Production Data Analysis of Shale Gas Wells with Abrupt Gas Rate or Pressure Changes

Abstract

Abstract Production data analysis (PDA) by using rate normalized pressure (RNP) and rate normalized pressure derivative (RNP’) is useful for transient rate and pressure analysis of shale gas wells with constant or smooth changing gas rate and pressure. However, some reasons may cause abrupt changes, fluctuation, or even loss of production data. The existing PDA methods can not well address this kind of issue. The paper analyzes the reasons that cause big changes of shale gas production rate and pressure. The reasons include well-interference, well shut-ins, and converting production from casing to tubing. Typical shale gas well cases in China are described. Three methods are proposed to address the non-smooth production data issue. For shale gas wells with severe well-interference from neighbouring fracturing wells or production wells, the segmented production data before well-interference is suggested as well-interference is like an imposed negative or positive force from outside, and this force disturbs the normal production performance only rely on the well's own energy. For wells with frequent shut-ins, a virtual equivalent time method is referenced. The process for this method firstly calculates the formation pressure distributions and the average formation pressures within the SRV area; sencondly, calculate the virtual equivalent time by use of the average formation pressure; thirdly, divided the whole production data into several interconnected segments by rearrange the vitual euivalent time into the actual time axis, and finally do the analysis by using the log-log plot of pressure and pressure derivate vs material balance time. For shale gas wells with converting production from casing to tubing, as there are abrupt rate and pressure changes at the converting point, the material balance time may be no more monotonically increasing with production time. We proposed the average material balance time method to solve this problem. For this method, we use average material balance time instead of the material balance time in the log-log plot of pressure and pressure derivative. Results shows that severe well-interference cause big disturblance and only data before well-interference is suggested for PDA. Both the PDA with average material balance time and PDA with virtual equivalent time can get much better match of production history and log-log plot of pressure and pressure derivative then the exsiting PDA method.