Systems and methods for cement evaluation

1. A method comprising:

receiving acoustic cement evaluation into a data processing system, wherein the acoustic cement evaluation data derives from one or more acoustic downhole tools used over a depth interval in a well having a casing, wherein the acoustic cement evaluation data comprises flexural attenuation and acoustic impedance, wherein the acoustic cement evaluation data has a first parameterization;

correcting at least a portion of the entire acoustic cement evaluation data to account for errors in the first parameterization using the data processing system, thereby obtaining corrected acoustic cement evaluation data, wherein the correction is estimated based on a relationship between the flexural attenuation and the acoustic impedance;

plotting acoustic cement data points, wherein each acoustic data point indicates a value of received acoustic impedance and flexural attenuation measured at a same depth,

processing the corrected acoustic cement evaluation data with an initial solid-liquid-gas model using the data processing system, wherein the initial solid-liquid-gas model includes one or more threshold ranges for identifying data points corresponding to liquid, solid or gas;

performing a posteriori refinement of the initial solid-liquid-gas model in the data processing system, thereby obtaining a refined solid-liquid-gas model, wherein the refined solid-liquid-gas model includes one or more refined threshold ranges for identifying data points corresponding to liquid, solid or gas; and

generating a well log track that indicates whether a material behind the casing is a solid, liquid, or gas by processing the corrected acoustic cement evaluation data using the refined solid-liquid-gas model in data processing system.

2. The method of claim 1 , wherein correcting the acoustic cement evaluation data comprises:

analyzing a subset of the acoustic cement evaluation data, wherein the subset of the acoustic cement evaluation data comprises at least some data points of the acoustic cement evaluation data points, the data points of the subset being beneath an evanescence point;

estimating a correction to the acoustic data that causes at least the subset of the acoustic cement evaluation data to more closely match expected nominal values; and

applying the correction to at least the portion of the entire acoustic cement evaluation data.

3. The method of claim 2 , wherein applying the correction comprises re-parameterizing at least the portion of the entire acoustic cement evaluation data to account for the errors in the first parameterization.

4. The method of claim 2 , wherein applying the correction comprises applying an offset to at least the portion of the entire dataset of the acoustic data to cause at least the subset of the acoustic cement evaluation data to more closely match the expected nominal values.

5. The method of claim 1 , wherein the initial solid-liquid-gas model with which the corrected acoustic cement evaluation data is processed comprises:

a first solid-liquid-gas model comprising a first gas threshold range, a first liquid threshold range, and a first solid threshold range;

a tight solid-liquid-gas model comprising a second gas threshold range, a second liquid threshold range, and a second solid threshold range, at least one of which is tighter than the corresponding first threshold ranges of the first solid-liquid-gas model; or

a solid-liquid-gas model that considers flexural attenuation values of the acoustic cement evaluation data only when a pulse-echo-derived acoustic impedance of the acoustic cement evaluation data is below an evanescence point; or

any combination thereof.

6. The method of claim 5 , wherein the second gas threshold range is not directly adjacent to the second liquid threshold range.

7. The method of claim 5 , comprising generating the tight solid-liquid-gas model, wherein the tight solid-liquid-gas model is generated at least in part by:

reducing noise properties propagated through the first solid-liquid-gas model; or

reducing an uncertainty value of a parameter used by the first solid-liquid-gas model, wherein the parameter comprises a well fluid density, a fluid velocity (VP), a well fluid acoustic impedance, or a thickness of the casing, or any combination thereof.

8. The method of claim 1 , wherein performing the posteriori refinement comprises:

overlaying a density distribution of at least some of the acoustic cement evaluation data onto a map of the solid-liquid-gas model; and

geographically refining solid-liquid-gas threshold boundaries of the map of the initial solid-liquid-gas models to determine the refined solid-liquid-gas model.

9. The method of claim 8 , wherein the solid-liquid-gas threshold boundaries are geographically refined using a polygon approach, a polynomial approach, or both a polygon and polynomial approach.

10. The method of claim 1 , wherein performing the posteriori refinement comprises:

overlaying a density distribution of at least some of the acoustic cement evaluation data onto a map of the initial solid-liquid-gas model; and

applying a statistical analysis of the acoustic cement evaluation data points to determine the refined solid-liquid-gas model.