Optimisation

Upstream optimisation problems are inherently non-linear. For example, gas-lift-gas allocation can not be done using a liner optimiser as the downstream pressure changes from injecting gas-lift-gas are not taken into account. The result is over injection of lift gas for less oil.

GAP is unique in industry and has a powerful non-linear optimiser that is capable of gas-lift-gas optimisation, ESP frequency optimisation, etc., and is capable of handling a variety of wells in the same network such as:

Naturally flowing oil wells
Gas-lifted wells
ESP operated wells
Condensate or gas producers
Water producers
Water or gas injectors
PCP wells
HSP wells

The Optimiser controls production rates using wellhead chokes, ESP operating frequencies or allocating lift gas to maximise hydrocarbon production while honouring constraints at the gathering system, well and reservoir levels.

GAP's powerful optimisation engine can, for example, allocate gas for gas lifted wells, alter the frequency of ESP pumps or set wellhead chokes for naturally flowing wells to maximize Revenue or Oil Production while honouring constraints at any level of the system.

GAP can also simultaneously model and optimise water and gas injection networks associated with the production system.

Solver

The "natural" response of the system may not be optimum. The network may be capable of higher production rates by altering certain conditions such as:

Back pressure of wells by applying wellhead chokes
Gas-lift-gas allocated to individual gas lifted wells
Frequency of operation of ESP fitted wells
Frequency of operation of pumps / compressors
Inline choke sizes

GAP's non-linear optimiser can do this task including simultaneously honouring system constraints such as:

Maximum oil / gas production
System constraints (min water production, max pressure constraint, velocity constraints, etc.)
Maximum H₂S, CO₂, ⁰API, etc.

GAP will also highlight the limiting elements with a green border after a system has been solved in optimisation mode thus immediately identifying system bottlenecks.

Limiting elements are also tracked during a prediction and can be identified and investigated easily by reloading a prediction snap-shot.