Formatted Title
Data Management, Visualization, and Predictive Analytics Tools to Streamline Decision-Making and Optimize Performance on Steam-Enhanced Extraction Sites
Background/Objectives
Data management and visualization tools are essential for effectively analyzing and making decisions on in situ thermal remediation projects. These tools allow for the organization and management of large quantities of data, as well as the ability to easily view and interpret those data. This has historically enabled professionals to identify patterns and trends and communicate informed decisions to field operators about how to make changes in the field in order to maximize overall performance of the system. However, while the use of these types of data visualization dashboards is ubiquitous throughout the thermal remediation industry, senior technical team members can oftentimes become a bottleneck in the process of evaluating performance data, determining optimization strategies, and communicating those strategies in actionable ways to field personnel. This bottleneck creates an opportunity to optimize the data review and decision-making process for faster and better performance. Recent innovations in the field of predictive analytics can be applied to the active thermal remediation system operational data to better and automatically determine optimization strategies. This can allow the field operators themselves to quickly evaluate the data and make operational changes to enhance performance without the delay in waiting for senior technical team members.
Approach/Activities
At a U.S. government facility in southern California, steam-enhanced extraction (SEE) was implemented at full scale at three separate bulk, field-constructed underground storage tank (BFCUST) sites where kerosene-based jet fuels were historically stored. At each of the sites, modelling software packages that are normally used for three-dimensional representation of geology and contaminant distribution were instead repurposed to visually demonstrate three-dimensional heating progress to identify areas requiring heating optimization. Additionally, software packages that have typically been used for business intelligence purposes were successfully repurposed to analyze, interpret, and visualize the vast quantity of data associated with thermal remediation projects in meaningful ways to the operators. These data not only included the readings from the network of over 300 subsurface temperature sensors, automated data collected from treatment system PLC logs, but also electronically collected field data from system operator’s readings, such as steam injection flow rates. Three-dimensional heating front modelling was then used in conjunction with predictive analytics to evaluate the data for the need for operational changes to maximize system performance.
Results/Lessons Learned
The approach of combining a three-dimensional representation of the heating front with predictive analytics allowed operators in the field to make informed and real-time decisions about how to best modify the system for maximum effectiveness. The predictive analytics module automatically determined which steam injection wells needed to be adjusted based on the proximity of those wells to the areas of the subsurface which were furthest behind the predicted heat up curve. This provided operators in the field with a useful tool to quickly prioritize which steam injection wells could be adjusted for the most benefit. Additionally, the operators had access to the automatically generated three-dimensional heating model to visually confirm which areas of the site were furthest behind the expected performance. Overall, this both changed and optimized the decision-making process from one that involved a remote senior technical lead analyzing weekly data and providing delayed instructions to the field operators to one that put the tools for decision-making into the hands of the operators themselves. The flexibility and speed in optimization that this process created resulted in faster than expected heat up of the site in most areas, because recalcitrant areas received a quick, focused, and effective operational response.