Process Plant Control With Operator-Aware AI Validation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing control systems for process plants, such as DCS, do not adequately consider the expertise of control room operators and often lack timely feedback, leading to inefficient automation, especially when inexperienced operators provide feedback or are unavailable.

Innovation Solution

A Distributed Control System (DCS) that uses AI to autonomously control processes by detecting operator availability through historical data and sensory parameters, such as facial expressions and body movements, and validates recommended control operations with experienced operators when available, allowing for autonomous operation even when operators are not directly involved.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If AI models are used to automate control room operations, then productivity is improved, but reliability deteriorates due to lack of operator expertise consideration

Engineering Contradiction:

Improveautomation efficiencyVSAvoidcontrol accuracy

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system implements feedback by having control room operators validate AI-recommended control operations. The validation feedback from operators is used to retrain the AI model, creating a closed-loop system that continuously improves accuracy while maintaining automation efficiency.

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system introduces an intermediary validation mechanism where operators review and validate AI recommendations before implementation. This intermediary step ensures that operator expertise is considered while maintaining the productivity benefits of AI automation.

Inventive Principle:

Principle #24Intermediary (Mediator)

2Manufacturing precision

If constant feedback from control room operators is received to train models, then manufacturing precision is improved, but loss of time increases due to operator unavailability

Engineering Contradiction:

Improvemodel accuracyVSAvoidoperator availability time

Core Design Contradiction:

Manufacturing precisionVSLoss of time

Solution Approach 1:

The system uses periodic action by detecting operator availability and requesting validation only during available periods. Historical data about operator schedules and availability patterns are used to timing validation requests appropriately, avoiding times when operators are busy with critical control tasks.

Inventive Principle:

Principle #19Periodic action

Solution Approach 2:

The system performs preliminary action by pre-detecting operator availability using historical data before requesting validation. This allows the system to plan validation requests in advance during periods when operators are likely to be available, reducing delays in model training.

Inventive Principle:

Principle #10Preliminary action

3Reliability

If validation from control room operators is requested, then reliability is improved, but productivity decreases due to operator involvement

Engineering Contradiction:

Improvevalidation accuracyVSAvoidautonomous control speed

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The system applies partial action by requesting validation only for specific control operations rather than all operations. Validation is requested selectively based on the AI model's confidence level and the criticality of the control operation, maintaining high productivity while ensuring reliability where needed.

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system uses dynamics by adaptively adjusting the validation requirement based on operator availability and process conditions. When operators are available and process conditions warrant it, validation is requested; when operators are unavailable or conditions are routine, the system operates autonomously without validation delays.

Inventive Principle:

Principle #15Dynamics

4Loss of information

If inexperienced operators provide feedback, then loss of information is reduced, but reliability deteriorates due to variable feedback quality

Engineering Contradiction:

Improvefeedback collectionVSAvoidfeedback consistency

Core Design Contradiction:

Loss of informationVSReliability

Solution Approach 1:

The system replaces the mechanical system of direct operator judgment with sensor-based detection of physiological states. Imaging units capture facial expressions, eye gaze, and body movements to objectively assess operator state, providing consistent feedback quality independent of operator experience level.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Data Source

PatentUS12013690B2Method and system for controlling a process in a process plant

Publication Date: 2024.06.18 ABB (SCHWEIZ) AG

AI summary

The present invention relates to a method and a system for controlling a process of a process plant. Hie system is configured to control a process in the process plant by recommending control operations and perform the control operations. The system obtains the control operations and detects availability of control room operators using historical data related to actions performed by the control room operators, sensory parameters of the control room operators and a status associated with the process. The system provides queries related to the control operations, the plant parameters, status of the process to the control room operators upon detecting the availability. Hie queries are validated by the control room operators. The system thereafter autonomously controls tire control room based on the validated data.