Use Cases
No continuous well rate measurement
It is important for hydrocarbon accounting process to have continuous well rate estimation. However, most of the time there are only periodic physical measurements of wells.
Continuous well rate measurement with Neuronet
Wells are equipped with several sensors such as pressure, temperature transmitters, choke valve opening indicators, downhole gauges with data transmission to surface.
Data from all these sensors can be used for building virtual flow metering models, in our case data driven models based on deep learning. After training and validation, the model is used as virtual flow meter. Such concept will help to obtain high degree of redundancy for your MPFM and test separator fleet.
Anomaly detection
It is important for hydrocarbon accounting process to have continuous well rate estimation. However, most of the time there are only periodic physical measurements of wells.
Estimation of per well production between tests is usually done by use of simple correlations such as PQ curves, linear regressions, etc.
But such correlations cannot handle transient behavior. What can be done?Measurement management “from-well-to-export”
So far, we have mentioned isolated problem in production measurement, which is well production measurement. However, the overall issue is broader, which encompasses fiscal measurements, measurements in processing plant and wells, so called “from-well-to-export” measurement management.
