| |
Energy
Applications
Integrated energy companies have a wealth of data
related to exploration, producing, trading, transportation, refining, and
marketing. A few E&P applications include:
-
Seismic Pattern Recognition: Predictive models can be used to detect
seismic events, pick first breaks, filter spikes, etc.
-
Drilling: During drilling predictive models can be used to detect
conditions in which the drill pipe might get stuck and determine the likelihood of
un-sticking it. Predictive models can also be used to monitor and
control drilling mud properties.
-
Well Log Analysis: Predictive models can integrate data from
wireline logs with core samples to predict key reservoir properties, such as
permeability, where core samples are unavailable.
-
Decline Curve Analysis: It can be a time consuming and error-prone
process to eye-ball the production curves of every well in an active field.
Predictive models can analyze production histories to forecast future decline
rates and detect failing equipment.
-
Pump Card Diagnosis: A dynamometer pump card is generated by
monitoring the pump load string through its full cycle. The shape of the
card is useful for diagnosing various problems with the pump. Predictive
models can be used to automate pump card diagnosis in real-time.
|