• In this post, we will close out our discussion on the detection of abnormal loss and gain. As you can see from Slide 1, the network lays quite a bit of emphasis on the modeled gain/loss rate (which is based on the filtered mud volume discussed in the last two posts), the flow out trend and driller memos. Slide 2 provides details on node characteristics that influence the abnormal gain belief. The network in Slide 5 similarly shows the major factors that influence the abnormal loss beliefs.

    Note that in Sentinel RT, we track influxes during the connections separately. The logic is quite straightforward – track connection gains for a set number of previous connections to detect anomalies in the current connection influx. Slide 4 shows an example where the connection influx belief spikes due to the connection flowback being sufficiently (pre-defined criteria) different from those in previous connections. The parameters for what constitute an anomaly can be adjusted as needed.

    Slide 6 shows an example of abnormal loss detection. You can see how the filtered mud volume filters out external mud removals and addition. The rate of change of the filtered mud volume channel along with loss in flow out, cause the abnormal loss belief to go up.

    To summarize, there are a multitude of models and approaches to detecting kicks and losses, and almost all of them track mud volume and flow out. We hope we have provided you with some additional ideas, such as how to filter the noise from the mud volume channel and how to use driller memos for more context. It must be said that driller memos are today an under-utilized input to drilling advisory software, but that is poised to change in the coming years.

    Click below for slides on this topic:

    Abnormal Loss Gain Network

    That is all on this topic for now. Next week, we will start discussing software tools to support directional drilling. So, as always, stay tuned.