A Statistical Investigation on Risk and Economy of Geothermal Well Drillings

Abstract

This paper presents a statistical investigation of drilling success in the exploration, development and operation phases of a geothermal power project. Drilling “success rate” is shown to be an awkward concept; drilling success may better be represented by the average cost per MW capacity secured from a drilling program. There exist several operational fields in which drilling success rate are lower than 50 %. The case histories of the Kamojang field in Indonesia and a project at The Geysers field in California are used to illustrate some of the concepts. It is shown that the average drilling success rate improves as more wells are drilled due both to its inherent statistical nature, which becomes better defined with an increasing sample size, as well as due to the “learning curve” effect. For the two case histories presented, the learning curve effect was found to be relatively minor compared to the statistical effect. A basic calculation on probability of well success rate with drillings of ten or less wells also shows that “learning curve” effect is relatively small. According to such probability calculations, expected power capacity MW per unit drilling cost is approximately same for three to five well drillings, assuming at least two successful wells. On the other hand the risk of no-drilling-success becomes as low as 3% if five wells are drilled with a drilling success rate for each well of 50%. Therefore it may be reasonable to plan at least five wells to be drilled to evaluate economic feasibility of a geothermal field development.