Hydraulic fracturing is usually discussed as an engineering design problem: fluid, proppant, pump rate, stages. My doctoral research started from a different observation. In tight gas reservoirs in the Southern North Sea, wells with similar fracture designs delivered very different results, and the difference tracked the rock, not the job.
What heterogeneity does to a fracture
A fracture designed for a uniform reservoir behaves differently when it meets high-permeability streaks and natural fractures. Fluid leaks off into the streaks, the fracture grows where the rock lets it rather than where the design intended, and the well's long-term deliverability depends on what the fracture connects to. Net-pressure behaviour during the treatment and the well test afterwards both carry this signature.
The research combined the two: fracture net-pressure modelling during the job and well-test interpretation after it, joined into a single measure of how much heterogeneity changed the outcome. It was published in the Journal of Petroleum Science and Engineering and extended in Energy & Fuels and Energies.
From physics to money
The step that matters for an investor is the last one. When the heterogeneity is carried through production forecasting and discounted cash flow, the result is not a single net present value but a distribution. In the field case in the Energies paper, the computed risk of commercial failure was about one in five, a risk that a single expected value would not show.
A fracturing programme should be sanctioned on its probability of losing money, not only on its expected value.
Why it still matters
The same logic applies wherever a standard completion is repeated across variable rock: tight gas, unconventional plays, geothermal stimulation and, increasingly, natural hydrogen. Price the heterogeneity before the campaign, and design the first wells to measure it. I presented this work at the UK Parliament in 2015, and it remains the foundation of how I approach any decision where the average hides the risk.
The papers are listed on the knowledge map.