Improve Model Accuracy using AI driven process simulation

Understand the power of first principles-driven hybrid models in industrial applications. Learn the advantages of integrating mechanistic principles with AI and machine learning techniques to create accurate, predictive models. Develop hybrid models tailored for critical processes such as distillation columns, reactors, heat exchangers, pressure changers, and separators. Additionally, use real-time plant data to enhance model accuracy to replace inadequately modeled relationships not fully captured by traditional engineering models.

Audience:

  • Process Engineers who would like to learn how to apply AI and machine learning to improve their modeling activities.

Training Details

  • Course Id:

    EHM124

  • Duration:

    2 day(s)

  • CEUs Awarded:

    1.4

  • Level:

    Introductory

Benefits

  • Discover how to build a predictive AI model with no specific AI background 
  • Leverage historical plant data to enhance first-principles models using AI/ML to improve modeling accuracy
  • Improve standardized calibration workflow using AI capabilities

Approach

  • Delivered through a blend of lectures, and case study discussions

Pre-requisites

Basic knowledge of Aspen Plus/Aspen HYSYS recommended

Agenda

  • Introduction to First Principles Driven Hybrid Models
    • Learn the general concepts and definitions of First Principles Driven Hybrid Models
    • Identify different types of workflows and their uses
    • Recognize First Principles Driven Hybrid Modeling architecture and use cases
  • Data Formatting and Pre-Processing
    • Become familiar with the format required for building First Principles Driven Hybrid Models from plant data
    • Identify different types of data preprocessing available
    • Understand the new AI training interface in the simulator
    • Learn how to import raw data into the simulator
    • Workshop 1: Format and Import Plant Data
  • Analyzing and Conditioning Raw Data
    • Understand the tools to analyze raw data
    • Identify trends and correlations within the data
    • Learn how to apply different conditioning techniques to raw data
    • Workshop 2: Analyze and Condition Raw Data
  • Building the Hybrid Model
    • Evaluate the need to train a hybrid model for your process
    • Build a hybrid model from conditioned plant data
    • Select dependent and independent variables to be used in the hybrid model
    • Identify a Neural Network Output to be trained in the hybrid model
    • Workshop 3: Evaluate and Train a Hybrid Model
  • Validating the Hybrid Model
    • Identify best practices to validate the hybrid model
    • Use the snapshot feature to try different data conditioning and NN configuration
    • Understand key parameters to evaluate the accuracy of the hybrid model before its deployment
    • Workshop 4: Validating the Hybrid Model
  • Deploying the Hybrid Model
    • Learn how to deploy hybrid models in the process simulator
    • Explore the automatic changes made to the simulation interface once the model has been deployed
    • Workshop 5: Deploying the Hybrid Model
  • Using and Sustaining the Hybrid Model
    • Enter the minimum input required for running the hybrid model
    • Recognize the usability of the model
    • Identify the need to re-train the model with newly updated data
    • Workshop 6: Using and Sustaining the Model
  • Additional use cases (optional):
    • Upstream example:
      • Friction Factor in Two-Phase Pipe in Aspen HYSYS
    • Downstream examples
      • Crude distillation unit in Aspen HYSYS
      • Distillation Column in Aspen Plus

 

Register for a Class

Date Class Type Location Price Language
Date(s): 04/15/2025 - 04/16/2025 Type: Public Classroom Location: Grand Hyatt Hotel
Al Khobar , Saudi Arabia 		Price: (USD) 1300.00 Language: English Register

Aspen Technology, Inc. awards Continuing Education Units (CEUs) for training classes conducted by our organization. One CEU is granted for every 10 hours of class participation.