Process Optimization Server for Dynamic Set-Point Updates
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for incremental process optimization between rigorous optimization runs often result in suboptimal solutions due to the use of conservative linear models and fixed cost coefficients, leading to large variations in setpoints, increased workload on tracking control systems, and potential product wastage.
Innovation Solution
A server system utilizing a linear model derived from a rigorous on-line modeling and equation-based optimization software (ROMEO) to generate incremental set-point changes, ensuring alignment with the latest model and cost coefficients, and incorporating customizable objective weights and variable bounds.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Speed
If simplified linear models and fixed cost coefficients are used for incremental optimization between rigorous optimization runs, then computation speed is improved, but optimization accuracy deteriorates leading to suboptimal solutions
Solution Approach 1:
The patent applies dynamics by making the linear model coefficients and cost coefficients dynamic rather than fixed. The system continuously updates these parameters based on current process conditions and rigorous optimization results, allowing the simplified linear model to adapt to changing operating conditions while maintaining computational speed. This resolves the contradiction by enabling the simplified model to achieve near-rigorous optimization accuracy without sacrificing computation speed.
Solution Approach 2:
The patent changes parameters by updating the linear model coefficients and cost coefficients between rigorous optimization runs based on current process data and previous rigorous optimization results. This parameter adaptation allows the simplified linear programming approach to maintain high optimization accuracy while preserving fast computation speed, directly resolving the contradiction between speed and accuracy.
2Reliability
If conservative linear models are used to ensure model validity, then model reliability is improved, but the region of applicability is restricted leading to suboptimal solutions
Solution Approach 1:
The system dynamically updates the linear model parameters and cost coefficients based on current operating conditions and rigorous optimization results. This dynamic adaptation expands the region of applicability beyond the conservative bounds of traditional linear models while maintaining reliability through continuous alignment with rigorous optimization directions and actual process behavior.
Solution Approach 2:
The patent implements feedback by using the results from rigorous optimization runs to update and refine the linear model parameters and cost coefficients. This feedback mechanism ensures that the simplified linear model remains reliable and accurate across a broader operating range, resolving the contradiction between model reliability and region of applicability.
3Device complexity
If incremental optimization is performed far from rigorous optimization directions, then computational simplicity is improved, but workload on tracking control system increases
Solution Approach 1:
The patent uses feedback from rigorous optimization results to guide incremental optimization directions. By aligning incremental optimization with rigorous optimization directions through updated cost coefficients and model parameters, the system maintains computational simplicity while minimizing tracking control workload, as the control system follows more direct paths to optimal setpoints.
Solution Approach 2:
The system changes parameters by updating cost coefficients and model parameters to reflect current process conditions and rigorous optimization directions. This ensures incremental optimization remains computationally simple while following efficient trajectories that reduce tracking control workload and time.
4Loss of time
If linear programming is used between rigorous optimization runs, then response time is improved, but solution optimality deteriorates
Solution Approach 1:
The patent makes the linear programming model dynamic by continuously updating coefficients and parameters based on current process conditions and rigorous optimization results. This dynamic approach maintains fast response time while improving solution optimality, as the simplified model adapts to reflect the true nonlinear process behavior and optimization directions.
Solution Approach 2:
The system changes parameters by updating linear model coefficients and cost coefficients between rigorous optimization runs. This parameter adaptation allows linear programming to maintain its computational speed advantage while achieving higher solution optimality by aligning with actual process conditions and rigorous optimization directions.
Data Source
Figure 1~2
Figure 3
Figure 4
AI summary
Some embodiments include a server system including a first logic module executable by a processor for receiving a data communication from an industrial control system coupled to a communications network. In some embodiments, the data communication comprises data or data streams associated with the industrial process. The program logic of the first logic module includes a model configured to receive a variable and iterate and converge an optimization problem to an optimization solution to a first level of optimization based at least in part on the variable and the data or data streams. A second logic module executable by the processor is operatively data-linked to the first logic module and utilizes a model to iteratively process data values of the optimization solution to a second level of optimization with an increased level of optimization over the first level of optimization.