Q: How is an artificial lift strategy managed in the Petro.ai DSU Design Service (DSUDS) model?
A: Industry observers estimate that 90% to 95% of the world’s producing wells require the use of an artificial lift system, the process of increasing downhole pressure to encourage more production to the surface. The horizontal, multi-stage development of unconventional wells is a unique challenge for operators who embrace the pressure challenge in very different ways.“ In low pressure regions, our clients may choose to be more aggressive and ramp up the well as quickly as possible,” commenting on the use of artificial lift, Kyle LaMotta explains the impact on the DSUDS model.
“Artificial lift is certainly a challenging part of the equation,” LaMotta continues. “There isn’t a way to predict artificial lift because it’s a decision that people make and there’s really no predictors of how the wells are going to flow back and what artificial lift to use. Nor is there data on artificial lift systems publicly available.
“Given the impact of various choke management and artificial lift strategies on Arps decline parameters, we need to use something more consistent when we’re creating the DSUDS models.
“So, when we’re training the models, we’ll calculate 12-month cumulative production (12-mo cum.) as the value that we’re trying to predict. At this point in a well’s life, the impact of choke management, flowback strategy, and artificial lift selection has flattened out.
“If we chose a 24-month time span, we’d be losing two years' worth of wells. Using a 12-month cum. is right there in that sweet spot where a lot of the early life things have been taken care of but it’s not so far out into the future that we’re losing a lot of wells.
“Importantly, 12 months is not so far into the future that we filter out too many new wells.”
“In terms of making the predictions, let’s say we’re designing a 4 well DSU, we can control the ramp up period in DSUDS. Since we’re predicting a 12-mo cum., we can hang a variety of Arps fits off this control point, effectively controlling for our customer’s early life strategy. If the client tells us they’re going to ramp the wells up over 60 days, then we can build in that ramp up, keep the 12-mo cum. the same but change the shape of the rate versus time curve.
“In making hundreds of projections across every shale basin in North America, the 12-mo cum. is the best balance of factors. When you look at either rate versus time or cum. production versus time, you’ll see deviation in the early life that would throw off other machine-learning powered predictions.
“The DSUDS process generates blind predictions of 12-mo cums. orders of magnitude more accurate than type curves, which gives us a great deal of confidence in time series and cash flow forecasts. Once the well stabilizes, the DSUDS prediction rate curve paints a line through actual measurements time and time again.”