US US20180240021A1

A concise summary of U.S. Patent Application Publication No. US20180240021A1, which has been granted as U.S. Patent No. 11,087,221 B2, is provided below.

Title: Well performance classification using artificial intelligence and pattern recognition

Assignee: Saudi Arabian Oil Company

Inventors: Badr M. Al-Harbi, Amell Ali Al-Ghamdi, Ali A. Al-Turki

Filing Date: February 20, 2017

Issue Date: August 10, 2021

Abstract:
A heterogeneous classifier based on actual reservoir and well data is developed to qualitatively classify oil well producer performance, and based on the classification drill a new well into a producing reservoir or adjust fluid flows in an existing well. The data includes perforation interval(s), completion type, and how far or close the perforated zones are located relative to the free water level or gas cap. The data also include geological data, such as major geological bodies like regional faults and fractures. The features may be prioritized before classification. The classifier utilizes four different techniques to apply pattern recognition on reservoir simulation vector data to classify the wells. Three of the classification techniques are supervised learning methods: Bayesian classification, dynamic time warping and Neural Network. The fourth classification is an unsupervised method, clustering, to automate well grouping into similar categories.

Overview of Independent Claims:

This patent has three independent claims, each directed to a different aspect of the invention: a method, a data processing system, and a data storage device.

Independent Claim 1 (Method): This claim outlines a method for deciding whether to drill a new well in a hydrocarbon reservoir. The process begins by feeding proposed well parameters, such as production rates and its physical configuration and location, into a data processing system. The system then creates a "classification model" that predicts fluid production, flow, and pressure. Using this model, a probabilistic estimate of the target well's production rates is generated. If this estimate is deemed acceptable, the well is then physically drilled in the reservoir.

Independent Claim 8 (Data Processing System): This claim describes a data processing system designed to carry out the method of the first claim. It specifies a system with a processor that receives the proposed well parameters. The processor is configured to build the classification model and then calculate the probabilistic estimate of the well's production rates. The system also includes an output display that indicates whether the estimated production rates are acceptable for proceeding with the drilling of the well.

Independent Claim 15 (Data Storage Device): This claim covers a non-transitory computer-readable storage medium (such as a hard drive or flash memory) that contains instructions for a data processing system. When these instructions are executed, they cause the system to perform the same essential steps as the method claim: receiving well parameters, creating a classification model, generating a probabilistic production estimate, and displaying whether the result is acceptable for drilling a new well.

A search of the U.S. Court of Appeals for the Federal Circuit (CAFC) dockets for 2026 did not reveal any cases associated with US20180240021A1 or its granted counterpart, US11087221B2.

No Record of Litigation for US Patent US20180240021A1

As of May 1, 2026, a thorough search of patent litigation databases and federal court records reveals no known litigation involving U.S. Patent Application Publication No. US20180240021A1, or its granted counterpart, U.S. Patent No. 11,087,221 B2.

Searches were conducted using the PACER (Public Access to Court Electronic Records) system for federal district courts, the docket for the U.S. Court of Appeals for the Federal Circuit (CAFC), and the portal of Unified Patents, a leading patent litigation database. None of these searches returned any results indicating that the patent has been asserted in a lawsuit.

Therefore, there are no plaintiffs, defendants, jurisdictions, case numbers, filing dates, or outcomes to report at this time.

Proceedings overview

The USPTO Open Data Portal indicates no AIA trial proceedings on file for US Patent US20180240021A1 or its granted counterpart, US11087221B2. This means that, as of the most recent data ingest, there are no active, concluded, or settled Inter Partes Review (IPR), Post-Grant Review (PGR), or Covered Business Method (CBM) proceedings against this patent. For a defendant, this means the patent's claims remain untested by PTAB challenges.

Strategic summary

As of May 29, 2026, all claims of US20180240021A1 remain UNTESTED by any AIA trial proceedings before the PTAB. There are no claims that have been canceled or sustained in such proceedings.

Since no AIA trials have been initiated, there is no estoppel landscape under 35 U.S.C. § 315(e)(2). Therefore, a potential defendant is not barred from raising any prior art grounds they deem relevant in a future PTAB challenge or district court litigation.

The absence of PTAB activity can be a mixed signal. On one hand, it could suggest that the patent has not yet been widely asserted, or that prior art challenges have not been identified by potential infringers. On the other hand, well-asserted patents often eventually attract IPRs. The lack of challenges might simply mean the patent has not been subjected to significant scrutiny.

Recommended next steps

If facing an assertion of US20180240021A1, a defendant should:

• Conduct a thorough prior art search, building upon the initial prior art identified during prosecution, to identify the strongest invalidity arguments.
• Evaluate the merits of filing an IPR petition, considering the claims' scope and the available prior art, as the patent's claims are currently untested.
• Monitor for any newly filed PTAB petitions against this patent, which would be publicly available through the USPTO's PTAB search tools.

Inventors

• Badr M. Al-Harbi (Saudi Arabian Oil Co.)
• Amell Ali Al-Ghamdi (Saudi Arabian Oil Co.)
• Ali A. Al-Turki (Saudi Arabian Oil Co.)

All inventors were employed by the original assignee, Saudi Arabian Oil Co., at the time of filing. There are no unusual patterns indicating inventors departing the original assignee.

Original assignee

The entity named on the issued patent is Saudi Arabian Oil Co. (also known as Saudi Aramco). Saudi Aramco is the world's largest oil producer and has a primary line of business in exploration, production, refining, distribution, and marketing of petroleum and natural gas. The company ships numerous products embodying the claims, specifically in the context of oil and gas exploration and production. Its current status is operating and publicly traded.

Assignment timeline

The USPTO Assignment Center (https://assignmentcenter.uspto.gov/) has no recorded assignments for US20180240021A1 or its granted counterpart US11087221B2 beyond the original assignment to Saudi Arabian Oil Company, which is noted on the face of the patent.

The Google Patents "Legal Events" section shows an assignment on 2017-02-20 to SAUDI ARABIAN OIL COMPANY with the context "ASSIGNMENT OF ASSIGNORS INTEREST (SEE DOCUMENT FOR DETAILS). Assignors: AL-HARBI, BADR M., AL-GHAMDI, Amell Ali, AL-TURKI, Ali A." This is the initial assignment from the inventors to the original assignee.

Timeline diagram

timeline
    title Ownership of US20180240021A1
    2017 : Filed by Saudi Arabian Oil Co
    2021 : Issued to Saudi Arabian Oil Co

NPE / troll-pattern signals

1. Shell-entity transfer - Not present. The only recorded assignment is from the inventors to Saudi Arabian Oil Company, a large operating company.
2. Known asserter in the chain - Not present. Saudi Arabian Oil Company is not a known patent asserter (NPE).
3. Repeat correspondent across the chain - Not present. There is only one initial assignment from the inventors to the operating company.
4. Cascading transfers - Not present. There is only one initial assignment.
5. Pre-litigation transfer - Not present. No litigation has been identified, and no transfers beyond the initial assignment exist.
6. Bankruptcy fire-sale - Not present. Saudi Arabian Oil Company is an active, publicly traded company.
7. Privateering - Not present. There is no indication of transfer to an NPE for assertion on the operating company's behalf.
8. Defensive aggregator (anti-NPE) - Not present. The patent remains with the original operating company.

Verdict

Insufficient data. There are no recorded assignments in the USPTO Assignment Center beyond the initial assignment from the inventors to the original assignee, Saudi Arabian Oil Co., as shown on the patent itself. This lack of subsequent transfer means there is no evidence to suggest NPE activity or defensive aggregation.

Analysis of Prior Art for US Patent Application US20180240021A1

A review of the citations for U.S. Patent Application Publication No. US20180240021A1 (granted as US11087221B2) reveals several key prior art references that were considered during its examination. These references relate to the use of computational models, simulations, and data analysis for predicting and optimizing hydrocarbon production from reservoirs. Below is an analysis of the most relevant citations and their potential impact on the claims of the '021 application.

Key Prior Art References and Potential Anticipation of Claims:

The core of the invention in US20180240021A1 lies in its method of using a machine-learning-based "classification model" to generate a "probabilistic estimate" of a target well's performance, which then informs the decision to drill the well. The independent claims (1, 8, and 15) broadly cover this method, the system that performs it, and a storage device containing the instructions for it. The analysis below focuses on how prior inventions might anticipate these core concepts.

---

1. US Patent No. 9,043,188 B2 - "System and method for forecasting production from a hydrocarbon reservoir"

• Full Citation: US Patent No. 9,043,188 B2, "System and method for forecasting production from a hydrocarbon reservoir," assigned to Chevron U.S.A. Inc.
• Filing Date: September 1, 2006
• Brief Description: This patent describes a method for forecasting hydrocarbon production by creating a "proxy model" from a limited number of detailed reservoir simulations. This proxy model, which can be a response surface or a neural network, is then used to rapidly estimate production outcomes for various operational scenarios without running full, time-consuming simulations for each case. The system can be used to optimize field development plans.
• Potential Anticipation of Claims:
  • Claim 1 (Method) & Claim 8 (System): The '188 patent discloses a method and system that align closely with the initial steps of the '021 application's claims. It involves receiving well and reservoir parameters ("operational scenarios"), forming a predictive model ("proxy model" or "neural network"), and using it to estimate production rates. The '188 patent's "proxy model" can be seen as analogous to the "classification model" in the '021 application. While the '188 patent focuses on "forecasting" and the '021 application on a "probabilistic estimate" for "classification" (good/bad), the underlying process of using a computationally derived model to predict well performance is substantially similar. The distinction may lie in the specific nature of the output (a probabilistic classification vs. a production forecast), but a strong argument for anticipation or at least obviousness could be made.
  • Claim 15 (Data Storage Device): The teachings of the '188 patent would inherently anticipate a data storage device containing instructions to carry out its described method, thus overlapping with the scope of claim 15.

---

2. US Patent No. 9,910,938 B2 - "Shale gas production forecasting"

• Full Citation: US Patent No. 9,910,938 B2, "Shale gas production forecasting," assigned to Schlumberger Technology Corporation.
• Filing Date: June 20, 2012
• Brief Description: This invention details a system for forecasting shale gas production by generating a large number of "realizations" (possible geological models) and running simulations on them. It then uses techniques like principal component analysis and clustering to analyze the simulation results and identify patterns that correlate with production outcomes. This allows for a probabilistic forecast of production.
• Potential Anticipation of Claims:
  • Claim 1 (Method) & Claim 8 (System): The '938 patent describes a workflow that includes generating simulation results and then using statistical analysis and clustering (an unsupervised learning method mentioned in the '021 application's abstract) to understand and predict well performance. This process of forming a data-driven model from simulation results to make a "probabilistic forecast" is a core element of claims 1 and 8. The use of clustering to group similar outcomes is particularly relevant to the "classification" aspect of the '021 application.
  • Claim 15 (Data Storage Device): Similar to the '188 patent, the system described in the '938 patent would be implemented via software on a storage device, making claim 15 potentially anticipated.

---

3. US Patent No. 10,138,717 B1 - "Predicting well performance with feature similarity"

• Full Citation: US Patent No. 10,138,717 B1, "Predicting well performance with feature similarity," assigned to Novi Labs, LLC.
• Filing Date: January 7, 2014
• Brief Description: This patent discloses a method for predicting the performance of a new, un-drilled well by identifying "analogous" or similar existing wells from a database. It calculates a similarity score based on various features (geological, completion, etc.) between the proposed well and the existing wells. The performance of the most similar existing wells is then used to predict the performance of the new well. This is framed as a machine learning approach.
• Potential Anticipation of Claims:
  • Claim 1 (Method) & Claim 8 (System): This reference is highly relevant. It explicitly describes a machine learning-based system for predicting the performance of a "target well" before it is formed. The process involves receiving parameters for the target well, using a model based on "feature similarity" (analogous to the '021 application's classification model built from training data), and generating a prediction. The '717 patent's approach is a form of pattern recognition, which is a key element of the '021 application. While it may not explicitly use the term "probabilistic estimate" in the same way, the use of multiple analogous wells to generate a prediction inherently contains a probabilistic element. This patent presents a strong challenge to the novelty of the core method and system claims.
  • Claim 15 (Data Storage Device): The '717 patent's method is designed for computer implementation, therefore anticipating the substance of claim 15.

Obviousness Analysis of US Patent Application US20180240021A1 under 35 U.S.C. § 103

This analysis evaluates whether the invention claimed in U.S. Patent Application Publication No. US20180240021A1 (the '021 application) would have been obvious to a Person of Ordinary Skill in the Art (POSITA) at the time the invention was made. The analysis is based on the prior art references identified in the preceding section.

Under 35 U.S.C. § 103, a patent claim is invalid if the differences between the claimed invention and the prior art are such that the subject matter as a whole would have been obvious at the time the invention was made to a person having ordinary skill in the art.

Definition of a Person of Ordinary Skill in the Art (POSITA)

For the purposes of this analysis, a POSITA would be an individual with a degree in petroleum engineering, geological sciences, or a related field, coupled with several years of experience in reservoir engineering. This individual would be familiar with numerical reservoir simulation, conventional well planning techniques, and the application of statistical and computational modeling, including early-stage machine learning techniques, to analyze reservoir data and forecast production.

Analysis of Independent Claims (1, 8, and 15)

The independent claims of the '021 application cover a method, system, and data storage device for the same core process:

1. Receiving proposed parameters for a target well (location, configuration, production rates).
2. Forming a classification model by processing reservoir simulation results, with the model indicating fluid production rates, flows, and pressures.
3. Forming a probabilistic estimate of the target well's production rates using the classification model.
4. Making a decision to form (drill) the well if the estimate is acceptable.

The central inventive concept is the use of a machine learning-based "classification model," trained on simulation data, to generate a "probabilistic estimate" of a new well's performance to guide a drilling decision. The prior art, when combined, suggests this concept would have been obvious.

---

Combination 1: US 10,138,717 B1 ('717 patent) in view of US 9,043,188 B2 ('188 patent)

This combination of prior art would render the claims of the '021 application obvious.

• Primary Reference: US 10,138,717 B1 ('717 patent)
  The '717 patent discloses the core of the claimed method: predicting the performance of a new, un-drilled "target well" using a machine learning model. It teaches receiving the proposed well's features (location, completion, geology) and using a model based on "feature similarity" to find analogous existing wells. The performance of these analogous wells is then used to predict the target well's performance. This directly teaches steps 1, 3, and 4 of the '021 claims. The prediction based on a group of similar wells is inherently a "probabilistic estimate," and the entire purpose is to inform the decision of whether to drill the well.

• Secondary Reference: US 9,043,188 B2 ('188 patent)
  The primary element missing from the '717 patent is the explicit step of building the predictive model from a large set of computerized reservoir simulation results. The '717 patent builds its model from a database of existing wells. The '188 patent directly addresses this by teaching the creation of a "proxy model" (such as a neural network) by running a limited number of detailed, time-consuming reservoir simulations. This proxy model is then used to rapidly forecast production for many different scenarios.

• Motivation to Combine:
  A POSITA, familiar with the '717 patent's approach of using a data-driven model to predict target well performance, would naturally seek the best and most comprehensive data sources to train such a model. While the '717 patent uses historical data from existing wells, a POSITA would recognize the limitations of this approach—namely, that the historical data may not cover the full range of desired operational or geological scenarios for a new well.

  The '188 patent provides a well-known solution to this problem: using detailed numerical simulations to generate a rich dataset that can be used to train a faster "proxy model." A POSITA would have been motivated to combine these teachings for a predictable result. They would replace the historical well data used for training in the '717 patent with the synthetic, but more comprehensive, simulation data taught by the '188 patent. This would allow the "feature similarity" model of the '717 patent to be trained on a wider and more controlled set of data, leading to a more robust predictive tool. This straightforward combination of a known machine learning framework ('717) with a known method for generating training data ('188) arrives directly at the invention claimed in the '021 application.

---

Combination 2: US 9,043,188 B2 ('188 patent) in view of US 9,910,938 B2 ('938 patent)

This combination also provides a strong basis for an obviousness rejection.

• Primary Reference: US 9,043,188 B2 ('188 patent)
  The '188 patent teaches nearly the entire claimed invention. It describes a system that uses a "proxy model" (which can be a neural network, a type of classification model) built from reservoir simulations to forecast production. This covers receiving well parameters (as part of defining an "operational scenario") and forming a predictive model from simulation results to estimate production rates. The goal is to optimize a field development plan, which inherently includes making decisions about which wells to drill.

• Secondary Reference: US 9,910,938 B2 ('938 patent)
  The '188 patent describes its output as a "forecast." The '021 application specifically claims a "probabilistic estimate" and a "classification model." The '938 patent explicitly addresses this by teaching a method to generate a probabilistic forecast of production. It does this by running simulations on many geological "realizations" and then using statistical analysis and clustering to analyze the results. Clustering is a form of classification—it groups wells into categories of similar performance. This directly teaches the "classification" and "probabilistic" aspects of the '021 claims.

• Motivation to Combine:
  A POSITA starting with the proxy model concept from the '188 patent would understand that a single-point forecast has limitations and that a probabilistic output would be more valuable for risk assessment and decision-making. The '938 patent provides a known method for achieving this, teaching the use of statistical analysis and clustering on simulation results to create a probabilistic forecast.

  A POSITA would be motivated to apply the probabilistic and classification techniques from the '938 patent to the proxy model framework of the '188 patent. This would enhance the '188 system by allowing it to not just provide a single production forecast, but to classify the likely outcome of a proposed well (e.g., "good" or "bad," as described in the '021 specification) and provide a probabilistic confidence level. This is a simple application of a known data analysis technique ('938) to a known modeling system ('188) to achieve the predictable improvement of a more robust, risk-aware output. The result of this combination is precisely the system claimed in the '021 application.

Conclusion

The independent claims of US20180240021A1 appear to be obvious over at least two different combinations of the cited prior art. The core concept of using a computer-generated model, trained on reservoir simulation data, to produce a probabilistic estimate of a target well's performance was a known and evolving practice in the field before the '021 application's priority date. The prior art teaches the use of machine learning models for this purpose, the generation of training data from simulations, and the framing of the output in probabilistic and classificatory terms. Combining these known elements to create the claimed system would have been a predictable and logical step for a Person of Ordinary Skill in the Art seeking to improve the efficiency and accuracy of well planning.

Analysis for US Patent US20180240021A1

As of May 1, 2026, the following details pertain to the patent prosecution history and family of U.S. Patent Application Publication No. US20180240021A1, which was granted as U.S. Patent No. 11,087,221 B2.

Patent Term

• Projected Expiration Date: The projected expiration date for U.S. Patent No. 11,087,221 B2 is March 8, 2039.
• Patent Term Adjustment (PTA): The standard 20-year term from its filing date of February 20, 2017, would have resulted in an expiration date of February 20, 2037. The adjusted expiration date of March 8, 2039, indicates a significant patent term adjustment granted by the USPTO, likely to compensate for delays during the patent prosecution process. Such adjustments are provided under 35 U.S.C. § 154(b) to account for administrative delays by the patent office.
• Patent Term Extension (PTE): There is no indication of any Patent Term Extension (PTE) for this patent. PTE is typically associated with delays in regulatory review for products like pharmaceuticals and is not applicable here.

Continuity and Divisional Applications

Based on the provided documentation and a search of the public record, there are no continuation or divisional applications that claim priority to U.S. Application No. 15/437,170 (the application that matured into the '221 patent). The application itself claims priority to its own filing date and does not appear to be a continuation or divisional of a prior U.S. application.

Patent Family

This U.S. patent is part of a larger international patent family, indicating that the assignee, Saudi Arabian Oil Company, sought protection for this invention in multiple jurisdictions. A patent family consists of a set of patent applications filed in various countries that are related to each other through a common priority claim.

The known members of the patent family for US20180240021A1 include:

• United States:

  • Application No.: US15/437,170 (Filing Date: 2017-02-20)
  • Publication No.: US20180240021A1
  • Granted Patent No.: US11087221B2 (Issue Date: 2021-08-10)

• World Intellectual Property Organization (WIPO/PCT):

  • Application No.: PCT/US2018/018098
  • Publication No.: WO2018152147A1

• European Patent Office (EP):

  • Application No.: EP18707562.7A
  • Publication No.: EP3583292A1

• China (CN):

  • Application No.: CN201880012935.8A
  • Publication No.: CN110325706A

• Saudi Arabia (SA):

  • Application No.: SA519402351A
  • Publication No.: SA519402351B1

Defensive Disclosure and Prior Art Generation

Document ID: D-9384-A1
Publication Date: 2026-05-01
Title: Systems and Methods for Predictive Classification Modeling Across Multiple Technical Domains
Subject Matter: This document discloses derivative inventions, alternative embodiments, and cross-domain applications of the core methodology described in US Patent 11,087,221 B2 (based on application US20180240021A1), thereby placing them in the public domain. The core methodology involves: (1) receiving parameters for a target system, (2) forming a classification model from simulation data, (3) generating a probabilistic estimate of the target system's performance, and (4) using the estimate to guide a decision.

---

Derivative Embodiment 1: Component Substitution with Advanced Neural Architectures

Enabling Description: The classification model described in the reference patent can be substantially improved by replacing the Multilayer Perceptron (MLP), Bayesian, and DTW models with a Transformer-based architecture. Time-series data from the reservoir simulator (e.g., pressure, flow rate, water cut at discrete time steps) is treated as a sequence, analogous to words in a sentence. Each time-step's vector of parameters (e.g., [pressure, temp, saturation]) is converted into a high-dimensional embedding vector. Positional encodings are added to these embeddings to retain temporal information. The entire sequence of embedded vectors is then processed by a multi-head self-attention mechanism within a Transformer encoder stack. This allows the model to learn complex, non-linear dependencies between distant time steps in the simulation, providing a more accurate classification of the well's long-term performance (e.g., "Good," "Bad," "Requires Intervention"). The final classification is produced by a linear layer and a softmax function applied to the output of the [CLS] token embedding from the Transformer. This approach is superior for identifying subtle patterns in long-duration simulations that simpler models would miss.

sequenceDiagram
    participant Sim as Reservoir Simulator
    participant Emb as Tokenizer & Embedder
    participant Trans as Transformer Encoder
    participant Classifier as Output Layer

    Sim->>Emb: Generate Time-Series Data (Vectors)
    Emb->>Emb: Convert each vector to an embedding
    Emb->>Emb: Add Positional Encodings
    Emb->>Trans: Pass sequence of encoded vectors
    Trans->>Trans: Apply Multi-Head Self-Attention
    Trans->>Classifier: Pass final hidden state of [CLS] token
    Classifier->>User: Output Probabilistic Classification (Good/Bad)

---

Derivative Embodiment 2: Operational Parameter Expansion to Cryogenic Fluid Sequestration

Enabling Description: The methodology is applied to the geological sequestration of cryogenically stored supercritical fluids, such as liquid nitrogen or captured carbon dioxide, in subterranean salt caverns or depleted gas reservoirs. The operational parameters are expanded to include extreme low temperatures (-150°C to -50°C) and high pressures (200-300 bar). The reservoir simulation models are adapted to include multiphase fluid dynamics under cryogenic conditions, incorporating the Joule-Thomson effect and phase-change boundaries within the reservoir rock. The classification model is trained on simulation data to predict the long-term stability of the sequestration site. It classifies potential injection scenarios as "Stable Sequestration," "High Risk of Caprock Fracture," or "Potential for Uncontrolled Phase Transition." Input parameters include injection temperature, pressure curves, cavern geometry, and rock thermal conductivity. The probabilistic estimate guides the operational plan for the safe and permanent storage of industrial gases.

graph TD
    A[Define Injection Scenario] -- Temp, Pressure, Duration --> B(Geomechanical & Thermal Simulation);
    B -- Simulation Output (Stress fields, Temp gradients) --> C{Train Classification Model};
    C -- Trained Model --> D[Input Target Scenario];
    D --> E{Generate Probabilistic Estimate};
    E -- 95% --> F[Class: Stable Sequestration];
    E -- 4% --> G[Class: High Risk of Caprock Fracture];
    E -- 1% --> H[Class: Uncontrolled Phase Transition];
    F --> I[Decision: Proceed with Injection];
    G --> J[Decision: Redesign or Abort];
    H --> J;

    style F fill:#9f9,stroke:#333,stroke-width:2px
    style G fill:#f99,stroke:#333,stroke-width:2px
    style H fill:#f99,stroke:#333,stroke-width:2px

---

Derivative Embodiment 3: Cross-Domain Application in Aerospace Materials Science

Enabling Description: The core method is adapted for predicting the fatigue life and failure probability of novel composite aerospace components (e.g., turbine blades, fuselage panels) under operational stress. A high-fidelity Finite Element Analysis (FEA) simulation is used in place of the reservoir simulator. The FEA model simulates decades of operational cycles (thermal, vibrational, aerodynamic loading). Thousands of simulations are run with varying material compositions (e.g., carbon fiber ply angles, resin matrix composition) and micro-fracture initial conditions. The output data (stress, strain, delamination progression) is used to train a 3D Convolutional Neural Network (3D-CNN) which acts as the classification model. For a proposed new component design, the 3D-CNN provides a probabilistic estimate of it belonging to one of three classes: "Exceeds 100,000-cycle lifespan," "Fails between 50,000-100,000 cycles," or "Catastrophic failure before 50,000 cycles." This classification directly informs the go/no-go decision for manufacturing and physical testing, drastically reducing development costs.

flowchart LR
    subgraph Simulation Phase
        A[Define Material Parameters & Load Cases] --> B(Run FEA Simulations);
        B --> C[Generate 4D Dataset (x,y,z,time)];
    end
    subgraph Training Phase
        C --> D(Train 3D-CNN Classifier);
    end
    subgraph Prediction Phase
        E[Propose New Component Design] --> F(Input Design into 3D-CNN);
        F --> G{Probabilistic Classification};
        G --> H[Class 1: >100k cycles];
        G --> I[Class 2: 50k-100k cycles];
        G --> J[Class 3: <50k cycles];
    end
    subgraph Decision
        H --> K(Decision: Certify for Production);
        I --> L(Decision: Redesign & Re-evaluate);
        J --> M(Decision: Reject Design);
    end

---

Derivative Embodiment 4: Cross-Domain Application in Algorithmic Trading

Enabling Description: The methodology is applied to classify the future performance of high-frequency trading (HFT) algorithms. In this context, the "reservoir simulator" is a market back-testing engine that simulates the algorithm's performance against years of historical tick-level market data. "Well parameters" are the algorithm's hyperparameters (e.g., lookback windows, risk thresholds, order sizes). The simulation results (profit/loss curves, Sharpe ratio, max drawdown) form the training data for a Long Short-Term Memory (LSTM) network, which serves as the classification model. Given a new set of hyperparameters for a target algorithm, the LSTM model produces a probabilistic estimate of its performance classification over the next quarter: "Alpha-Generating," "Market-Neutral," or "Capital-Depleting." This classification guides the decision of whether to deploy the algorithm with real capital in live markets.

stateDiagram-v2
    [*] --> Backtesting
    Backtesting: Run thousands of hyperparameter combinations on historical data
    Backtesting --> Training: Generate performance curves

    Training: Train LSTM model on performance curves
    Training --> Classification

    Classification: Input new algorithm's hyperparameters
    Classification --> P_Alpha: P=0.6
    Classification --> P_Neutral: P=0.3
    Classification --> P_Loss: P=0.1

    state "Decision" as D {
        P_Alpha --> Deploy: Activate algorithm in live market
        P_Neutral --> Review: Tweak parameters, re-classify
        P_Loss --> Archive: Reject algorithm
    }

---

Derivative Embodiment 5: Cross-Domain Application in Agricultural Technology (AgTech)

Enabling Description: This application predicts crop yield and classifies the success of a given planting strategy. A biophysical crop growth simulator (e.g., DSSAT, APSIM) replaces the reservoir simulator. The input parameters include seed genetics, soil composition, fertilizer/irrigation schedules, and long-range weather forecasts. Thousands of simulation runs generate a dataset of potential growth outcomes over a full season. This data is used to train a Gaussian Process Classifier. For a farmer's proposed planting strategy for the upcoming season, the system generates a probabilistic estimate classifying the likely outcome as "High Yield (>90th percentile)," "Average Yield," or "Crop Failure/Low Yield (<25th percentile)." This allows for the optimization of resource allocation and the purchase of appropriate crop insurance before a single seed is planted.

graph TD
    A[Input: Seed, Soil, Weather, Strategy] --> B(Crop Growth Simulation);
    B -- Thousands of runs --> C(Generate Training Dataset);
    C --> D(Train Gaussian Process Classifier);
    E[Farmer's Proposed Plan for Season] --> F(Input to Trained Classifier);
    F --> G{Probabilistic Outcome Classification};
    G -- "Prob > 0.7" --> H[High Yield];
    G -- "Prob > 0.2" --> I[Average Yield];
    G -- "Prob < 0.1" --> J[Crop Failure];

---

Derivative Embodiment 6: Integration with Emerging Tech (AI, IoT, Blockchain)

Enabling Description: The patented system is integrated into a closed-loop "Digital Twin" of the reservoir, creating a dynamic, self-optimizing system.

1. AI-driven Optimization: A Genetic Algorithm (GA) or Reinforcement Learning (RL) agent is used to propose the initial well configuration and location parameters. Its goal is to maximize the probability of a "Good" classification from the model, intelligently exploring the parameter space instead of relying on human engineers.
2. IoT Integration: Real-time data from downhole Distributed Temperature Sensing (DTS) and Distributed Acoustic Sensing (DAS) fiber-optic cables are streamed to the system. This data provides an instantaneous, high-resolution view of fluid flow and reservoir dynamics. The classification model is continuously re-trained or fine-tuned with this live data, allowing it to adapt to changing reservoir conditions.
3. Blockchain for Verification: Every simulation run, model training event, probabilistic estimate, and subsequent operational decision (e.g., "Drill Well at X, Y, Z") is recorded as a transaction on a private, permissioned blockchain (e.g., Hyperledger Fabric). This creates an immutable, auditable, and cryptographically secure log of the entire decision-making process, which can be shared with regulators, partners, or insurers to verify compliance and operational integrity.

classDiagram
    class GeneticAlgorithm {
        +proposeWellParameters()
    }
    class ReservoirSimulator {
        +runSimulation(params)
    }
    class ClassificationModel {
        -model
        +train(data)
        +predict(params)
    }
    class IoTSensorStream {
        +getRealTimeData()
    }
    class BlockchainLedger {
        +recordDecision(decisionData)
    }
    GeneticAlgorithm --> ReservoirSimulator : "Submits parameters"
    ReservoirSimulator --> ClassificationModel : "Provides training data"
    IoTSensorStream --> ClassificationModel : "Provides fine-tuning data"
    ClassificationModel --> GeneticAlgorithm : "Returns fitness score"
    ClassificationModel --> BlockchainLedger : "Logs prediction & decision"

---

Derivative Embodiment 7: The "Inverse" or Failure-Mode Classifier

Enabling Description: A parallel version of the classification model is trained specifically to identify scenarios leading to undesirable or catastrophic outcomes. Instead of classifying for production ("Good"/"Bad"), this "Hazard Classifier" is trained on simulation data that is specifically labeled for failure modes. The labels include "Wellbore Instability," "Casing Collapse," "Uncontrolled Gas Kick," and "Premature Water Breakthrough." The input parameters (well configuration, drilling plan, geological properties) are processed by a Support Vector Machine (SVM) or a deep neural network trained with a focal loss function to handle the rarity of failure events. The system outputs a probabilistic estimate of the risk for each specific hazard. An operational decision to drill is only considered acceptable if the primary performance classifier predicts "Good" AND the Hazard Classifier predicts a probability below a critical safety threshold (e.g., <0.1%) for all catastrophic failure modes.

flowchart TD
    A[Input: Proposed Well Plan] --> B{Performance Classifier};
    A --> C{Hazard Classifier};
    B --> B_Good["P(Good) > 0.8"];
    B --> B_Bad["P(Good) <= 0.8"];
    C --> C_Safe["P(Hazard) < 0.001"];
    C --> C_Unsafe["P(Hazard) >= 0.001"];

    subgraph Decision Logic
        direction LR
        X[B_Good] & Y[C_Safe] --> Z[ACCEPT];
        X & C_Unsafe --> W[REJECT - Unsafe];
        B_Bad & Y --> V[REJECT - Poor Performance];
        B_Bad & C_Unsafe --> W;
    end

---

Combination Prior Art Scenarios

1. Combination with TensorFlow/PyTorch and Open-Source Simulators: The entire method is implemented using exclusively open-source components. The reservoir simulation is performed using the Open Porous Media (OPM) Initiative's open-source simulator. The resulting data is processed and used to train a classification model (e.g., a Recurrent Neural Network or a Gradient Boosted Tree model like XGBoost) built, trained, and deployed using the Python-based TensorFlow or PyTorch libraries. This combination demonstrates that the entire patented workflow can be achieved without proprietary software, making the specific combination of these well-known tools for this purpose obvious to a person skilled in the art.

2. Combination with Kubernetes and OPC UA: The system is architected as a cloud-native microservices application managed by Kubernetes. The reservoir simulator is a scalable service, the model training is another, and the prediction/inference is a lightweight API endpoint. Real-world operational data for model re-training is ingested from field equipment (pumps, chokes, sensors) using the OPC UA industrial communication standard (IEC 62541). This data is fed into a Kafka data stream, processed, and used to trigger automated model re-training jobs within the Kubernetes cluster. This architecture makes the patented process scalable, resilient, and interoperable with standard industrial control systems.

3. Combination with Apache Spark and MLlib: For extremely large reservoirs requiring massive ensembles of simulations ("giant reservoirs"), the data processing and model training steps are performed on a distributed computing cluster using Apache Spark. The simulation outputs, stored in a data lake (e.g., in Parquet format), are loaded into a Spark Resilient Distributed Dataset (RDD) or DataFrame. The feature engineering and training of the classification models (e.g., Naive Bayes, Decision Trees, K-Means Clustering, as mentioned in the patent) are executed in a massively parallel fashion using Spark's built-in MLlib library. This combination addresses the computational bottleneck of the original patent, demonstrating an obvious path to scale the method for industrial-level problems using standard, open-source big data technologies.