مانا بترو
SR-01متقدمةالموجة 3الذكاء الاصطناعيالمجال

محاكاة المكمن بالنموذج البديل الذكي

في جملة واحدة

A learned stand-in for the full reservoir simulator that answers in seconds instead of hours, so an engineer can test fifty scenarios instead of three.

المشكلة

A full-physics run on a multi-million-cell model takes six to thirty hours. Optimisation, uncertainty quantification and history matching all need hundreds of runs, so in practice they are replaced by a handful of hand-picked cases and engineering judgement.

كيف تعمل

A neural surrogate is trained on a designed ensemble of full-physics runs, learning the mapping from control and property inputs to pressure and saturation response. It is used only inside the training envelope, and every proxy answer that drives a decision is confirmed with one full-physics run before it leaves the team.

البيانات المطلوبة

  • A validated full-physics simulation deck for the field
  • Compute budget for 200–500 training runs of that deck
  • Historical production and injection by well
  • Property model realisations covering the intended uncertainty range

ما الذي تحصل عليه

  • Sub-second forecast of field and well response to a control scenario
  • Uncertainty envelope across the property ensemble
  • Ranked scenario table for the reservoir management meeting

ما لا تستطيع فعله

A proxy cannot extrapolate. Outside the range of controls and properties it was trained on it fails quietly rather than loudly, which is the dangerous failure mode, so the system refuses queries outside the training hull instead of answering them. It also needs an existing trustworthy simulation deck — it does not replace building one.

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

قدرات ذات صلة

  • SR-02متقدمة

    المطابقة التاريخية الآلية

    Tunes the reservoir model until it reproduces what the field actually produced, automatically and with the uncertainty kept rather than discarded.

    ذكاء تحت السطح والمكامنالذكاء الاصطناعيالمجال
  • SR-11متوسطة

    التنبؤ مع تحديد عدم اليقين

    Produces P10, P50 and P90 production forecasts whose spread is calculated, not chosen to look reasonable.

    ذكاء تحت السطح والمكامنالذكاء الاصطناعي
  • SR-10متقدمة

    تحديد مواقع آبار التطوير

    Searches for where the next wells should go, with drilling feasibility and surface constraints inside the search rather than applied afterwards.

    ذكاء تحت السطح والمكامنالذكاء الاصطناعيالمجال
  • SR-05متقدمة

    النمذجة التوليدية بين الآبار

    Fills the rock between wells with geologically plausible detail instead of a smooth interpolation that no outcrop ever looked like.

    ذكاء تحت السطح والمكامنالذكاء الاصطناعي