Multi-Modal MPC Optimization Using Adjusted Operating Condition Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current optimization algorithms for processing plants, particularly those with multi-modal processing units, face issues due to underlying assumptions in MPC architectures that are violated when multiple operational modes are considered, leading to incorrect results and inefficiencies.
Innovation Solution
A computer-implemented method and apparatus that generate adjusted operating condition data by subtracting default operating condition data from identified operating condition data at specific timestamp intervals, incorporating additional operating condition data for other operational modes, to derive corrective responses for optimizing processing unit operations without modifying the underlying control scheme.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If current optimization algorithms operate based on underlying assumptions of MPC architecture, then the control scheme remains simple and stable, but the algorithms produce incorrect results when multiple operational modes are considered
Solution Approach 1:
The patent applies dynamics by making the operating condition data time-varying and mode-specific. Instead of using static default assumptions, the system dynamically adjusts operating condition data based on timestamp intervals and specific operational modes, allowing the optimization algorithm to adapt to changing plant conditions while maintaining mathematical consistency with the MPC architecture.
Solution Approach 2:
The patent changes the parameters of operating condition data from fixed default values to variable values that depend on timestamp intervals and operational modes. By modifying these parameters dynamically, the system resolves the contradiction between maintaining simple MPC assumptions and accurately representing multi-modal plant operations.
2Adaptability or versatility
If default operating condition data is used for optimization, then the MPC architecture assumptions are satisfied, but the optimization does not account for actual plant operations in different operational modes
Solution Approach 1:
The patent segments the optimization process by dividing timestamp intervals into subintervals, each associated with specific operational modes. This segmentation allows the system to handle complex multi-modal operations by breaking them into manageable pieces, where each subinterval can be processed using standard MPC assumptions while collectively covering all operational modes.
Solution Approach 2:
The patent introduces intermediary operating condition data that acts as a bridge between default operating condition data and actual plant operations. This intermediary data structure allows the system to incorporate multi-modal information while maintaining compatibility with the underlying MPC architecture, effectively mediating between simplicity and accuracy.
3Reliability
If the underlying control scheme is modified to accommodate multi-modal operation, then accurate optimization for multiple modes is achieved, but the control scheme complexity increases and default assumptions are violated
Solution Approach 1:
The patent applies partial action by modifying only the operating condition data parameters rather than the entire control scheme. This partial modification is sufficient to achieve accurate multi-modal optimization while leaving the core MPC architecture and its default assumptions intact, thereby avoiding excessive complexity.
Data Source
AI summary
Embodiments generate improved optimizations of processing unit(s) associated with a particular dynamic control scheme (e.g., an MPC architecture). Embodiments perform the improved optimizations based on default assumptions associated with the scheme, for example status operation of a particular mode. Some embodiments generate default operating condition data associated with a particular timestamp interval, the default operating condition data at a first value representing a default assumption based on a first operational mode, identify first operating condition data associated with the first operational mode, representing the first value at particular subinterval(s) of the particular timestamp interval, generating intermediary operating condition data by removing the first operating condition data from the default operating condition data, generating adjusted operating condition data by including second operating condition data associated with a second subinterval of the subinterval of the particular timestamp interval, and optimizing based at least in part on the adjusted operating condition data.