مانا بترو
PR-01أساسيةالموجة 1الذكاء الاصطناعي

التنبؤ بالإنتاج

في جملة واحدة

Forecasts what each well will produce next month, and tells you which wells it is not confident about.

المشكلة

Monthly production plans are built by extrapolating last month by hand across several hundred wells. There is no separation between a well that is declining and a well that was simply shut in for nine days, so plan-versus-actual variance is argued about rather than explained.

كيف تعمل

Well-level models combine decline behaviour, operating conditions — wellhead pressure, choke, lift setting — and downtime patterns, with an ensemble that falls back to a simple decline fit where the data is thin. The forecast is delivered with a per-well confidence class, and low-confidence wells are listed explicitly rather than buried in the field total.

البيانات المطلوبة

  • Daily or monthly production by well with downtime recorded separately from rate
  • Wellhead pressure and choke setting history
  • Well event log: workovers, lift changes, shut-ins with reasons
  • Well test data with test dates and durations

ما الذي تحصل عليه

  • Well, reservoir and field forecast for the next 1–12 months with confidence class
  • Ranked list of wells whose behaviour has changed since last month
  • Plan-versus-actual variance report with the cause attributed to rate, downtime or allocation

ما لا تستطيع فعله

The forecast assumes operating conditions continue. It does not anticipate an unplanned facility shutdown, a compressor trip or a pipeline constraint, and those are the events that dominate large monthly misses. Where the field has poor downtime recording, forecast quality is capped by that and we report the cap instead of a headline accuracy number.

كل صفحة قدرة تذكر حداً حقيقياً. إن لم يستطع المورّد إخبارك أين يتوقف منتجه، فلم يُقل لك الحق عن أين يبدأ.

قدرات ذات صلة