CN108827643A - A kind of high-temperature component of gas turbine fault early warning method for considering to arrange warm temperature field rotation - Google Patents

Abstract

A fault early warning method for high temperature components of a gas turbine considering the rotation of exhaust temperature field, the invention relates to a fault early warning method for high temperature components of a gas turbine. The present invention solves the problems of after-the-fact diagnosis in the existing method, when the system alarms, the components have already been seriously damaged and the accuracy of fault early warning is low. Compared with the known real-time monitoring method of the same kind at present, the method of the invention makes full use of the data information of each measuring point of exhaust temperature. Considering the influence of different working conditions on the exhaust temperature, the interference of the high-temperature gas rotation of the gas turbine on the outlet temperature field distribution can be eliminated, and the abnormal evolution process can be detected more accurately, and the fault can be detected in the early stage of the fault, improving the accuracy of fault warning . When the exhaust temperature decreases and deteriorates with a slope of 0.008, the method of the invention detects the fault 50 minutes earlier than the traditional method. The invention is used in the field of gas turbine failure prediction.

Description

A fault early warning method for gas turbine high temperature components considering the rotation of exhaust temperature field

technical field

The invention relates to the field of failure prediction of gas turbines, and relates to a method for early warning of failures of high temperature components of gas turbines in consideration of the rotation of exhaust temperature field.

Background technique

As a new type of power equipment, gas turbine has the advantages of compact structure, stable operation, safety and reliability, quick start and load driving, and high thermal efficiency. It has been widely used in aviation, ground and ships. Therefore, the gas turbine Anomaly detection is of great significance to production practice. During the operation of the gas turbine unit, more than 50% of the faults in the abnormal detection of the gas turbine combustion system are related to the combustion chamber. Since the combustor, combustor and other components work in the high temperature area of 1600 ° C for a long time, the working environment is harsh, and once the equipment is defective, it may pose a threat to the safety of downstream nozzles and moving blade components. Therefore, it is necessary to monitor the working condition of the combustion chamber.

Once the combustion system fails, it will cause a difference in the outlet temperature of the combustion chamber. Therefore, we can monitor the operation status of the combustion system by detecting the outlet temperature of the combustion chamber. However, conventional temperature measuring elements cannot work in such a high temperature area for a long time. Therefore, several exhaust temperature measuring thermocouples are evenly arranged circumferentially in the turbine exhaust passage of the unit. The temperature measured by the thermocouples is the exhaust temperature of the gas turbine. temperature. The user indirectly monitors the working conditions in the combustor by monitoring the exhaust temperature of the gas turbine. In actual operation, when the combustion tube is abnormal, the result of exhaust temperature will also be abnormal, so it can be judged whether the working condition of the combustion tube is abnormal according to the exhaust temperature. The distribution of thermocouples is shown in Figure 1.

The MARK VI combustion monitoring system developed by GE defines S as the allowable dispersion of exhaust gas temperature, and considers that S is a function of the average exhaust temperature T ₄ ^* at the outlet of the gas turbine and the temperature T ₄ ^* at the outlet of the compressor. The specific function is an experience formula:

In this formula, temperature is measured in °F. The 100 at the right end of the formula has brackets, which means that this item is only added under variable working conditions.

In addition, the MARK VI combustion monitoring system also defines: S1 is the difference between the highest reading and the lowest reading of the exhaust temperature thermocouple; S2 is the difference between the highest reading and the second lowest reading of the exhaust temperature thermocouple; S3 is the difference between the highest reading and the 3rd lowest reading of the discharge temperature thermocouple.

Based on the above formulas and definitions, the discrimination principle of MARK VI combustion monitoring and protection system is shown in Figure 2. In Fig. 2, K ₁ , K ₂ , and K ₃ are three parameters defined based on experience. Typically: K ₁ =1.0; K ₂ =5.0; K ₃ =0.8;

In practical application, it is found that this method cannot detect the abnormal evolution process, and cannot make judgments on the changing trend of the combustion state. There is a serious "post-event" diagnosis phenomenon, that is, the combustion system has been seriously damaged when the detection system sends out an alarm.

At the same time, the existing technologies all assume that the turbine flow path of the unit is uniform, but in the actual operation process, different working conditions will cause the high-temperature gas to rotate, as shown in Figure 3, so that the gas turbine outlet temperature field distribution of difference. The difference caused by the rotation is an interference factor for the difference in the outlet temperature of the combustion chamber when the combustion system fails. In the early warning, the influence caused by the rotation may be greater than the abnormality of the combustion cylinder, as shown in Figure 4. As a result, abnormalities in the combustion chamber cannot be detected. In addition, different working conditions will also lead to the scaling of the temperature of different measuring points. Therefore, it is necessary to consider the influence of different working conditions on the exhaust temperature and eliminate the influence of rotation on the gas turbine outlet temperature distribution.

Contents of the invention

The purpose of the present invention is to solve the problem of after-the-fact diagnosis in the existing method. When the system alarms, the components have been seriously damaged and the accuracy of fault early warning is low, and a gas turbine considering the rotation of the exhaust temperature field is proposed. High temperature component failure early warning method.

A fault early warning method for high temperature components of a gas turbine considering the rotation of the exhaust temperature field includes the following steps:

Step 1: The exhaust temperature of the gas turbine is measured by n thermocouples evenly distributed in the circumferential direction of the turbine outlet. During the normal operation time of the gas turbine [t ₁ , t _m ], the exhaust temperature data sequence of n measuring points at m times is obtained. Respectively constitute the exhaust temperature data series and T _4,1 , T _4,2 ,..., T _4,n , the exhaust temperature data sequence at time t ₁ As the reference sequence of the training set; the exhaust temperature data sequence at time t _m As the comparison sequence to be processed in the training set; T ₄ is the exhaust temperature;

said is the exhaust temperature data sequence at time t ₂ , T _4,1 is the exhaust temperature data sequence of the first measuring point, T _4,2 is the exhaust temperature data sequence of the first measuring point, T _4,n is the nth Exhaust temperature data series of measuring points;

Step 2: Define the comparison sequence to be processed in the training set Reference sequence with training set The deviation of e is e, and the minimum value of e is determined by the Euclidean distance discriminant method, and the exhaust temperature data sequence after rotation correction at m moments when e takes the minimum value is obtained and the rotation-corrected exhaust temperature data sequence T′ _4,1 , T′ _4,2 ,…, T′ _4,n of n measuring points;

Step 3: Determine the relationship between the rotation-corrected exhaust temperature data series T′ _4,1 , T′ _4,2 , ..., T′ _4,n of n measuring points and the average temperature of n measuring points;

Step 4: Exhaust temperature data sequence corrected by rotation at m moments determined in step 3 and the relationship between the average temperature of n measuring points to obtain the predicted temperature values T _4pre,1 ,T _4pre,2 ,...,T _4pre,i ,...,T _{4pre,n of n} measuring points;

Step 5: According to the temperature prediction values T _4pre,1 ,T _4pre,2 ,…,T _4pre,i ,…,T _4pre,n of n measuring points obtained in Step 4, get the error band ΔT ₁ of each measuring point ,ΔT ₂ ,…,ΔT _i ,…,ΔT _n ; where ΔT _i is the error band of the i-th measuring point;

Step 6: When monitoring, re-execute steps 1 to 5. If the error band of the i-th measuring point is within the range of ΔT _i , it means that the gas turbine unit is not faulty, otherwise the gas turbine unit is faulty.

During monitoring, the exhaust temperature data measured at each moment is substituted into the predicted value T _{4pre,1 of measuring point 1} . The difference between the predicted value T _4pre,1 of measuring point 1 and the actual value (corrected value) T ₄ ' _,1t of measuring point 1 when the unit is not faulty at each moment of monitoring is ΔT ₁ , which means that the unit is not faulty. If it exceeds this range , it means that the unit is malfunctioning.

The difference between the predicted value and the actual value ΔT ₂ , ΔT ₃ ,...,ΔT _n of the measuring points 2 to n when the unit has no faults is respectively obtained. If the deviation is within the range of ΔT _i , it means that the unit has no faults. If it exceeds this range, It means that the unit is malfunctioning.

The high-temperature components refer to the flame cylinder, the transition section, the first stage turbine nozzle, the moving blade and other components in the combustion chamber of the gas turbine.

The beneficial effects of the present invention are:

The current existing methods have the phenomenon of post-event diagnosis. When the system alarms, the components have been seriously damaged. In order to realize the early warning of the failure of the high temperature component of the gas turbine and find the failure early, the present invention provides an early warning method for the failure of the high temperature component of the gas turbine considering the rotation of the exhaust temperature field.

Compared with the known real-time monitoring methods of the same kind at present, the method of the invention can make full use of the data information of each measuring point of exhaust temperature. At the same time, considering the influence of different working conditions on the exhaust temperature, the interference of the high-temperature gas rotation of the gas turbine on the outlet temperature field distribution can be eliminated, and the abnormal evolution process can be detected more accurately, which can be detected in the early stage of the fault and improve the effectiveness of fault early warning. accuracy. Therefore, it can reduce the loss and possibility caused by the gas turbine failing to be discovered in time. When the exhaust temperature decreases and deteriorates with a slope of 0.008, the method of the invention detects the fault 50 minutes earlier than the traditional method.

Description of drawings

Figure 1 is a diagram of the distribution of thermocouples;

Fig. 2 is the discrimination schematic diagram of combustion monitoring;

Figure 3 is the phenomenon diagram of rotation under different working conditions; C ₀ in the figure is the absolute velocity of the airflow at the inlet of the stator blade (m/s), C ₁ is the absolute velocity of the airflow at the inlet of the moving blade (m/s), and C ₂ is The absolute velocity of the airflow at the outlet of the rotor blade (m/s), u is the implicated velocity of the rotor blade (m/s), α ₁ is the absolute velocity direction angle of C ₁ , β ₁ is the relative velocity direction angle, α ₂ is the absolute velocity direction angle of C ₂ Velocity direction angle, _β2 is the relative velocity direction angle;

Figure 4 is a diagram of exhaust temperature distribution before and after rotation;

Fig. 5 is a training set data diagram, in which the abscissa is time, and the ordinate is power;

Figure 6 is a normal result graph of the test set;

Figure 7 is a graph of the abnormal results of the test set.

Detailed ways

Embodiment 1: A method for early warning of failure of high temperature components of a gas turbine considering the rotation of the exhaust temperature field includes the following steps:

Step 1: The exhaust temperature of the gas turbine is measured by n thermocouples evenly distributed in the circumferential direction of the turbine outlet. During the normal operation time of the gas turbine [t ₁ , t _m ], the exhaust temperature data sequence of n measuring points at m times is obtained. Respectively constitute the exhaust temperature data series and T _4,1 , T _4,2 ,..., T _4,n , the exhaust temperature data sequence at time t ₁ As the reference sequence of the training set; the exhaust temperature data sequence at time t _m As the comparison sequence to be processed in the training set;

is the exhaust temperature data sequence of m moments, T _4,1 , T _4,2 ,..., T _4,n is the exhaust temperature data sequence of n measuring points;

said is the exhaust temperature data sequence at time t ₂ , T _4,1 is the exhaust temperature data sequence of the first measuring point, T _4,2 is the exhaust temperature data sequence of the first measuring point, T _4,n is the nth Exhaust temperature data series of measuring points;

Step 2: Define the comparison sequence to be processed in the training set Reference sequence with training set The deviation of e is e, and the minimum value of e is determined by the Euclidean distance discriminant method, and the exhaust temperature data sequence after rotation correction at m moments when e takes the minimum value is obtained and the rotation-corrected exhaust temperature data sequence T′ _4,1 , T′ _4,2 ,…, T′ _4,n of n measuring points;

Step 3: Determine the relationship between the rotation-corrected exhaust temperature data series T′ _4,1 , T′ _4,2 , ..., T′ _4,n of n measuring points and the average temperature of n measuring points;

Step 4: Exhaust temperature data sequence corrected by rotation at m moments determined in step 3 and the relationship between the average temperature of n measuring points to obtain the predicted temperature values T _4pre,1 ,T _4pre,2 ,...,T _4pre,i ,...,T _{4pre,n of n} measuring points;

Step 5: According to the temperature prediction values T _4pre,1 ,T _4pre,2 ,…,T _4pre,i ,…,T _4pre,n of n measuring points obtained in Step 4, get the error band ΔT ₁ of each measuring point ,ΔT ₂ ,…,ΔT _i ,…,ΔT _n ; where ΔT _i is the error band of the i-th measuring point;

Step 6: When monitoring, re-execute steps 1 to 5. If the error band of the i-th measuring point is within the range of ΔT _i , it means that the gas turbine unit is not faulty, otherwise the gas turbine unit is faulty.

During monitoring, the exhaust temperature data measured at each moment is substituted into the predicted value T _{4pre,1 of measuring point 1} . The difference between the predicted value T _4pre,1 of measuring point 1 and the actual value (corrected value) T′ _4,1t of measuring point 1 when there is no fault in the monitoring unit at each moment is ΔT ₁ , which means that the unit has no fault. If it exceeds this range , it means that the unit is malfunctioning.

The difference between the predicted value and the actual value ΔT ₂ , ΔT ₃ ,...,ΔT _n of the measuring points 2 to n when the unit has no faults is respectively obtained. If the deviation is within the range of ΔT _i , it means that the unit has no faults. If it exceeds this range, It means that the unit is malfunctioning.

The training set data is shown in Figure 5, and the test set results are shown in Figures 6 and 7.

Specific embodiment 2: The difference between this embodiment and specific embodiment 1 is that in the step 1, the temperature exhaust data series at m times are obtained And the exhaust temperature data series T _4,1 , T _4,2 ,..., T _4,n of n measuring points are specifically:

Exhaust temperature data series at m time points for:

...

in are the temperature values from the first measuring point to the nth measuring point at time _t1 respectively,

are the temperature values from the first measuring point to the nth measuring point at time t _m respectively;

The exhaust temperature data series T _4,1 , T _4,2 ,..., T _4,n of n measuring points are specifically:

...

in are the temperature values of the first measuring point from _t1 to _tm , respectively, are the temperature values at the second measuring point from t ₁ to t _m respectively, are the temperature values at the nth measuring point from _t1 to _tm , respectively.

Other steps and parameters are the same as those in Embodiment 1.

Specific embodiment 3: The difference between this embodiment and specific embodiment 1 or 2 is that in the step 2, the rotation-corrected exhaust temperature data sequence at m moments when e takes the minimum value is obtained The specific process of the exhaust temperature data sequence T′ _4,1 , T′ _4,2 , ..., T′ _4,n after rotation correction of n measuring points is:

When the effect of rotation is not considered, the sequence of comparisons to be processed in the training set Reference sequence with training set The deviation e ₀ of is:

When rotating a survey point, the sequence of comparisons to be processed in the training set Reference sequence with training set The deviation e1 _of is:

when spinning

The sequence of comparisons to be processed for the training set when rotating two survey points Reference sequence with training set The deviation _e2 of is:

When rotating N measurement points, the comparison sequence to be processed in the training set Reference sequence with training set The deviation e _N of is:

where e=[e ₀ ,e ₁ ,...,e _N ];

Take the smallest value in e:

e _min ＝argmin(e ₁ ,e ₂ ,…,e _N )

argmin represents the minimum value;

When N measuring points are rotated, e takes the minimum value, and the above operation is carried out for the temperature exhaust sequence at different times, and the corrected exhaust temperature data sequence can be rotated at m times:

...

in Respectively, the temperature values after rotation correction from the first measuring point to the nth measuring point at time _t1 , Respectively, the temperature values after rotation correction from the first measuring point to the nth measuring point at time t _m ;

For example, at time _t1 , when the sequence to be processed (comparison) of the training set rotates N measurement points, the deviation e takes the minimum value, at this time It can be seen that at time t _m ,

The exhaust temperature data series T ₄ ' _,1 , T ₄ ' _,2 ,..., T ₄ ' _,n of n measuring points after rotation correction are specifically:

...

in are the temperature values after rotation correction of the first measuring point from time t ₁ to time t _m respectively, are the temperature values after rotation correction of the second measuring point from time t ₁ to time t _m respectively, Respectively, the temperature values after rotation correction of the nth measuring point from time t ₁ to time t _m .

Other steps and parameters are the same as those in Embodiment 1 or Embodiment 2.

Embodiment 4: The difference between this embodiment and one of Embodiments 1 to 3 is that in the step 3, the exhaust temperature data sequence T ₄ ' _,1 and T ₄ ' _,2 after rotation correction of n measuring points are determined ,..., the relationship between T ₄ ' _,n and the average temperature of n measuring points is specifically:

In the combustion chamber, the distribution of the outlet exhaust temperature reflects the operation of the combustion chamber. At the same time, in order to eliminate the scaling of different measuring point temperatures under different working conditions, the average temperature is used for calculation. The temperature of each measuring point at the outlet satisfies a linear relationship, that is, for each measuring point, there is a linear relationship between the exhaust temperature after rotation correction and the average temperature:

Where T′ _4,i is the exhaust temperature data sequence after rotation correction of the i-th measuring point, i=[1,2,…,n]; is the average temperature of n measuring points, α _i is the average temperature fitting parameter, and β _i is the constant term fitting parameter;

i is the exhaust temperature measuring point number, and there are n measuring points in total. α _i , β _i are the fitting parameters, and the determination method of α _i , β _i is as follows: Assume that the gas turbine has n thermocouples uniformly distributed at the exhaust end, and for any measuring point i, i=1,2,...n select the unit The data during normal operation for a period of time is used as the training set;

at each time point in the time period The values of the parameters α _i and β _i corresponding to the measuring point are determined by the least square method.

Other steps and parameters are the same as those in Embodiments 1 to 3.

Embodiment 5: The difference between this embodiment and Embodiment 1 to Embodiment 4 is that the temperature prediction values T _4pre,1 , T _4pre,2 ,..., T _4pre,i of n measuring points are obtained in the step 4 ,...,T _4pre,n is calculated as:

According to the functional relationship in step 3, the average exhaust temperature under normal operating conditions at time t is calculated as Substituting in, the temperature prediction value of each measuring point can be determined:

Other steps and parameters are the same as in one of the specific embodiments 1 to 4.

Embodiment 6: The difference between this embodiment and one of Embodiments 1 to 5 is that the temperature prediction values T _4pre,1 , T _4pre,2 ,..., T _4pre,i ,…,T _4pre,n , the calculation formula of the error band ΔT ₁ ,ΔT ₂ ,…,ΔT _i ,…,ΔT _n of each measuring point is:

ΔT _i =T _4pre,i -T′ _4,1t

Where T′ _4,1t is the temperature value after rotation correction at the first measuring point at time t, t=[t ₁ ,t ₂ ,…,t _m ].

The difference between the theoretical value T _4pre,i of measuring point i and the measured value T′ _4,1t of measuring point i at time t can be calculated as ΔT _i , ΔT _i =T _4pre,i -T′ _4,1t . This error is called the error band of the i measuring point. When calculating, the exhaust temperature data at each moment are substituted into step 4 to obtain the time series of exhaust temperature prediction values of each measuring point when the unit has no faults, and then use ΔT _i =T _4pre,i -T′ _{4, 1t} calculates the error band of the measuring point.

Other steps and parameters are the same as one of the specific embodiments 1 to 5.

The present invention can also have other various embodiments, without departing from the spirit and essence of the present invention, those skilled in the art can make various corresponding changes and deformations according to the present invention, but these corresponding changes and deformations are all Should belong to the scope of protection of the appended claims of the present invention.

Claims (6)

1. A gas turbine high-temperature component failure early warning method considering the exhaust temperature field rotation, characterized in that: the gas turbine component failure early warning method considering the exhaust temperature field rotation comprises the following steps: Step 1: The exhaust temperature of the gas turbine is measured by n thermocouples evenly distributed in the circumferential direction of the turbine outlet. During the normal operation time of the gas turbine [t ₁ , t _m ], the exhaust temperature data sequence of n measuring points at m times is obtained. Respectively constitute the exhaust temperature data series and T _4,1 , T _4,2 ,..., T _4,n , the exhaust temperature data sequence at time t ₁ As the reference sequence of the training set; the exhaust temperature data sequence at time t _m As the comparison sequence to be processed in the training set; said is the exhaust temperature data sequence at time t ₂ , T _4,1 is the exhaust temperature data sequence of the first measuring point, T _4,2 is the exhaust temperature data sequence of the first measuring point, T _4,n is the nth Exhaust temperature data series of measuring points; Step 2: Define the comparison sequence to be processed in the training set Reference sequence with training set The deviation of e is e, and the minimum value of e is determined by the Euclidean distance discriminant method, and the exhaust temperature data sequence after rotation correction at m moments when e takes the minimum value is obtained and the rotation-corrected exhaust temperature data sequence T′ _4,1 , T′ _4,2 ,…, T′ _4,n of n measuring points; Step 3: Determine the relationship between the rotation-corrected exhaust temperature data series T′ _4,1 , T′ _4,2 , ..., T′ _4,n of n measuring points and the average temperature of n measuring points; Step 4: Exhaust temperature data sequence corrected by rotation at m moments determined in step 3 and the relationship between the average temperature of n measuring points to obtain the predicted temperature values T _4pre,1 ,T _4pre,2 ,...,T _4pre,i ,...,T _{4pre,n of n} measuring points; Step 5: According to the temperature prediction values T _4pre,1 ,T _4pre,2 ,…,T _4pre,i ,…,T _4pre,n of n measuring points obtained in Step 4, get the error band ΔT ₁ of each measuring point ,ΔT ₂ ,…,ΔT _i ,…,ΔT _n ; where ΔT _i is the error band of the i-th measuring point; Step 6: When monitoring, re-execute steps 1 to 5. If the error band of the i-th measuring point is within the range of ΔT _i , it means that the gas turbine unit is not faulty, otherwise the gas turbine unit is faulty. 2. According to claim 1, a gas turbine high-temperature component failure warning method considering the rotation of the exhaust temperature field is characterized in that: in the step 1, the exhaust temperature data sequence at m moments is obtained And the exhaust temperature data series T _4,1 , T _4,2 ,..., T _4,n of n measuring points are specifically: Exhaust temperature data series at m time points for: ... in are the temperature values from the first measuring point to the nth measuring point at time _t1 respectively, are the temperature values from the first measuring point to the nth measuring point at time t _m respectively; The exhaust temperature data series T _4,1 , T _4,2 ,..., T _4,n of n measuring points are specifically: ... in are the temperature values of the first measuring point from _t1 to _tm , respectively, are the temperature values at the second measuring point from t ₁ to t _m respectively, are the temperature values at the nth measuring point from _t1 to _tm , respectively. 3. According to claim 1 or 2, a gas turbine high-temperature component fault warning method considering the rotation of the exhaust temperature field, is characterized in that: in the step 2, the m moments when e takes the minimum value are obtained after the rotation correction Exhaust temperature data series The specific process of the exhaust temperature data sequence T′ _4,1 , T′ _4,2 , ..., T′ _4,n after rotation correction of n measuring points is: When the effect of rotation is not considered, the sequence of comparisons to be processed in the training set Reference sequence with training set The deviation e ₀ of is: When rotating a survey point, the sequence of comparisons to be processed in the training set Reference sequence with training set The deviation e1 _of is: When rotating When rotating two measurement points, the sequence of comparisons to be processed in the training set Reference sequence with training set The deviation _e2 of is: When rotating N measurement points, the comparison sequence to be processed in the training set Reference sequence with training set The deviation e _N of is: where e=[e ₀ ,e ₁ ,...,e _N ]; Take the smallest value in e: e _min ＝arg min(e ₁ ,e ₂ ,…,e _N ) arg min means take the minimum value; When N measuring points are rotated, e takes the minimum value, and the exhaust temperature data series after rotation correction at m moments: ... in Respectively, the temperature values after rotation correction from the first measuring point to the nth measuring point at time _t1 , Respectively, the temperature values after rotation correction from the first measuring point to the nth measuring point at time t _m ; at this time Determine the moment t _m , The exhaust temperature data sequence T′ _4,1 , T′ _4,2 ,…, T′ _4,n of n measuring points after rotation correction is specifically: ... in are the temperature values after rotation correction of the first measuring point from time t ₁ to time t _m respectively, are the temperature values after rotation correction of the second measuring point from time t ₁ to time t _m respectively, Respectively, the temperature values after rotation correction of the nth measuring point from time t ₁ to time t _m . 4. According to claim 3, a gas turbine high-temperature component failure warning method considering the rotation of the exhaust temperature field is characterized in that: in the step 3, the exhaust temperature data sequence T′ ₄ after the rotation correction of n measuring points is determined _,1 , T′ _4,2 ,…, the relationship between T′ _4,n and the average temperature of n measuring points is specifically: Where T′ _4,i is the exhaust temperature data sequence after rotation correction of the i-th measuring point, i=[1,2,…,n]; is the average temperature of n measuring points, α _i is the average temperature fitting parameter, and β _i is the constant term fitting parameter; The least square method determines the values of α _i , β _i . 5. According to claim 4, a gas turbine high-temperature component failure warning method considering the rotation of the exhaust temperature field, is characterized in that: the temperature prediction values T _4pre,1 and T _4pre, of n measuring points are obtained in the step 4. ₂ ,…,T _4pre,i ,…,T _4pre,n are calculated as follows: 6. According to claim 5, a gas turbine high-temperature component failure warning method considering the rotation of the exhaust temperature field is characterized in that: in the step 5, the temperature prediction values T _4pre,1 of the n measuring points obtained according to the step 4 ,T _4pre,2 ,…,T _4pre,i ,…,T _4pre,n , the calculation formula of the error band ΔT ₁ ,ΔT ₂ ,…,ΔT _i ,…,ΔT _n of each measuring point is: ΔT _i =T _4pre,i -T′ _4,it Where T′ _4,it is the temperature value of the i-th measuring point after rotation correction at time t, t=[t ₁ ,t ₂ ,…,t _m ].