historymatching

History matching is the process of adjusting a mathematical model of a petroleum reservoir so that its simulated history reproduces observed field data. It is used to align model predictions with measurements such as production rates, cumulative production, bottom-hole pressures, water cut, gas-oil ratio, and, increasingly, time-lapse seismic responses.

In history matching, the reservoir model is treated as an inverse problem. Parameters subject to calibration include permeability fields and distributions, porosity, relative permeability curves, fluid properties, initial conditions, and sometimes features like fault geometry or well-bottom-hole locations. The objective is to minimize the misfit between observed data and model predictions, often while respecting geological plausibility and uncertainty.

The typical workflow involves building a geological and numerical model, running a reservoir simulator to produce predictions, comparing these predictions with historical data, and iteratively updating parameters to reduce discrepancies. The process may use a formal objective function and may incorporate weights for different data types or times. History matching can be performed manually, by deterministic optimization, or by stochastic and data-assimilation approaches.

Common methods include gradient-based optimization with adjoint models, and stochastic techniques such as ensemble Kalman filters, ensemble smoothers, or other Bayesian updating schemes that generate multiple plausible realizations. The goal is not only to fit past data but also to quantify and propagate uncertainty into future forecasts.

Applications include improved production forecasts, reserve estimation, and strategic field development planning. Limitations include non-uniqueness of solutions, high computational cost, data noise, and the risk of overfitting, which can lead to overly optimistic or biased projections.