The energy sector is undergoing a digital transformation, and modern Artificial Intelligence has evolved into a critical asset for upstream operations. From optimizing reservoir characterization to mitigating drilling risks, AI acts as a force multiplier for technical decision-making.
These prompts are rigorously tested and optimized for deployment across major Large Language Models (LLMs), including ChatGPT, Gemini, Claude, and DeepSeek. While specific models demonstrate unique strengths—such as DeepSeek’s aptitude for logic or Claude’s handling of large technical contexts—the following library provides a universal foundation for Petroleum Engineers seeking to enhance efficiency and precision in both office and field environments.
1. Optimizing History Matching Strategy
Best for: DeepSeek or Claude (Due to high reasoning capabilities for complex logical workflows).
History matching is notoriously iterative and time-consuming. Use this prompt to generate a structured sensitivity analysis plan based on mismatched parameters.
Act as a Senior Reservoir Simulation Engineer. I am performing a history match on a [Sandstone/Carbonate] reservoir with a [Water/Gas] drive mechanism. The simulation model is currently matching pressure data well but is significantly underestimating water cut in the breakdown phase for Wells [X, Y, and Z].
Based on relative permeability curves and aquifer influx parameters, propose a step-by-step sensitivity analysis strategy to identify the parameters responsible for the water breakthrough delay. Rank the parameters by likelihood of impact and suggest specific adjustments for the relative permeability endpoints (Krw, Kro) and exponent values.
The Payoff: Drastically reduces the trial-and-error phase by providing a prioritized list of tuning parameters tailored to specific mismatch symptoms.
2. Designing Drilling Fluid Programs
Best for: ChatGPT (Versatile for synthesizing chemical properties and operational constraints).
Drafting mud programs requires balancing inhibition, rheology, and cost. This prompt helps outline a robust fluid system for challenging formations.
Act as a Drilling Fluids Specialist. I need a preliminary mud program for a deviated well section (angle [X] degrees) passing through reactive shale formations prone to swelling and sloughing. The estimated bottom hole temperature is [X] degrees Celsius.
Recommend a Water-Based Mud (WBM) formulation that maximizes inhibition and hole cleaning. List the key additives (inhibitors, viscosifiers, filtration control agents) with recommended concentration ranges. Include a brief risk assessment regarding differential sticking and suggested mitigations.
The Payoff: Generates a comprehensive starting point for mud system design, ensuring all critical chemical interactions and borehole stability risks are addressed early.
3. Automating Decline Curve Analysis (DCA) Interpretation
Best for: DeepSeek (Strong performance in mathematical logic and pattern recognition).
While AI cannot replace standard DCA software, it can validate your logic and suggest hyperbolic vs. exponential fits based on production narratives.
Act as a Production Engineer. I have production data for a fractured tight oil well. The data shows an initial high decline rate followed by a transition to boundary-dominated flow.
Analyze the theoretical implications of using a modified Arps Hyperbolic decline versus a Stretched Exponential Production Decline (SEPD) model for this flow regime. Provide the mathematical justification for which model likely yields a more conservative Estimated Ultimate Recovery (EUR) and explain how the 'b-factor' should be constrained to avoid overestimation in long-term forecasts.
The Payoff: Provides a theoretical “second opinion” on reserve estimation methodologies, helping prevent the common pitfall of overestimating EUR in unconventional plays.
4. Troubleshooting Electrical Submersible Pump (ESP) Failures
Best for: Claude (Excellent for processing detailed technical context and diagnosing complex systems).
Rapid diagnosis of downhole failures restores production faster. This prompt acts as a diagnostic checklist.
Act as an Artificial Lift Engineer. An ESP in a high-GOR well has tripped on 'Underload'. The intake pressure chart shows a cyclical "sawtooth" pattern prior to the trip. The drive frequency is fixed at [X] Hz.
Diagnose the most likely cause of this failure (e.g., gas locking, pump-off, or broken shaft) based on the amperage and pressure signature described. Outline a troubleshooting procedure to confirm the root cause and suggest 3 specific parameter changes to the Variable Speed Drive (VSD) to stabilize the well upon restart.
The Payoff: Accelerates root cause analysis for artificial lift failures, minimizing deferment and preventing repeat trips.
5. Generating PVT EOS Regression Strategies
Best for: Gemini (Strong at handling multi-variable data concepts and synthesis).
Tuning an Equation of State (EOS) is critical for accurate simulation. This prompt helps structure the regression approach.
Act as a PVT Specialist. I need to tune a Peng-Robinson Equation of State model for a volatile oil sample. The lab data includes CCE (Constant Composition Expansion) and DLE (Differential Liberation) experiments.
Outline a regression strategy. Specify which component properties (Tc, Pc, Omega, or Volume Shift) should be grouped and regressed for the heavy fractions (C7+) versus the light ends. Explain the order of operations to match saturation pressure first, followed by liquid density and GOR, while maintaining thermodynamic consistency.
The Payoff: formalized the EOS tuning workflow, ensuring that simulation fluid models are physically consistent and computationally stable.
6. Writing HSE Risk Assessments for Perforating Ops
Best for: Claude (Nuanced understanding of safety protocols and professional documentation).
Safety documentation is non-negotiable. This prompt creates a rigorous Job Safety Analysis (JSA) foundation.
Act as a Drilling & Completions HSE Manager. Draft a specific Risk Assessment and Job Safety Analysis (JSA) for a Wireline Conveyed Perforating operation on a high-pressure offshore rig.
Focus on three critical hazards: 1) Unintentional detonation due to radio silence failure, 2) Pressure control equipment (PCE) failure during run-in-hole, and 3) Dropped objects during gun assembly. For each, provide the 'Potential Consequence', 'Initial Risk Rating', and specific 'Control Measures' required by API RP 67 standards.
The Payoff: Produces a high-compliance safety document draft that aligns with industry standards, saving hours of administrative work.
7. Enhanced Oil Recovery (EOR) Screening
Best for: ChatGPT (Good for broad knowledge retrieval and comparative analysis).
Screening reservoirs for EOR suitability requires matching field characteristics with recovery methods.
Act as a Reservoir Engineer. I am screening a mature oil field for Enhanced Oil Recovery. The reservoir parameters are: Depth [X] ft, Permeability [X] mD, Oil Viscosity [X] cP, and Temperature [X] °C. The formation water salinity is high ([X] ppm).
Evaluate the technical feasibility of Polymer Flooding versus CO2 Miscible Flooding for these specific parameters. Create a comparison table highlighting the pros, cons, and fatal flaws for each method considering the high salinity and temperature constraints.
The Payoff: Quickly filters out non-viable recovery methods, allowing engineers to focus feasibility studies on the most promising EOR techniques.
8. Interpreting Open-Hole Log Anomalies
Best for: DeepSeek or Gemini (Capable of cross-referencing geological logic).
When log data is ambiguous, AI can offer geological hypotheses.
Act as a Petrophysicist. I am analyzing open-hole logs in a shaly-sand sequence. I observe a 'crossover' effect where the Density porosity is significantly higher than Neutron porosity, but the Gamma Ray is also high (indicating shale).
Propose three geological or mineralogical explanations for this specific log signature (e.g., presence of heavy minerals, gas effect masking, or specific clay types like chlorite vs. illite). Recommend one additional logging tool or core analysis test to confirm the lithology.
The Payoff: Assists in identifying complex lithologies or tool artifacts that standard interpretation models might misinterpret as “gas crossover” or “bad hole.”
9. Economic Sensitivity Analysis for Field Development
Best for: Gemini (Effective at structuring variables for financial modeling).
Project economics depend on volatile inputs. This prompt structures the CAPEX/OPEX stress test.
Act as a Petroleum Economist. We are evaluating a greenfield offshore development project with a CAPEX of $[X] million.
Draft a structure for a Spider Diagram sensitivity analysis. Identify the top 5 variables (e.g., Oil Price, Drilling Days per Well, Facility Uptime) that typically have the highest impact on NPV and IRR. For each variable, define a reasonable uncertainty range (+/- %) based on current industry volatility, and explain how these sensitivities should be presented to investment stakeholders.
The Payoff: ensures that economic models account for the most critical financial drivers, leading to more robust Final Investment Decisions (FID).
10. Casing Design Burst & Collapse Verification
Best for: DeepSeek (Precision with engineering constraints and load cases).
Validating casing design logic ensures well integrity under worst-case scenarios.
Act as a Drilling Engineer. I am verifying the design for a 9-5/8" Intermediate Casing string set at [X] ft TVD.
List the mandatory load cases that must be calculated for both Burst and Collapse to meet API 5C3 standards. Specifically, detail the assumptions for the "Lost Circulation with Gas Kick" load case regarding internal and external pressure profiles. Provide the formula concept for calculating the resultant differential pressure at the casing shoe.
The Payoff: Acts as a technical safeguard, ensuring no critical load cases are overlooked during the casing design phase.
Pro-Tip: Context Injection for Technical Accuracy
To get the most out of these models, practice Context Injection. Never simply ask “How do I fix a stuck pipe?” Instead, inject the specific wellbore geometry, BHA (Bottom Hole Assembly) configuration, and the sequence of events leading to the incident.
Example: “I am stuck at [Depth]. The BHA includes [List Tools]. We were [Sliding/Rotating] when torque spiked. Current circulation pressure is [Pressure]. Based on this, determines the sticking mechanism.”
Providing this raw data allows the AI to function less like a search engine and more like a senior technical consultant sitting in the doghouse with you.
The integration of AI into Petroleum Engineering is not about replacing engineering judgment; it is about augmenting it. By utilizing these prompts, you streamline the computational and administrative heavy lifting, freeing up mental bandwidth for the complex decision-making that defines high-value reservoir management. Start implementing these workflows today to build a more agile and data-driven operation.