JPH0676960B2 - Method and apparatus for evaluating remaining life of mechanical structure subjected to repeated load - Google Patents

Description

Description: TECHNICAL FIELD The present invention relates to a method for evaluating the remaining life of a machine structure subjected to cyclic load and its apparatus, and particularly to a cyclic load or a fluctuating load under a high temperature atmosphere. The present invention relates to a method and apparatus for evaluating the remaining life of a mechanical structure to be received.

[Background of the Invention]

Turbine for a power plant is an example of a mechanical structure that receives a repeated load under a high temperature atmosphere. The turbine for a power generation plant is repeatedly started and stopped, or is accompanied by a change in load. The mechanical structures that make up such turbines are simultaneously subject to thermal fatigue and creep damage. As a result, the accumulation of damage resulting from fatigue and creep causes cracks in the structural member and reduces the structural strength. In such a case, unless the strength residual life of the mechanical structure is evaluated, there is a risk that the equipment may be damaged and a power plant accident may occur. Especially turbine casings,
It is important to evaluate the remaining life of the rotor and various valves in terms of strength.

Conventionally, when designing this kind of equipment, a large safety factor was set based on the creep strength of the material used, and its strength reliability was made high. Therefore, the plant was operated with almost no evaluation of the remaining life of the equipment. However, at present, about half of the thermal power plants in operation have exceeded the design service life, and it is time to decide whether these plants can be used as they are or whether they should be replaced with new plants. Establishment of evaluation method is desired.

On the other hand, due to changes in the power demand pattern, operation is performed such that a power plant designed for base load is started and stopped daily for variable load, and in such a plant with severe operating conditions, It is estimated that the useful life will naturally be shortened. In such cases, it is extremely important to evaluate how much remaining life is.

If the strength of the turbine casing, the main valve, and the control valve and the degree of damage to the components can be determined by non-destructive inspection, the reliability of the plant will be greatly improved, but this is still possible. Not enough has been proposed.

Slightly, for turbine rotors, measure the steam and gas temperatures around the rotor and the temperature on the inside surface of the casing,
A system has been proposed in which the thermal stress generated in the rotor is calculated from the measured values, and the turbine operation is controlled based on the rate of change and absolute value of this thermal stress to prevent severe thermal stress from being generated in the rotor. JP-B 55-21169, JP-B 58-25842
issue).

However, this system prevents the turbine rotor from being subjected to severe stress and prevents the life of the rotor from becoming shorter than the designed service life, and the remaining life cannot be known. Especially, in the past, it is impossible to evaluate the remaining life of a plant which has been operated harshly.

Next, "SAFER", which adopts linear fracture mechanics as a residual life reserve code of a steam turbine rotor, is known (Research report of Central Electric Power Company, Research report No. 283021, December 1983).

Further, as a nondestructive aging deterioration diagnosis technique, an example in which an electrochemical method is applied to a steam turbine has been reported (Electric field technology, Vol. 23, No. 261).

All of these conventional techniques have a certain degree of reliability under certain conditions, but under other conditions, they may take a binary value or lack reliability.

[Object of the Invention]

It is an object of the present invention to provide a residual life evaluation method and a device therefor capable of nondestructively and highly accurately determining the residual life of a mechanical structure subjected to a repeated load or a fluctuating load.

[Outline of Invention]

The present invention, under the repeated load of the creep region in a high temperature atmosphere, based on a new finding that a minute crack occurs on the structure surface at the beginning of life, and the progress of this crack and the life are closely related, It is characterized in that the remaining life is judged from the maximum length of cracks generated on the surface of the structure.

Further, in obtaining the maximum crack length existing on the surface of the structure, the maximum crack length on the sampling surface of a limited area is obtained by statistical processing.

Example of Invention

Hereinafter, the principle of the remaining life evaluation method of the present invention will be described.

Fig. 1 shows the number of repetitions until crack initiation and the number of repetitions until breakage of a smooth test piece of austenitic stainless steel (SUS316) used for main steam pipes of steam turbines in a 650 ° C atmosphere. Is. In FIG. 1, the vertical axis is the strain range, and the horizontal axis is the number of repetitions until breakage. In this case, the strain rate ε is 10 ⁻³ / sec, and the strain application method, that is, the load waveform, is a triangular waveform that monotonically increases at the strain rate and monotonously decreases when a predetermined strain range Δε is reached. In this figure, the white square mark indicates the number of repetitions when the maximum crack length on the surface of the test piece reached 0.05 mm, and the black square mark indicates the number of repetitions when the test piece broke. .

In other words, when the strain range is 1%, a crack of 0.05 mm occurs on the surface after N _f has been repeated about 200 times, and fracture has occurred at 1000 times N _f .

As is clear from this figure, no matter what value the strain range Δε has, a crack of 0.05 mm always occurs at a repetition rate of 10 to 20% of the repetition rate leading to fracture. In other words, 0.
If a minute crack of about 05 mm occurs, if the cyclic load is applied as it is, the number of repetitions of the minute crack is 5
It shows that fracture occurs at a repetition number of 10 times.

The development of minute cracks is a major factor in the life of the test piece until it breaks, and it is considered that the accuracy of the remaining life evaluation will be greatly improved if the crack growth process can be quantitatively grasped.

Therefore, the progress of microcracks is shown in FIG. 2 when the strain range Δε is shown as a parameter. The test piece used is SU
S316, strain rate ε was 2 × 10 ^-3 / sec, and the load waveform was triangular.

FIG. 2 is experimental data showing the progress of microcracks used as a parameter for evaluating the remaining life according to the present invention.

The vertical axis of FIG. 2 represents the crack length generated on the surface of the test piece on a logarithmic scale, and the horizontal axis represents the number of repetitions N.
As is clear from this figure, the progress process in any strain range can be linearly approximated in the figure by taking a logarithmic scale on the vertical axis. That is, it is apparent from FIG. 2 that the logarithm of the crack length and the number of repetitions can be approximated to a direct relationship.

By the way, the life in a high temperature atmosphere largely depends on the strain rate and the load waveform. Therefore, next, with respect to various load waveforms, the crack progress situation of the aforementioned microcracks will be described with reference to FIG. In this figure, the load load waveform is FF, SS-S, S
-F, F-S, Hold390sec is shown, but
The meaning of each waveform is as follows.

(I) Strain is applied at FF ... = 2 × 10 ⁻³ / sec and oppositely given at the same speed (ii) S−S ... = Strain is added at 5 × 10 ⁻³ / sec and the same speed is applied (Iii) S−F ... Strain is applied at + 5 × 10 ⁻³ / sec in the plus direction and 2 × 10 ⁻⁵ / sec in the opposite direction.

(Iv) F-S ... Reverse of (iii) above (v) Hold 390sec ...
Hold for 0 sec and apply the same strain in the opposite direction. As shown in the results in Fig. 3, Hold390 sec, which is generally considered to cause significant creep damage, and most of the life until fracture even with a sawtooth load waveform. Is due to the development of minute cracks. Even in the damage state in which creep and fatigue are superimposed in this way, if the number of cycles N reaching the life is, for example, 1500 between the logarithm log2a of the crack length and the number of cycles N, most of the life, that is, the number of cycles is about 200. Approximately linearly approximated between ~ 1500 times, the following equation holds.

log2a = C · N (1) Therefore, the crack growth rate da / dN is Is led.

If the initial crack length is 0.05 mm in the crystal grain size and the final crack length is about 10 mm in the fracture crack of the test piece in the equation (2) for the microcrack propagation law, Therefore, Can be asked.

By the way, since the crack initiation life N ₀ (the number of repetitions until a crack of 50 μm occurs) is extremely early in the fracture life,
Ignore this and set N ₀ = 0. This is a rational assumption because the mechanical structural member has machining scratches and cracks of several μm to several tens μm before use due to surface roughness.

By the way, if equation (4) is substituted into equation (1), Becomes That is, if the crack length is known, the life ratio N / N _f ,
That is, it indicates that the remaining life can be estimated. But,
If the effect of C'in the equation (5) is large, a large error may occur in the estimation of the remaining life.

Therefore, it is necessary to confirm by experiment whether or not the above-mentioned influence (error) of C ′ on the remaining life estimation. The experimental results are shown in FIG. FIG. 4 shows the relationship between the crack length and the life ratio obtained by actually measuring the progress of microcracks under various conditions such as strain range and type of load waveform. The meaning of each mark in this figure is the same as in FIG. 3, and the distortion range of each case is shown in parentheses.
In FIG. 4, RT shows that the strain was given by the load waveform F-F at room temperature, and Precrept (10 kg / mm ² , 100 hr) is
This means that the test piece was one that caused creep under the conditions in parentheses.

Despite the type of loading conditions, the relationship between the log2a and N / N _f was found that a linear relationship is established in the range of variation of Factor of 2.

That is, the experimental data shown in FIG.
This proves that within the range of the variation of 2, there is almost no influence of the above C '.

This new finding is the basis of the remaining life evaluation method of the present invention. By the way, the crack length in FIG. 4 is a main crack, that is, a crack propagation behavior that causes a test piece as a mechanical structural material used for a mechanical structure to break. But,
Many small cracks occur on the surface of the actual structural member,
They repeatedly grow or coalesce to grow into large cracks, which eventually lead to fracture. Therefore, since the number of cracks until the structural member breaks changes, it is necessary to grasp the change in the number of cracks in the process of reaching the life.

Experimental data on the change in the number of cracks will be described with reference to FIG. FIG. 5 shows a change in the number of cracks generated in a certain area (5 × 5 mm ² ) on the surface of the test piece. From this figure, each crack independently grows and grows up to 400 repetitions, but thereafter, the number of cracks in a certain area tends to decrease due to coalescence growth of the cracks. The behavior of these microcrack groups directly corresponds to the actual damage evolution. Therefore, statistically treating these minute crack groups,
In order to evaluate the remaining life from the relationship between the crack length and the life ratio shown in the figure, a method of estimating the maximum crack length of the distributed cracks will be introduced.

Before that, in order to confirm that the distribution of the crack length can be handled mathematically and statistically, the results of Weibull plotting of the crack length are shown in FIG. The vertical axis of FIG. 6 indicates the cumulative frequency F.
(2a), the horizontal axis is the crack length. This figure shows that the cumulative frequency is about 92% when the crack length is 0.5 mm and the number of cycles N is 502.
The number of cracks having a crack length of 0.5 mm or less is about 92% of the total number of cracks. If the number of repetitions N becomes large, the one having a long crack length, for example, 1.0 to
It is confirmed that there is a higher probability that 6.0 mm cracks are present more than when the number of repetitions N is small. This means that, as described with reference to FIG. 5, as the number of repetitions N increases, the cracks coalesce and the length of one crack becomes longer, and the total number of cracks becomes smaller.

As is clear from these facts, the crack length distribution can be approximated by the Weibull distribution. Therefore, it is confirmed that the distribution of crack length can be handled mathematically and statistically.

By the way, the strength of a mechanical structure is determined by the length of the largest distributed crack. Therefore, if the largest one of the detected microcracks is defined as the main crack, the size of the main crack is considered to be a parameter indicating the degree of damage at that time. Therefore, when the damage degree of an actual device is obtained, an extreme value statistical method for predicting the maximum crack length in the evaluation target region from the crack distribution existing in the sampling surface of the device is used.

That is, the distribution of the maximum value follows the double exponential Gumbel distribution represented by the following equation.

Here, x is the extreme value variable, the maximum crack length in one sample λ: Position parameter α: Scale parameter λ, α is determined by a linear unbiased estimator from the sample data using the extreme value calculation method. Can be used as Alternatively, a method using probability paper can be considered as the graphic calculation method. As an example, a method using probability paper is shown in FIG. FIG. 7 is a diagram for explaining a method for estimating the maximum size of microcracks on the surface to be evaluated. The white circles in the figure represent the maximum size of the microcracks within the sampling area. Since the number of sampling points (locations) in this case is 7, there are 7 circles. Then, the cumulative frequency F _y (y) is plotted on the vertical axis, and the crack length 2a is plotted on the horizontal axis on an arithmetic scale to obtain an extreme value statistical paper. It is possible to confirm that the relationship between the crack length and the cumulative frequency can be linearly approximated by plotting the above-mentioned circles on this extreme value statistical paper and using these plotted data. The vertical axis on the right side is used when predicting the maximum crack length in the target area in the reproduction period T. In this case, T is given by the ratio of the evaluation target area of the actually evaluated device to the area of one sampling, that is, (evaluation target area) / (area of one sampling). Then, the maximum crack length of the microcracks in the evaluation target area is calculated using the previously calculated approximation line and the reproduction time T. In the case of FIG. 7, since the reproduction period T = 6, the crack length is about 0.95 mm.

An example of comparison between the predicted value of the maximum crack length thus obtained and the actual measured value is shown in FIG. In FIG. 8, the black circles represent the life ratio obtained based on the predicted value of the maximum crack length, and the open squares represent the life ratio obtained from the maximum crack length detected from the entire evaluation surface. The predicted value and the measured value are in good agreement.

Therefore, it is demonstrated from this figure that the method of estimating the maximum crack length by applying the extreme value statistical method to the microcrack distribution is highly accurate, and from this predicted value, the crack length and the life ratio The remaining life can be obtained from the relationship.

Although the length of the fracture crack has been described as 10 mm, depending on the device, even a larger size may withstand use. Therefore, it is advisable to preset the crack length, which is the limit of use, depending on the equipment to be applied.

As described above, the life expectancy evaluation method shown in the present embodiment firstly samples the crack distribution at a plurality of locations from the evaluation target area (area) of the mechanical structure, and from this crack distribution, the maximum crack of the microcracks is sampled. The length is detected, and the maximum crack length of the microcracks in the total surface area of the mechanical structure is estimated from the maximum crack length and the sampling area by using the extreme value statistical method as described in FIG. Then, the remaining life is evaluated by using the maximum crack length and the relationship characteristic between the crack length and the life ratio described in FIG.

Furthermore, the material, the material strength, the repetition number N _f up to the limit _cracking, i.e. advance since rupture life are different, the test piece of the same material, determine the N _f in a certain strain range, life shown in FIG. 4 If the characteristics of the ratio N / N _f and the crack length are obtained, the life ratio at that time can be obtained from the maximum crack length.

Next, an apparatus for carrying out the evaluation method described above will be described. Here, the casing of the steam turbine was selected for evaluation.

FIG. 9 shows a high pressure turbine casing of a steam turbine. It comprises an upper casing 1 and a lower casing 2 as shown. The high-temperature and high-pressure main steam 5 flows axially in the casing through the control valve chamber 3 and enters a reheater (not shown) as high-pressure exhaust 6, where it is heated again to high temperature and high pressure and then reheated. Enter the medium pressure stage as steam 7. After passing through the intermediate pressure stage, most of the steam is discharged to the low pressure stage as medium pressure exhaust and the rest is discharged as bleed air 9 to the outside of the casing.

Such a casing is a mechanical structure that is exposed to high-temperature and high-pressure steam, and when starting and stopping, a large thermal stress is generated due to the rapid temperature difference between the member thicknesses caused by the high-temperature steam fluid. Therefore, the casing simultaneously suffers fatigue damage due to repeated thermal stress due to start and stop and creep damage due to long-term load of high internal pressure at high temperature.

In the present invention, the degree of damage and the remaining life of the member are evaluated based on these data by detecting microscopic damages that occur on the surface of the member of the high-temperature equipment that is subject to creep and fatigue damage. In embodying the present invention in a steam turbine casing such as that described above, first, a region that is considered to be most damaged is set to limit an inspection target portion. This can be determined by calculating the maximum stress generation position by stress analysis of the entire casing or by past results. When the target area is set, some points of damage are randomly sampled in the area. Here, as an example, the inlet portion of the main steam pipe, which is one of the overstress generation positions of the steam turbine casing, is taken up. Figure 10 shows the main steam pipe inlet. In the figure, 12 is the main steam pipe 11 and casing 1.
Since it is a welded part that connects the parts to each other and is exposed to the high-speed steam 5, thermal stress based on the temperature difference in the member thickness direction is repeatedly applied at each start and stop. In particular, it is considered that the material 13 and the structure 13 near the welded portion 12 become discontinuous portions and an excessive load is repeated, and the damages due to creep and fatigue are larger than those of other portions. A block diagram of the basic configuration when the device of the present invention is applied to such a part is shown in FIG.
Shown in Figure 11. In the figure, 14 is a detector for detecting a minute crack distribution of several μm to several mm existing on the surface to be evaluated, which has a function of randomly sampling within a certain area of the object to be evaluated. . Figure 12 shows an example of the distribution of microcracks. This is an example of observing the surface of a member that has undergone creep fatigue damage at 650 ° C in the early life. 15 finds the crack length of the extreme value from the distribution of one microcrack existing in the input sampling, and from the extreme value of some sampling, the extreme value statistical method It is an instrument that performs a calculation to estimate the maximum crack length existing in the area to be evaluated. As such calculation software, the parameters of the extreme value statistical distribution are estimated using the linear unbiased estimator method. In addition, this operation unit obtains observed sampling data by performing X ² tests and Kolmogorov-Smirnov tests by statistics on the sampling data. Furthermore, the instrument performs a goodness-of-fit test to determine whether this data can be regarded as a sample from a certain distribution, and determines the number of data and the appropriateness of sampling. If it is inappropriate, the detector 14
It also has a function to give feedback so that sampling can be performed. If it can be confirmed that the number of data and sampling are appropriate, the estimated value is used as a judgment unit for evaluating the degree of damage.
The remaining life is calculated and is output on the display device 17. As the output value, the estimated maximum crack length, its dispersion value, and the remaining life based on it are displayed.

FIG. 13 has a magnifying lens system for the main steam pipe weld
This is an example in which a TV camera is used as a detection unit. Detector 18
Is a high-sensitivity image sensor that has a magnifying lens 19 that can be replaced by 20 to 100 times, and detects the input field of view and converts it into a video signal.
A small industrial television camera 21 having 20 (for example, a solid-state image sensor) is installed in the center of the inside. Illumination light sources 22 are provided on both sides of the image sensor 20 to facilitate observation of minute cracks in the target portion, and the surface of the target portion is illuminated at an angle of about 45 ° through a mirror 22a below the detection unit head. To do. The input video signal from the detection unit 18 enters the image analysis system 23 and undergoes image processing. A certain threshold value is set for the detected input image as shown in FIG. 12 and a gray image of the image is processed to form a binary image 24. After that, edge detection processing is performed and the lengths of all the microcracks on the image are obtained and digitized 25. The digitized data 26 is statistically processed by an arithmetic unit, for example, a general personal computer 27, and the maximum crack length existing in the evaluation target part is obtained from several observation samplings to evaluate the remaining life based on the value. Perform CRT2
Output the result with 8.

As shown in FIG. 14, the detector of the present invention can be miniaturized by using a TV camera solid-state image pickup device 18 with a magnifying lens 19. Further, the processing unit of the input signal from the detection unit 29 may have a built-in microprocessor group,
A compact remaining life evaluation device 30 provided with the display device 31 can be configured. By doing so, the operator can sample the damage of the detection unit 18 at an arbitrary position and can operate the display unit 31 while monitoring the input original image of the detection unit 18.

As another embodiment, the one shown in FIG. 15 has a magnetic particle flaw detector for detecting a fine crack distribution with high resolution 14th.
It is a combination with the device shown. This is because, when the target portion is magnetized, if there is a defect on the surface, the magnetic flux is blocked by the defect, the magnetic flux leaks to the surface where the defect exists, and a small magnetic pole is formed at that portion. If you apply fine iron powder to this part, the iron powder will be attracted to the magnetic poles,
Agglomeration pattern due to iron powder is formed on the defect. In order to make this aggregate pattern easy to see, iron powder sprinkled with a fluorescent dye is used, and a method of irradiating with ultraviolet rays and inspecting is used.

A fluorescent magnetic powder solution is applied to an object to magnetize the object part by the electrodes 34 and 33, and the detection part 18 is provided on the electrode part, and a fluorescent light source is used as a light source for irradiating the object part.
According to this embodiment, there is an effect that it becomes easier to observe the distribution of microcracks.

〔The invention's effect〕

As described above, according to the present invention, the remaining life is evaluated based on the microscopic damage that occurs from the beginning of the life of the mechanical structure member, so that the accident of the mechanical structure can be prevented and the reliability of the plant can be improved. There is an effect that it can be guaranteed.

[Brief description of drawings]

FIG. 1 is a characteristic diagram showing the number of repetitions until breakage of the test piece and the number of repetitions until microcracks are generated on the surface, FIGS. 2 and 3 are characteristic diagrams showing the progress of cracks, and FIG. A characteristic diagram showing the relationship between the crack length and the life ratio, FIG. 5 is a characteristic diagram showing the relationship between the number of surface cracks and the number of repetitions, and FIGS. 6 and 7 are characteristic diagrams showing the relationship between the cumulative frequency and the crack length. FIG. 8 is a characteristic diagram of crack length and life ratio, FIG. 9 is a sectional view of a turbine casing, FIG. 10 is a detailed view of a main steam pipe portion, and FIG. 11 is an apparatus implementing the present invention. Block diagram, FIG. 12 is a diagram showing a crack distribution on the surface to be inspected, FIG. 13 is a diagram showing a schematic configuration of the device of the present invention, FIGS. 14 and 15 are other embodiments of the present invention. It is a schematic diagram showing. 18 …… Detector, 20 …… Image sensor, 23 …… Image analysis system, 27 …… Calculator, 28 …… Display device.

 ─────────────────────────────────────────────────── ─── Continuation of the front page (72) Inventor Saburo Usami 502 Jinritsu-cho, Tsuchiura-shi, Ibaraki Machinery Research Institute Ltd. (72) Hiro Miyata 502 Kintate-cho, Tsuchiura-shi, Ibaraki Nitate Manufacturing Co., Ltd. Inside the Mechanical Research Laboratory (72) Inventor Toritani Hajime 3-1-1, Saiwaicho, Hitachi City, Ibaraki Hitachi Ltd. Hitachi Factory (72) Inventor Kuniyoshi Tsubouchi 502 Kandachicho, Tsuchiura-shi, Ibaraki Hitate Co., Ltd. Machinery Research Laboratory (72) Inventor Ryoichi Kaneko 3-1-1 Sachimachi, Hitachi City, Ibaraki Prefecture Hitachi Ltd. Hitachi Factory (56) References JP 57-151841 (JP, A)

Claims (2)

[Claims] 1. A method for evaluating the remaining life of a mechanical structure subjected to a cyclic load, comprising: (a) the number of repetitions of a load until a critical crack occurs in the mechanical structure material used for the mechanical structure and the mechanical structure material; The relationship between the life ratio, which is the ratio of the number of times the load is received, and the maximum crack length in the microcrack group that occurs in the mechanical structure material is obtained in advance, and (b) from a plurality of sampling surfaces of the mechanical structure. Each of the maximum crack lengths is obtained, (c) the maximum crack length on the surface of the mechanical structure is estimated from the sampled maximum crack length using a statistical method, and (d) the estimated maximum crack length is estimated. A method for evaluating the remaining life of a mechanical structure subjected to cyclic loading, wherein the remaining life is calculated from the previously obtained relationship between the maximum crack length and the life ratio using the maximum crack length. 2. An apparatus for evaluating a remaining life of a mechanical structure subjected to a repeated load, comprising: (a) selecting a plurality of sampling surfaces on the surface of the mechanical structure from among a group of microcracks generated on the sampling surface of the mechanical structure. A detector for detecting the maximum crack length, and (b) a calculator for estimating the maximum crack length of the microcracks on the surface using a statistical method from the maximum crack length detected by the detector, (C) Using the maximum crack length estimated by the calculator,
The maximum crack length in the microcrack group that occurs in the mechanical structural material used for the mechanical structure, the number of repetitions of the load until a critical crack occurs in the mechanical structural material, and the repetition of the load that the mechanical structural material receives And a display device for displaying the determination result of the determination means, the determination means determining the remaining life of the mechanical structure based on the relationship characteristic with the life ratio, which is a ratio of the number to the number. Remaining life assessment system for machine structures subjected to repeated loads.