Analyzer, analysis method, analyzer program, and analysis learning device

1. An analyzer that analyzes a gaseous measurement sample on a basis of spectrum data obtained by irradiating the gaseous measurement sample with continuous light having a plurality of wave numbers, the analyzer comprising:

a total analysis value calculation part that, on a basis of a total analysis value of a gaseous reference sample of which the total analysis value of multiple predetermined components are preliminarily obtained, calculates a total analysis value of the multiple components in the gaseous measurement sample from the spectrum data of the gaseous measurement sample, wherein the total analysis value of the gaseous reference sample is a quantitate property of the gaseous reference sample and wherein the total analysis value calculation part includes

a correlation data storage part that stores data indicating a correlation, obtained using machine learning, between spectrum data of the gaseous reference sample and the total analysis value of the gaseous reference sample, wherein each spectrum of the spectrum data is a single spectrum containing contributions of all components in the gaseous reference sample, and

a calculation main body part that applies the correlation in the correlation data storage part to the spectrum data of the gaseous measurement sample to calculate the total analysis value of the multiple components in the gaseous measurement sample.

2. The analyzer according to claim 1 , wherein

the total analysis value calculation part further comprises

a correlation calculation part that, on a basis of the spectrum data of the gaseous reference sample and the total analysis value of the gaseous reference sample, calculates the correlation between the spectrum data and the total analysis value to store it in the correlation data storage part.

3. The analyzer according to claim 2 , further comprising

a reception part that receives data indicating the total analysis value of the gaseous reference sample, the total analysis value being measured by one other type of analyzer, wherein

the correlation calculation part is one that calculates the correlation on a basis of the total analysis value received by the reception part and the spectrum data of the gaseous reference sample, the spectrum data being obtained by the analyzer.

4. The analyzer according to claim 2 , wherein

the correlation calculation part is one that calculates the correlation from respective total analysis values and respective pieces of spectrum data of multiple gaseous reference samples, and when a total analysis value of a new reference sample and its spectrum data are acquired, takes into account the total analysis value of the new reference sample and its spectrum data to update the correlation.

5. The analyzer according to claim 1 , wherein

the correlation is calculated with temperature and/or pressure of the gaseous reference sample as parameters, and

the calculation main body part calculates the total analysis value of the gaseous measurement sample on a basis of the correlation specified with temperature and/or pressure of the gaseous measurement sample as parameters.

6. The analyzer according to claim 5 , wherein

the correlation is calculated with surrounding situation data related to an engine or a catalyst as a parameter, and

the calculation main body part calculates the total analysis value of the gaseous measurement sample on a basis of the correlation specified with the surrounding situation data related to the engine or the catalyst as a parameter.

7. The analyzer according to claim 1 , wherein

the spectrum data is one in which an influence of another component interfering with the multiple components is reduced or removed.

8. The analyzer according to claim 1 , further comprising

a main analysis part that analyzes an individual single component contained in the gaseous measurement sample on a basis of the spectrum data obtained by irradiating the gaseous measurement sample with the light.

9. The analyzer according to claim 1 , wherein

the gaseous measurement sample or the gaseous reference sample is exhaust gas of a vehicle, and the components to be analyzed are hydrocarbons.

10. The analyzer according to claim 9 , wherein

the total analysis value of the multiple predetermined components is concentration of total hydrocarbons contained in the exhaust gas.

11. The analyzer according to claim 1 ,

the analyzer being one of an FTIR type.

12. The analyzer according to claim 11 , wherein

the one other type of analyzer is an FID analyzer.

13. An analysis method that analyzes a gaseous measurement sample on a basis of spectrum data obtained by irradiating the gaseous measurement sample with light, the analysis method comprising:

on a basis of a total analysis value of a gaseous reference sample of which the total analysis value of multiple predetermined components are preliminarily obtained, calculating a total analysis value of the multiple components in the gaseous measurement sample from the spectrum data of the gaseous measurement sample, wherein the total analysis value of the gaseous reference sample is a quantitate property of the gaseous reference sample;

storing data indicating a correlation, obtained using machine learning, between spectrum data of the gaseous reference sample and the total analysis value of the gaseous reference sample, wherein each spectrum of the spectrum data is a single spectrum containing contributions of all components in the gaseous reference sample; and

applying the correlation to the spectrum data of the gaseous measurement sample to calculate the total analysis value of the multiple components in the gaseous measurement sample.

14. The analysis method of claim 13 , further comprising

when a total analysis value of a new gaseous reference sample and its spectrum data are acquired, taking into account the total analysis value of the new gaseous reference sample and its spectrum data to update the correlation.

15. An analyzer that analyzes a gaseous measurement sample on a basis of spectrum data obtained by irradiating the gaseous measurement sample with continuous light having a plurality of wave numbers, the analyzer comprising:

a total analysis value calculation part that, on a basis of a total analysis value of a gaseous reference sample of which the total analysis value of multiple predetermined components are preliminarily obtained, calculates a total analysis value of the multiple components in the gaseous measurement sample from the spectrum data of the gaseous measurement sample, wherein the total analysis value of the gaseous reference sample is a quantitate property of the gaseous reference sample and wherein the total analysis value calculation part includes

a correlation data storage part that stores data indicating a correlation, obtained using machine learning, between spectrum data of the gaseous reference sample and the total analysis value of the gaseous reference sample, wherein each spectrum of the spectrum data is a single spectrum containing contributions of all components in the gaseous reference sample,

a calculation main body part that applies the correlation in the correlation data storage part to the spectrum data of the gaseous measurement sample to calculate the total analysis value of the multiple components in the gaseous measurement sample, and

a correlation calculation part that, on a basis of the spectrum data of the gaseous reference sample and the total analysis value of the gaseous reference sample, calculates the correlation from respective total analysis values and respective pieces of spectrum data of multiple gaseous reference samples to store it in the correlation data storage part, and when a total analysis value of a new gaseous reference sample and its spectrum data are acquired, takes into account the total analysis value of the new gaseous reference sample and its spectrum data to update the correlation.