Uncertainty and Risk in Energy Systems

Figure 1. Predicted mean surface heat flow in the Appalachian Basin of New York, Pennsylvania, and West Virginia.

The large-scale deployments envisioned for most new energy projects are also likely to have large uncertainties. Thus an important part of any energy planning process, whether at demonstration or at commercial scale, is to anticipate and represent the uncertainty in the quality and characteristics of the resource (e.g. surface heat flow, Figure 1 and Figure 2), the cost of resource extraction, possible changes and improvements in technologies for extraction and production of energy, and the value of the product in the marketplace reflecting future market prices. Investors will be concerned about all of these factors.

 

Figure 2. Variability in surface heat flow along cross section A-A’ in Figure 1. Where error lines do not overlap vertically, there is a significant difference in predicted mean surface heat flow.

Public concerns may relate to the risk of induced earthquakes, subsidence, vibrations, diminishing access to sources of water, regular or accidental discharge of fluids to water bodies or the  groundwater, air emissions, or potential accidents, at a minimum. Although no project is without risks, the public will want to be well informed about the potential risks, their likelihood, and what is being done to address potential problems. How the public perceives risks can be important to the outcome of a project.

Ongoing efforts in the Energy Institute attempt to be sensitive to the many uncertainties of concern to investors and the public. Current efforts in geothermal energy are directed at prediction of thermal field properties (e.g. heat flow, temperature-at-depth, depth-to-temperature) in the Appalachian Basin of New York, Pennsylvania, and West Virginia using bottomhole temperatures (BHTs) from oil and gas well-log data. Data quality is assessed using spatial outlier detection methods on the thermal field properties calculated from BHT data at each well. Uncertainty in the thermal field properties at each well is determined by cross-examination of reliable wells within the basin for which detailed rock stratigraphy is known. The thermal field is predicted using a kriging algorithm that provides the predicted mean (Figure 1) and the spatial standard error of predicted mean. Representation of the mean resource and the error about the mean in a single figure is accomplished by use of a cross section (Figure 2), which allows a decision maker to examine the relative certainty of the thermal field in locations that have favorable resources.

Faculty Involved: Lindsay Anderson, Mircea Grigoriu, Patrick Reed, Gennady Samorodnitsky, Jery Stedinger, Jeff Tester