Early warning method for abnormal operation of hot component of gas turbine

Abstract

According to the early warning method for the abnormal operation of the hot component of the gas turbine, the early warning model for the abnormal operation of the hot component of the gas turbine is established based on historical data, real-time early warning for the abnormal operation of the hot component of the gas turbine is achieved, and the abnormal condition of the hot component of the gas turbine can be found in advance. The method comprises the specific steps of firstly obtaining historical data of gas turbine exhaust temperature measuring points and power generation loads, then analyzing the relation between the historical data of all the measuring points of the gas turbine hot component, and establishing a gas turbine hot component abnormity early warning model; and finally, according to the gas turbine hot part abnormity early warning model and the current real-time data of the gas turbine, whether operation of the gas turbine hot part is abnormal or not and whether early warning prompt is sent out or not are judged.

Description

Technical Field

[0001] The invention belongs to the field of gas turbine generator sets, and in particular relates to an early warning method for abnormal operation of a hot component of a gas turbine. Background Art

[0002] The hot component is one of the most important components of the gas turbine, and also one of the most expensive components. The safety of its operation has an important impact on the safe operation of the entire power plant. According to statistics, among various gas turbine failures, more than 50% of the failures are related to high-temperature hot components of the gas turbine. The output power of modern heavy-duty gas turbines is getting higher and higher, and the initial temperature of combustion is getting higher and higher. The initial temperature of E-class is about 1200℃, the initial temperature of F-class is about 1350℃, and the initial temperature of H / G-class is about 1500℃. In the future, the initial temperature of the gas turbine will reach more than 1600℃. The higher the initial temperature, the worse the operating environment of the hot component, and the greater the loss caused by the failure. Therefore, it is particularly important to carry out early abnormal warning for the hot components of the gas turbine.

[0003] Currently, gas turbine power plants mainly rely on the combustion protection system provided by the gas turbine manufacturer to monitor hot components. However, from the actual application effect, when the monitoring system issues an alarm, the hot components of the gas turbine are often severely damaged, or the gas turbine can only reduce the load or shut down when the alarm is issued. Moreover, the alarm and gas turbine shutdown signal are only given when the damage of the hot components reaches a certain extent. Summary of the invention

[0004] Purpose of the invention: The purpose of the invention is to improve the monitoring level of the operating status of the thermal components of the gas turbine, establish a model for judging the abnormal operation of the thermal components of the gas turbine according to the historical data of the operation of the gas turbine, and realize the early warning of the abnormal operation of the thermal components of the gas turbine. The method of the invention comprises the following steps:

[0005] Step 1: Check the engine operation log or communicate with the operator to determine whether the engine hot parts have failed during this overhaul period and when the failure occurred;

[0006] Step 2, collecting the fault-free historical data of the gas turbine exhaust temperature measurement point and the gas turbine load measurement point in the historical database;

[0007] Step 3: Establish an abnormal warning model for thermal components of the gas turbine based on the fault-free historical data;

[0008] Step 4: Provide real-time warning for the thermal components of the gas turbine according to the warning model.

[0009] In step 2, the engine exhaust temperature (exhaust temperature for short) measuring points are numbered in a clockwise or counterclockwise order. The engine exhaust temperature measuring points can be numbered consecutively according to the naming rules of the exhaust temperature in the engine TCS. For example, the engine #2 exhaust temperature measuring point is preceded by the engine #1 exhaust temperature measuring point, and followed by the engine #3 exhaust temperature measuring point, which can be recorded as t01, t02, ..., tn, for a total of n engine exhaust temperature measuring points.

[0010] In step 2, the historical database is the historical database of the DCS (Distributed Control System) control system, or the real-time database of the SIS (Supervisor Information System) monitoring system; the gas turbine exhaust temperature is the gas turbine blade channel temperature or the gas turbine exhaust temperature; the fault-free historical data requires excluding the historical data within 3 months before the fault occurs, and the total data length can cover all months of 1 year.

[0011] Step 3 includes:

[0012] 3.1. Extract historical data of all gas turbine exhaust temperature measurement points and gas turbine load measurement points from the fault-free historical data at a sampling rate of 1 minute (or 30 seconds) to form a sample data set;

[0013] 3.2. Check the integrity of the data and delete incomplete sample data. If a sample lacks data for a certain gas turbine exhaust temperature measurement point or load measurement point, delete the entire sample.

[0014] 3.3, delete the sample data during the shutdown and startup process from the sample data set. The deletion rule is to delete the sample data when the gas turbine load is lower than 30% (or 40%) of the rated load of the gas turbine, and finally obtain a valid sample data set;

[0015] 3.4, for each sample data, perform the following steps:

[0016] 3.4.1. Sort all the gas turbine exhaust temperature measurement points in the sample data in order from large to small (or from small to large);

[0017] 3.4.2. For the sorted sample data, delete the largest 1 to 3 gas turbine exhaust temperature measurement point values, and delete the smallest 1 to 3 gas turbine exhaust temperature measurement point values; the number of deletions is selected based on the total number of gas turbine exhaust temperature measurement points. If the total number of gas turbine exhaust temperature measurement points is more than 20, 3 can be selected; if the total number of gas turbine exhaust temperature measurement points is less than 10, 1 can be selected; in other cases, 2 can be selected.

[0018] 3.4.3, calculate the average value of the remaining gas turbine exhaust temperature measurement points as the exhaust temperature average value tp_avg of this sample;

[0019] 3.4.4, calculate the deviation tp_dev between each gas turbine exhaust temperature measurement point and the exhaust temperature average value tp_avg. If the deviation is higher than the exhaust temperature average value, it is positive; if the deviation is lower than the exhaust temperature average value, it is negative;

[0020] 3.5. Statistic all sample calculation results to find the maximum deviation and minimum deviation of each gas turbine exhaust temperature measurement point, which are used as the key attributes of the gas turbine exhaust temperature measurement point. The maximum deviation is named as the upper deviation tp_max, and the minimum deviation is named as the lower deviation tp_min.

[0021] 3.6, select the threshold coefficient k = 1.3, and determine the alarm threshold of each gas turbine exhaust temperature measurement point according to the following method, so as to determine the early warning model:

[0022] 3.6.1, calculate the average value of the deviation dev_avg, dev_avg is equal to half of the sum of the upper deviation tp_max and the lower deviation tp_min;

[0023] 3.6.2, calculate the range of deviation, range is equal to the upper deviation tp_max minus the lower deviation tp_min;

[0024] 3.6.3, according to the value of the threshold coefficient k, enlarge the deviation range:

[0025] krange=k*range,

[0026] Where krange represents the deviation range after enlargement;

[0027] 3.6.4, upper deviation alarm threshold th_upper = dev_avg + 0.5*krange, lower deviation alarm threshold th_lower = dev_avg - 0.5*krange, determine the non-alarm value range: [dev_avg - 0.5*krange, dev_avg + 0.5*krange].

[0028] Step 4 includes:

[0029] 4.1, read the real-time data of the current gas turbine load and gas turbine exhaust temperature measurement points;

[0030] 4.2, if the engine load is lower than the rated load of the engine by 30% (or 40%), it is determined that the calculation conditions are not met, and no early warning judgment is made, waiting for the next real-time data of the engine load and engine exhaust temperature measurement point to be read;

[0031] 4.3. If the engine load is equal to or higher than 30% (or 40%) of the engine rated load, calculate as follows:

[0032] 4.3.1, sort the values ​​of the exhaust temperature measurement points of the gas turbine according to their size from high to low (or from low to high) to obtain an ordered exhaust temperature list;

[0033] 4.3.2, delete the largest 1 to 3 measurement point values ​​and the smallest 1 to 3 measurement point values ​​from the ordered temperature discharge list. The number of deleted points is the same as that in 3.4.2;

[0034] 4.3.3, calculate the average value of the remaining engine exhaust temperature measurement points as the exhaust temperature average value cur_avg for this calculation;

[0035] 4.3.4, calculate the deviation tp_dev between each gas turbine exhaust temperature measurement point and the average value;

[0036] 4.3.5, for each gas turbine exhaust temperature measurement point, compare whether the deviation tp_dev is within the normal value range, that is, the real-time deviation value is higher than the lower deviation threshold th_lower or lower than the upper deviation threshold th_upper, otherwise it is not within the normal value range;

[0037] 4.3.6, if the exhaust temperature real-time deviation value is within the normal range, it indicates that the current operation is normal, otherwise it is marked as abnormal and waits for the next calculation;

[0038] 4.4, according to the abnormal mark, real-time alarm is carried out according to the following rules:

[0039] 4.4.1. If the data of a single gas turbine exhaust temperature measurement point is marked as abnormal for 30 to 60 seconds, an alarm signal will be issued; if high sensitivity is required, 30 seconds can be selected; if a low false alarm rate is required, 60 seconds can be selected;

[0040] 4.4.2, If two or more measuring points with consecutive numbers report abnormalities at the same time, there is no need to wait and an alarm signal will be issued directly.

[0041] In step 3.1, the gas turbine load and the exhaust temperature measurement point values ​​of each gas turbine at the same time constitute a sample data.

[0042] In step 4.4, when a real-time warning is issued, the gas turbine load and the exhaust temperature measurement point values ​​of each gas turbine read at the same time are input into the warning model as a sample.

[0043] The present invention also provides an electronic device, comprising a processor and a memory, wherein the memory stores program code, and when the program code is executed by the processor, the processor executes the steps of the described method.

[0044] The present invention also provides a storage medium storing a computer program or instruction. When the computer program or instruction is run on a computer, the steps of the method described are executed.

[0045] The present invention has the following beneficial effects: the present invention is based on analyzing the health history data of the gas turbine and establishing a gas turbine thermal component operation abnormality judgment model, thereby realizing a gas turbine thermal component operation abnormality early warning method, which does not require additional hardware equipment and can detect the operation abnormality of the gas turbine thermal component in advance. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] The present invention will be further described in detail below in conjunction with the accompanying drawings and specific embodiments, and the above and / or other advantages of the present invention will become more clear.

[0047] Figure 1 The present invention is a general flow chart of the abnormal operation warning of the thermal components of the gas engine.

[0048] Figure 2 A flowchart of the historical data preprocessing process is modeled for the present invention.

[0049] Figure 3 A flowchart of the abnormal warning modeling of the thermal components of the gas engine according to the present invention.

[0050] Figure 4 This is a flowchart of the real-time early warning of the hot parts of the gas engine of the present invention. DETAILED DESCRIPTION

[0051] like Figure 1 As shown, the implementation case of the present invention provides an early warning method for abnormal operation of thermal components of a gas turbine. The implementation object of this case is an M701F4 gas turbine. The gas turbine has 20 inter-turbine temperature measuring points as exhaust temperature measuring points of the gas turbine.

[0052] Step 1, such as Figure 2 As shown, historical data is obtained and a valid sample set is generated.

[0053] 1.1. The gas turbine operates normally between 0:00 on February 1, 2023 and 0:00 on February 1, 2024. The historical data of this period is selected as the original data for analysis.

[0054] 1.2, Select features, select 1 gas turbine load measurement point, 20 gas turbine exhaust temperature measurement points, a total of 21 measurement points.

[0055] 1.3, according to the sampling interval of 1 minute, export the historical data of 21 measurement points from 0:00 on February 1, 2023 to 0:00 on February 1, 2024, and export the data in the format of csv file. Each line in the csv file is a sample, except for the first column which is the timestamp, it contains 21 columns of data corresponding to 21 measurement points.

[0056] 1.4, Check the integrity of the exported data, check if there are any missing columns in the csv file, and if there are any rows with missing values.

[0057] 1.5, data validity check, filter historical data according to the value of the gas turbine load measurement point; delete the samples in the historical data where the gas turbine load is less than 100.

[0058] 1.6, the gas turbine exhaust temperature measurement points are numbered clockwise according to the equipment installation situation. The gas turbine exhaust temperature measurement points in the DCS measurement point table have been numbered in clockwise order, so the corresponding gas turbine exhaust temperature measurement point No. 1 is renamed T01, the No. 2 measurement point is renamed T02, ..., and the gas turbine exhaust temperature measurement point No. 20 is named T20. The column header names of the gas turbine exhaust temperature measurement point columns in the CSV file are updated to T01 to T20 and rearranged in order.

[0059] 1.7, so far, the valid sample set has been completed.

[0060] Step 2: Generate an early warning model based on the valid sample set, such as Figure 3 shown.

[0061] 2.1, the total number of valid samples is 155002 rows, and the number of gas turbine exhaust temperature measurement points is 20.

[0062] 2.2, loop each sample:

[0063] 2.2.1. Sort the values ​​of the gas turbine exhaust temperature measurement points in the sample from large to small.

[0064] 2.2.2. Delete the three largest and three smallest values ​​from the sorted list of gas turbine exhaust temperature measurement point values.

[0065] 2.2.3. Calculate the average value of the remaining engine exhaust temperature measurement points as the sample average value avg.

[0066] 2.2.4. Calculate the deviations of all 20 gas turbine exhaust temperature measurement points from the sample average value avg and record them as T01_dev, T02_dev, …, T20_dev.

[0067] 2.3. Count the maximum and minimum values ​​of each exhaust temperature deviation value in all samples: Taking T01_dev as an example, find the maximum value from T01_dev of all samples and record it as T01_dev_max, and find the minimum value and record it as T01_dev_min.

[0068] 2.4, for each gas turbine exhaust temperature measurement point, take the threshold coefficient k = 1.3 to determine its deviation threshold range. The following takes the gas turbine exhaust temperature measurement point T01 as an example for explanation.

[0069] 2.4.1. Calculate the deviation range of the gas turbine exhaust temperature measuring point T01: range = T01_dev_max – T01_dev_min.

[0070] 2.4.2. Calculate the mean deviation of the exhaust temperature measuring point T01 of the combustion engine: dev_avg = (T01_dev_max + T01_dev_min) / 2.

[0071] 2.4.3. Calculate the upper threshold of the engine exhaust temperature measuring point T01: th_max = dev_avg + range*k / 2.

[0072] 2.4.4, calculate the lower limit threshold of the engine exhaust temperature measuring point T01: th_min = dev_avg-range*k / 2.

[0073] 2.5, after the threshold range [th_min, th_max] of each gas turbine exhaust temperature measurement point is determined, the early warning model is determined.

[0074] Step 3: Based on the early warning model, read the real-time data and issue a real-time early warning, such as Figure 4 shown.

[0075] 3.1. Read the real-time values ​​of the gas turbine load measurement point and each gas turbine exhaust temperature measurement point (renamed) from the DCS real-time database once per second, recorded as Load, T01, T02, ..., T20, as input data.

[0076] 3.2. According to the engine load value, determine whether to continue the calculation: if Load>=100, perform subsequent calculations; otherwise, return to the previous step and wait for 1 second before reading the next real-time data.

[0077] 3.3, follow the steps below to process the engine exhaust temperature measurement points

[0078] 3.3.1. Arrange the gas turbine exhaust temperature measurement points in descending order.

[0079] 3.3.2, delete the largest 3 measurement point values ​​and the smallest 3 measurement point values.

[0080] 3.3.3. Calculate the average value avg of the remaining combustion engine exhaust temperature measurement point values.

[0081] 3.3.4. Calculate the deviation between each gas turbine exhaust temperature measurement point and the average value avg. Taking T01 measurement point as an example, its deviation is T01_dev=T01-avg.

[0082] 3.3.5, determine whether the deviation value of each combustion engine exhaust temperature measuring point is within the threshold range [th_min, th_max]. If so, set the measuring point status to normal, otherwise set the measuring point status to abnormal.

[0083] 3.4, make alarm judgment according to the following steps.

[0084] 3.4.1, check the status of each gas turbine exhaust temperature measuring point to see if there is any abnormal measuring point; if not, return to 3.1 for the next calculation; if there is an abnormal point, continue.

[0085] 3.4.2, count the number of abnormal status measurement points. If the number of measurement points is greater than or equal to 2, proceed to the next step, otherwise jump to 3.4.6.

[0086] 3.4.3, check whether multiple abnormal state measuring points are adjacent to each other. The adjacent judgment rule is explained by taking measuring point T01 as an example. T02 and T20 are adjacent to T01, and the other measuring points are not adjacent to T01.

[0087] 3.4.4, if multiple abnormal state measuring points have adjacent measuring points, the adjacent abnormal measuring points are set to adjacent alarm state, and return to 3.1 for the next calculation.

[0088] 3.4.5. If there are no adjacent measuring points for multiple abnormal status measuring points, proceed to the next step.

[0089] 3.4.6, for each abnormal measurement point, calculate the duration of its abnormal state, recorded as time.

[0090] 3.4.7, if time is greater than 1 minute, the abnormal measuring point status is set to continuous abnormal alarm status, otherwise the measuring point abnormal status is maintained.

[0091] 3.4.8, after all abnormal measurement points have been processed, return to 3.1 and proceed to the next step of calculation.

[0092] 3.5. At this point, a cycle of real-time warning is completed.

[0093] The present invention provides an early warning method for abnormal operation of a hot component of a gas turbine. There are many methods and ways to implement the technical solution. The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the protection scope of the present invention. All components not specified in this embodiment can be implemented by existing technologies.

Claims

1. A method for early warning of abnormal operation of a gas turbine hot component, characterized in that: The following steps are involved: Step 1: Check the engine operation log to determine whether the engine hot parts have failed during this overhaul period and the time when the failure occurred; Step 2, collecting the fault-free historical data of the gas turbine exhaust temperature measurement point and the gas turbine load measurement point in the historical database; Step 3: Establish an abnormal warning model for thermal components of the gas turbine based on the fault-free historical data; Step 4: Provide real-time warning for the thermal components of the gas turbine according to the warning model.

2. The method according to claim 1, characterized in that In step 2, the gas turbine exhaust temperature measuring points are numbered in a clockwise or counterclockwise order.

3. The method according to claim 2, characterized in that In step 2, the historical database is a historical database of a DCS control system, or a real-time database of a SIS monitoring system; the gas turbine exhaust temperature is a gas turbine blade channel temperature or a gas turbine exhaust temperature.

4. The method according to claim 3, characterized in that Step 3 includes: 3.

1. Extract the historical data of all gas turbine exhaust temperature measurement points and gas turbine load measurement points from the fault-free historical data to form a sample data set; 3.2, check the integrity of the data and delete incomplete sample data; 3.3, delete the sample data during the shutdown and startup process from the sample data set, and finally obtain a valid sample data set; 3.4, for each sample data, perform the following steps: 3.4.1, sort all the gas turbine exhaust temperature measurement points in the sample data; 3.4.

2. For the sorted sample data, delete the largest 1 to 3 gas turbine exhaust temperature measurement point values ​​and the smallest 1 to 3 gas turbine exhaust temperature measurement point values; 3.4.3, calculate the average value of the remaining gas turbine exhaust temperature measurement points as the exhaust temperature average value tp_avg of this sample; 3.4.4, calculate the deviation tp_dev between each gas turbine exhaust temperature measurement point and the exhaust temperature average value tp_avg. If the deviation is higher than the exhaust temperature average value, it is positive; if the deviation is lower than the exhaust temperature average value, it is negative; 3.

5. Statistic all sample calculation results to find the maximum deviation and minimum deviation of each gas turbine exhaust temperature measurement point, which are used as the key attributes of the gas turbine exhaust temperature measurement point. The maximum deviation is named as the upper deviation tp_max, and the minimum deviation is named as the lower deviation tp_min. 3.6, select the threshold coefficient k, and determine the alarm threshold of each gas turbine exhaust temperature measurement point according to the following method, so as to determine the early warning model: 3.6.1, calculate the average value of the deviation dev_avg, dev_avg is equal to half of the sum of the upper deviation tp_max and the lower deviation tp_min; 3.6.2, calculate the range of deviation, range is equal to the upper deviation tp_max minus the lower deviation tp_min; 3.6.3, according to the value of the threshold coefficient k, enlarge the deviation range: krange=k*range, Where krange represents the deviation range after enlargement; 3.6.4, upper deviation alarm threshold th_upper = dev_avg + 0.5*krange, lower deviation alarm threshold th_lower = dev_avg - 0.5*krange, determine the non-alarm value range: [dev_avg - 0.5*krange, dev_avg + 0.5*krange].

5. The method according to claim 4, characterized in that Step 4 includes: 4.1, read the real-time data of the current gas turbine load and gas turbine exhaust temperature measurement points; 4.

2. If the engine load is lower than a certain range of the engine rated load, it is determined that the calculation conditions are not met, and no early warning judgment is made, waiting for the next engine load and engine exhaust temperature measurement point real-time data to be read; 4.3, If the engine load is equal to or higher than a certain range of the engine rated load, calculate according to the following steps: 4.3.1, sort the values ​​of the exhaust temperature measurement points of the gas turbine according to their size from high to low, and obtain an ordered exhaust temperature list; 4.3.2, delete the largest 1 to 3 measurement point values ​​and the smallest 1 to 3 measurement point values ​​from the ordered temperature discharge list; 4.3.3, calculate the average value of the remaining engine exhaust temperature measurement points as the exhaust temperature average value cur_avg for this calculation; 4.3.4, calculate the deviation tp_dev between each gas turbine exhaust temperature measurement point and the average value; 4.3.5, for each gas turbine exhaust temperature measurement point, compare whether the deviation tp_dev is within the normal value range, that is, the real-time deviation value is higher than the lower deviation threshold th_lower or lower than the upper deviation threshold th_upper, otherwise it is not within the normal value range; 4.3.6, if the exhaust temperature real-time deviation value is within the normal range, it indicates that the current operation is normal, otherwise it is marked as abnormal and waits for the next calculation; 4.4, according to the abnormal mark, real-time alarm is carried out according to the following rules: 4.4.

1. If the data of a single gas turbine exhaust temperature measurement point is marked as abnormal for a certain period of time, an alarm signal will be issued; 4.4.2, If two or more measuring points with consecutive numbers report abnormalities at the same time, there is no need to wait and an alarm signal will be issued directly.

6. The method according to claim 5, characterized in that In step 3.1, the gas turbine load and the exhaust temperature measurement point values ​​of each gas turbine at the same time constitute a sample data.

7. The method according to claim 6, characterized in that In step 4.4, when a real-time warning is issued, the gas turbine load and the exhaust temperature measurement point values ​​of each gas turbine read at the same time are input into the warning model as a sample.

8. An electronic device, characterized in that: The method comprises a processor and a memory, wherein the memory stores program codes, and when the program codes are executed by the processor, the processor executes the steps of the method according to any one of claims 1 to 7.

9. A storage medium, characterized in that: A computer program or instruction is stored, and when the computer program or instruction is run on a computer, the steps of the method according to any one of claims 1 to 7 are executed.

Images