Temperature measurement system and abnormality detection method

1. A temperature measurement system comprising:

an optical fiber;

a temperature distribution measurement apparatus configured to detect backscattered light by causing light to enter the optical fiber, and acquire a temperature distribution of the optical fiber in a length direction of the optical fiber based on a result of the detection; and

a data processing apparatus configured to store the temperature distribution acquired by the temperature distribution measurement apparatus in the data processing apparatus, perform signal processing on a difference temperature distribution obtained by computing a difference between a current temperature distribution and a past temperature distribution, and determine whether or not abnormality is present based on a result of the signal processing, wherein

when determining that abnormality is present based on the result of the signal processing, the data processing apparatus sets an abnormality region and reference temperature regions before and after the abnormality region along the length direction of the optical fiber, the abnormality region being a region where abnormality is considered to have occurred based on the result of the signal processing, the reference temperature regions being regions where no abnormality is considered to have occurred based on the result of the signal processing, and

the data processing apparatus then performs signal processing which corrects a peak value in the abnormality region by using a temperature difference between the current temperature distribution and the past temperature distribution in the reference temperature regions as a reference.

2. The temperature measurement system according to claim 1 , wherein when performing the signal processing which corrects the peak value, the data processing apparatus uses a non-negative constraint condition or a non-positive constraint condition derived from a result of integration of the difference temperature distribution.

3. The temperature measurement system according to claim 2 , wherein the data processing apparatus puts out an alert when a peak value of the difference temperature distribution obtained by the correction by the signal processing which corrects the peak value exceeds a preset threshold.

4. The temperature measurement system according to claim 1 , wherein the data processing apparatus determines whether or not abnormality is present by taking into consideration normal periodic temperature change.

5. The temperature measurement system according to claim 1 , wherein the signal processing includes:

performing a differentiation process on the difference temperature distribution.

6. The temperature measurement system according to claim 1 , wherein the signal processing includes:

acquiring a differentiated temperature distribution by performing a differentiation process on the difference temperature distribution, and

applying an FIR (Finite Impulse Response) filter to the differentiated temperature distribution.

7. The temperature measurement system according to claim 2 , wherein the signal processing which corrects the peak value is super-resolution signal processing including

correcting the difference temperature distribution sequentially a plurality of times such that a squared error between convolution of the difference temperature distribution and a transfer function, and the difference temperature distribution decreases at each correction, and

each time the correction is performed, replacing difference temperatures in the reference temperature regions after the correction with an estimated difference temperature such that the constraint condition is met.

8. The temperature measurement system according to claim 1 , wherein data on the past temperature distribution is a sum of weighted pieces of data on a plurality of past temperature distributions.

9. The temperature measurement system according to claim 1 , wherein the optical fiber is laid along a pipe.

10. The temperature measurement system according to claim 1 , wherein the optical fiber is laid along electronic equipment.

11. The temperature measurement system according to claim 1 , wherein the optical fiber is laid around a plant.

12. An abnormality detection method comprising:

by using a temperature distribution measurement apparatus, acquiring backscattered light by causing light to enter an optical fiber from one end of the optical fiber; and

by using a data processing apparatus, storing an intensity distribution of the backscattered light acquired by the temperature distribution measurement apparatus in the data processing apparatus, performing signal processing on a difference temperature distribution obtained by computing a difference between a current temperature distribution and a past temperature distribution, and determining whether or not abnormality is present based on a result of the signal processing, wherein

when determining that abnormality is present based on the result of the signal processing, the data processing apparatus sets an abnormality region and reference temperature regions before and after the abnormality region along a length direction of the optical fiber, the abnormality region being a region where abnormality is considered to have occurred based on the result of the signal processing, the reference temperature regions being regions where no abnormality is considered to have occurred based on the result of the signal processing, and

the data processing apparatus then performs signal processing which corrects a peak value in the abnormality region by using a temperature difference between the current temperature distribution and the past temperature distribution in the reference temperature regions as a reference.

13. The abnormality detection method according to claim 12 , wherein when performing the signal processing which corrects the peak value, the data processing apparatus uses a non-negative constraint condition or a non-positive constraint condition derived from a result of integration of the difference temperature distribution.

14. The abnormality detection method according to claim 13 , wherein the data processing apparatus puts out an alert when a peak value of the difference temperature distribution obtained by the correction by the signal processing which corrects the peak value exceeds a preset threshold.

15. The abnormality detection method according to claim 12 , wherein the signal processing includes:

performing a differentiation process on the difference temperature distribution.

16. The abnormality detection method according to claim 12 , wherein the signal processing includes:

acquiring a differentiated temperature distribution by performing a differentiation process on the difference temperature distribution, and

applying an FIR (Finite Impulse Response) filter to the differentiated temperature distribution.

17. The abnormality detection method according to claim 13 , wherein the signal processing which corrects the peak value is super-resolution signal processing including

correcting the difference temperature distribution sequentially a plurality of times such that a squared error between convolution of the difference temperature distribution and a transfer function, and the difference temperature distribution decreases at each correction, and

each time the correction is performed, replacing difference temperatures in the reference temperature regions after the correction with an estimated difference temperature such that the constraint condition is met.

18. The abnormality detection method according to claim 12 , wherein data on the past temperature distribution is a sum of weighted pieces of data on a plurality of past temperature distributions.