| Title | Estimation of drilling risk of non-Artesian wells with the Monte Carlo method |
|---|---|
| Authors | Pall Valdimarsson, Miao Yu, Caixia Sun |
| Year | 2023 |
| Conference | World Geothermal Congress |
| Keywords | Risk Mitigation, Economics and Financing, Drilling and Completion Technology, Simulation, Monte Carlo |
| Abstract | The drilling of a geothermal well has no guaranteed outcome. The drilling risk is taken in this paper as the risk of investing in a well which will not be economical for the system which it is intended to supply. This paper is also only treating non-Artesian geothermal wells, where there is a water level down in the well in an undisturbed condition. The parameters which are unknown prior to drilling the well are the drilling cost, the water temperature, the depth down to the undisturbed water level in the well, and the productivity index of the formation. The productivity index of the formation is understood to be the flow which can be taken from the well for each unit of drawdown of the well water level from the undisturbed level. A life cycle cost model is made for the well, treating the well cost, the water temperature, the undisturbed water level, and the productivity index as stochastic variables with a given density function. A simple triangular density function is defined for each of the stochastic variables, having zero density below the minimum boundary (“worst case”) and above the maximum boundary (“best case”), with a maximum density value in between (“most likely”). The density curve is linear between these points. These points can usually be defined by heuristics, which means in reality the guess of the good and qualified specialist. When a set of values has been drawn from the stochastic variables, the cost of the well pump can be estimated and the NPV of the pump electricity consumption. Each set of values will give some water production from the well which is of value to the downstream system. This value of the product is also defined by the market, so assumptions on how the water will be used and the income therefrom have to be made. The final risk estimation is then based on creating many (thousand) sets of values from the stochastic variables and present the probability density function of the score value, which is conveniently taken as the ration between lifetime income (NPV) and the well cost (including pump and NPV of electricity cost). The final result is then an estimate of the probability of that the well is economic for any given case. |