System And Method For Quantifying Uncertainty Of Predicted Petroleum Fluid Properties

A method, comprising:

• receiving optical spectra data into a processor, the optical spectra data being representative of optical characteristics of a formation fluid;
• predicting, via the processor, a parameter corresponding to one component of multiple components of the formation fluid based on the optical spectra data;
• and
• determining, via the processor, an uncertainty in the predicted parameter corresponding to the one component of the multiple components of the formation fluid based on the optical spectra data.

A method, comprising:

• receiving optical spectra data into a processor, the optical spectra data being representative of optical characteristics of a formation fluid flowing through a downhole formation fluid sampling tool;
• predicting, via the processor, a parameter corresponding to one or more components of the formation fluid based on the optical spectra data and a model derived from a calibration dataset;
• determining, via the processor, a measurement uncertainty in the received optical spectra data;
• identifying, via the processor, a model uncertainty in the model derived from the calibration dataset;
• and
• determining, via the processor, an uncertainty in the predicted parameter based on the measurement uncertainty and the model uncertainty.