively analyze the multi-dimensional information such as real-time detection parameters, performance indicators, safety thresholds, etc. of the equipment to form a comprehensive health evaluation index of the real-time status of the equipment;

[0185] Therefore, through multi-source information fusion technology and in-depth data analysis, the health assessment system is further established and the health degree is calculated using the life assessment results of core components. This can effectively integrate historical and real-time data, and accurately map the relationship between the health status of equipment and operating parameters, thereby achieving a comprehensive and dynamic assessment of the health level of gas-fired generator sets, thereby playing a role in efficient and accurate gas turbine fault diagnosis, and ultimately ensuring the long-term stable operation and efficient management of gas turbine equipment.

[0186] On the other hand, the present invention further discloses a computer-readable storage medium storing a computer program, wherein when the computer program is executed by a processor, the processor executes the steps of the above method.

[0187] On the other hand, the present invention further discloses a computer device, including a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the above method.

[0188] In another embodiment provided in the present application, a computer program product comprising instructions is also provided, which, when executed on a computer, enables the computer to execute the cloud-edge collaborative fault diagnosis method for any gas generator set in the above-mentioned embodiments.

[0189] It is understandable that the system provided by the embodiment of the present invention corresponds to the method provided by the embodiment of the present invention, and the explanation, examples and beneficial effects of the relevant contents can refer to the corresponding parts in the above method.

[0190] The embodiment of the present application also provides an electronic device, including a processor, a communication interface, a memory and a communication bus, wherein the processor, the communication interface and the memory communicate with each other through the communication bus.

[0191] Memory, used to store computer programs;

[0192] The processor is used to implement the cloud-edge collaborative fault diagnosis method for the above-mentioned gas generator set when executing the program stored in the memory.

[0193] The communication bus mentioned in the above electronic device can be a Peripheral Component Interconnect (PCI) bus or an Extended Industry Standard Architecture (EISA) bus, etc. The communication bus can be divided into an address bus, a data bus, a control bus, etc.

[0194] The communication interface is used for communication between the above electronic device and other devices.

[0195] The memory may include a random access memory (RAM) or a non-volatile memory (NVM), such as at least one disk memory. Optionally, the memory may also be at least one storage device located away from the aforementioned processor.

[0196] The above-mentioned processor can be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), etc.; it can also be a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, and discrete hardware components.

[0197] In the above embodiments, all or part of them may be implemented by software, hardware, firmware or any combination thereof.

[0198] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

[0199] In addition, it should be noted that if the embodiments of the present invention involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative position relationship, movement status, etc. between the components in a certain specific posture. If the specific posture changes, the directional indication will also change accordingly.

[0200] In addition, if there are descriptions involving "first", "second", etc. in the embodiments of the present invention, the descriptions of "first", "second", etc. are only used for descriptive purposes and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the meaning of "and / or" appearing in the full text includes three parallel schemes. Taking "A and / or B" as an example, it includes scheme A, or scheme B, or schemes that A and B meet at the same time. In addition, in the embodiments of the present invention, "multiple" refers to more than two. In addition, the technical solutions between the various embodiments can be combined with each other, but it must be based on the ability of ordinary technicians in the field to implement. When the combination of technical solutions is contradictory or cannot be implemented, it should be considered that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

Claims

1. A cloud-edge collaborative fault diagnosis method for a gas generator set, characterized in that: Follow these steps on your computer device: S1. Obtain the fault criteria and symptoms of gas turbines based on domain knowledge, and perform fault diagnosis and analysis on gas turbines based on technology fusion based on multiple criteria and symptoms, and establish a logical rule database for different faults; S2. Establish a modular simulation model of gas turbines, and perform energy efficiency diagnosis and optimization operations on the simulation model based on a logical rule database; S3. Analyze the failure impact mechanism of the optimized simulation model, and evaluate the life of the core components of the gas turbine based on the failure impact mechanism of the simulation model; S4. Conduct an overall health assessment of the gas turbine based on the life assessment results of each core component.

2. The cloud-edge collaborative fault diagnosis method for a gas generator set according to claim 1, characterized in that: In the step S1, the specific operation process of performing fault diagnosis and analysis on the gas turbine based on technology fusion according to multiple criteria and symptoms and establishing a logic rule database for different faults includes: S11. Analyze the signs and criteria of typical failure modes in gas generator sets in combination with expert domain knowledge, and comprehensively analyze the results to form a standard fault diagnosis information library; S12. Based on the analysis results of the faults by domain experts, various fault judgment modes and fault symptom criteria are obtained as fault knowledge, and the quantitative indicator criteria and qualitative symptom criteria of each fault are integrated to realize multi-dimensional intelligent diagnosis of each fault; S13. Based on the multi-dimensional intelligent diagnosis results of each fault, a preliminary model of logic rules for different faults is established, and the logic rules are centrally managed to establish a logic rule database for different faults.

3. The cloud-edge collaborative fault diagnosis method for a gas generator set according to claim 1, characterized in that: The specific operation process of analyzing the signs and criteria of typical failure modes in the gas generator set in step S11 includes: S111. Summarize the actual faults and classify them; S112. Determine the degree of membership between each fault symptom and the fault cause based on expert experience or statistical methods, and establish a fault membership table; S113. Establish a fuzzy relationship matrix based on the membership table and calculate the fuzzy relationship to draw a conclusion as the fault diagnosis result.

4. The cloud-edge collaborative fault diagnosis method for a gas generator set according to claim 2, characterized in that: The specific operation process of integrating the quantitative index criteria and the qualitative symptom criteria of each fault in step S12 includes: S121. Determine the main fault warning range through the preliminary parameters of the gas generator set, disassemble and analyze the gas generator set equipment to build an intelligent warning model; S122 develops diagnostic model monitoring parameters based on the analysis of qualitative signs for typical fault types, builds a parameter prediction model based on the intelligent early warning model, and compares the prediction results with the actual operation results to obtain changes in parameters such as increase, decrease, and fluctuation, and converts the monitoring parameters into fault signs; S123. According to the quantitative indicators of typical faults, based on the characteristic simulation of gas turbine equipment and the gas path performance degradation analysis, a gas turbine real-time performance simulation model is established to monitor the degradation of key indicators in real time; S124. Establish a logical reasoning model based on the fault diagnosis information database and integrate the above qualitative and quantitative signs.

5. The cloud-edge collaborative fault diagnosis method for a gas generator set according to claim 1, characterized in that: The specific operation process of the S2 step includes: S21. For key components of gas generator sets, mathematical models are established respectively, and the mathematical models are modified based on the characteristics of the components; S23. Analyze the impact of air intake system degradation on the thermal performance of the gas turbine, and optimize the performance of the key components of the gas turbine and the simulation model simultaneously based on the actual operating data of the gas turbine.

6. The cloud-edge collaborative fault diagnosis method for a gas generator set according to claim 1, characterized in that: The specific operation process of analyzing the fault impact mechanism of the optimized simulation model in step S3 includes: S31. Establish a numerical simulation model for NOx emissions during the startup process of F-class heavy-duty gas turbines; S32. Analyze the influence of fuel distribution mode on NOx emission concentration under PPM mode based on numerical simulation model and obtain the influence mechanism; S33. Based on the numerical simulation model, the dynamic pressure changes at specific positions of the combustion chamber under high load conditions, low load conditions, and high load conditions with the same nozzle flow in PPM mode are analyzed to optimize the influencing mechanism.

7. The cloud-edge collaborative fault diagnosis method for a gas generator set according to claim 1, characterized in that: The specific operation process of performing the life assessment of the core components of the gas turbine in step S3 includes: L1. Using the standard parameters of the gas turbine and the actual operating condition data, after data filtering, the turbine gas path section of the entire gas turbine is thermodynamically analyzed to calculate some missing information, quantify the heat load and load of the core components of the gas turbine, the appropriate benchmark parameters of the core components of the gas turbine, and create boundary conditions for the CFD analysis of the core components. The boundary conditions include the distribution of the values ​​of gas path temperature, pressure and mass flow; L2. Perform CFD analysis on the core components of the gas turbine. Based on the multi-level CFD calculation results, combined with the geometric files and material parameters of the core components, conduct thermal fluid-solid coupling analysis under different working conditions, realize load processing and data conversion of temperature load and aerodynamic load, obtain accurate cooling characteristics and strength characteristics, and then obtain the stress and strain distribution of the blades through finite element analysis; L3. By considering the creep, high cycle fatigue and low cycle fatigue loads of core components under basic load conditions, as well as the coupling effects of loads, a life analysis algorithm model of creep, high cycle fatigue and low cycle fatigue loads of core components is constructed based on the influence mechanism, and the life loss of core components is analyzed to obtain its estimated remaining life and reliability status, so as to determine the failure mode, potential failure location and component life of the components; L4. Establish a probabilistic analysis model based on the estimated remaining life and reliability status, conduct probabilistic analysis on the parameter variability and uncertainty during the operation of the gas turbine and their impact, and obtain deterministic calculation results; L5. Based on the deterministic calculation results in the above steps, the physical field is obtained as the nominal field for probability analysis, and parametric modeling, Monte Carlo-based Latin hypercube sampling, and variance reduction are performed in sequence for probability simulation; L6. Based on the crack length, depth, fatigue and creep crack growth rate, and crack instability length of the component, the remaining life of the component is further determined based on the damage tolerance analysis and probabilistic simulation results of the structural integrity, and the optimal maintenance interval is formulated.

8. The cloud-edge collaborative fault diagnosis method for a gas generator set according to claim 1, characterized in that: The specific operation steps of step S4 include: S41. Establish a health assessment system, preset assessment parameters, and construct a health function, where the characteristic numbers in the preset function are a, b, c, d, e, and f; S42. Classify and statistically analyze the monitoring values ​​of the evaluation parameters of the gas turbine generator set in normal working state and faulty working state, and select appropriate values ​​as the values ​​of the health function characteristic numbers a, b, c, d, e, and f based on the statistical results; S43. Establish the membership function of each system and each evaluation parameter to obtain the membership status; S44. Establish a factor set consisting of n influencing factors of the evaluation object: U={u1,u2,u3,…,u n } Create a review set consisting of m evaluation results as the evaluation level set: V={v1,v2,v3,…,v m } The comment set is a combination of a series of qualitative evaluations, such as excellent, good, fair, poor, and extremely poor. The number of elements in the comment set should be appropriate. Too many elements will make it unclear to distinguish between the levels, and too few elements will make the evaluation unclear. Generally, there are 4 to 5 elements in the comment set; S45. The evaluation matrix R representing the fuzzy relationship between the factor set and the evaluation set is obtained as follows: Among them, r ij (i∈[1,n],j∈[1,m]) represents the sum of the factors u i To see things being rated as level v j degree of affiliation; S46. The fuzzy set A on U is used to represent the weight distribution of each factor. The weight distribution of each factor here is realized by using the hierarchical analysis method, which is recorded as: <h2 style=";text-align:left;direction:ltr">A = {a1, a2, a3,…, a<h2 style=";text-align:left;direction:ltr"> n <h2 style=";text-align:left;direction:ltr">} Where a i (i∈[1,n]) represents the i-th factor u i The weight of , then the fuzzy comprehensive evaluation membership result is expressed as: in, The fuzzy synthesis operator is selected according to the actual situation of the problem; S47. For multi-level problems, the calculation starts from the lowest parameter layer, and recursively proceeds layer by layer to finally obtain the evaluation result of the target layer; S48. The evaluation results are processed in the membership function to obtain the health of the system.

9. A computer-readable storage medium, characterized in that: A computer program is stored, and when the computer program is executed by a processor, the processor is caused to perform the steps of the method according to any one of claims 1 to 8.

10. A computer device, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and when the computer program is executed by the processor, the processor executes the steps of the method according to any one of claims 1 to 8.

Images