Challenges in reservoir modeling accuracy

I’ve been diving deep into the intricacies of reservoir modeling lately, and it’s incredible how much variability can arise from our data inputs. For instance, I worked on a field in the North Sea where the porosity measurements were off by nearly 10%, drastically affecting our production forecasts. Has anyone else experienced similar discrepancies in their models, and how do you address them?

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍‌⁠‌‍‍‍​⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠‌‌⁠⁠‌⁠‌​‌‍⁠⁠‌⁠​​‌‍‍‌‌‍​⁠​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​‍​‍‌‍⁠‍‌‍‌‌‌⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‍​⁠​​​⁠‌⁠​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​⁠‍‌‌‍⁠‍‌‍‌​‌‌​​‌⁠​‍‌‌‌⁠‌​‌‌‌⁠‌⁠​⁠‌‍‌​‌⁠‌‌‌⁠‌​‍​‌‍‌‌​⁠‌‌‌‍⁠‍​⁠‌​​‍​‍‌⁠⁠‌​

I can totally relate to the porosity issue — had a similar experience with a field in the Gulf of Mexico where our core samples were inconsistent. What helped us was cross-verifying our data with geophysical logs, which revealed discrepancies we didn’t catch initially. It definitely improved our forecasts, but it takes more time and effort upfront.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍‌⁠‌‍‍‍​⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠‌‌​⁠​​​⁠‌⁠​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‍​⁠​‌​⁠​‌​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍‌‌​‌‌​⁠‍‌​‍‍​⁠‌​‌‌‌​‌‌​‍‌⁠‌‍‌​​‍‌‍​‌‌⁠‌​‌‌​‌‌⁠‌​‌​‍​‌‍‍‌‌‍‍⁠‌‍‍‌​‍​‍‌⁠⁠‌

I’ve faced similar issues with porosity affecting forecasts too, especially in projects where we relied heavily on initial core samples. One thing that helped us immensely was integrating real-time monitoring tools during drilling phases to validate the data on the fly. Sometimes, that extra layer of verification can really save time and resources in the long run.

‌⁠‍⁠​‍​‍‌⁠‌​​‍​‍​⁠‍‍​‍​‍‌‍‌⁠‌‍‌⁠‌‍‍‍​⁠​‍​‍​‍​‍⁠​​‍​‍‌‍‍⁠​‍​‍​⁠‍‍​‍​‍‌⁠​‍‌‍‌‌‌⁠​​‌‍⁠​‌⁠‍‌​‍​‍​‍⁠​​‍​‍‌‍‍‌‌‍‌​​‍​‍​⁠‍‍​⁠‌‌​⁠​​​⁠‌⁠​⁠‍‌​‍⁠​​‍​‍‌‍‌​​‍​‍​⁠‍‍​‍​‍​⁠​‍​⁠​​​⁠​‍​⁠‌‍​⁠​​​⁠​‍​⁠​‌​⁠‌​​‍​‍​‍⁠​​‍​‍‌‍‍​​‍​‍​⁠‍‍​‍​‍​⁠‌​‌‍⁠‍​⁠​​‌⁠‍​​‍⁠‌‌‍​‌‌‌‌⁠‌​‍‍​⁠​​‌​⁠⁠‌⁠​‌​‍⁠‌‌​​‍​⁠​‌‌‌‌‌‌‍‌‍​‍​‍‌⁠⁠‌