Michael Bortz

The critical spot eraser-a method to interactively control the correction of local hot and cold spots in IMRT planning.

Common problems in inverse radiotherapy planning are localized dose insufficiencies like hot spots in organs at risk or cold spots inside targets. These are hard to correct since the optimization is based on global evaluations like maximum/minimum doses, equivalent uniform doses or dose-volume constraints for whole structures. In this work, we present a new approach to locally correct the dose of any given treatment plan. Once a treatment plan has been found that is acceptable in general but requires local corrections, these areas are marked by the planner. Then the system generates new plans that fulfil the local dose goals. Consequently, it is possible to interactively explore all plans between the locally corrected plans and the original treatment plan, allowing one to exactly adjust the degree of local correction and how the plan changes overall. Both the amount (in Gy) and the size of the local dose change can be navigated. The method is introduced formally as a new mathematical optimization setting, and is evaluated using a clinical example of a meningioma at the base of the skull. It was possible to eliminate a hot spot outside the target volume while controlling the dose changes to all other parts of the treatment plan. The proposed method has the potential to become the final standard step of inverse treatment planning.

Challenges in process optimization for new feedstocks and energy sources

Abstract Current and future challenges of optimization in the process industry are discussed. The gap between academic research and industrial workflow is analyzed. Moreover, issues arising from the shift from conventional fossil fuels as both feedstock and energy source to nonconventional feedstocks (shale gas, tar sands, CO2 and biomass) and penetration of intermittent renewable energy are discussed. This manuscript focuses mainly on offline model-based optimization of design and operation, including the generation and selection of promising process alternatives for new feedstocks in conceptual design, multi-objective optimization, the estimation of thermodynamic parameters of new intermediates and the optimization of process operation under the volatile availability of the new feedstocks and energy sources. Moreover, a number of opportunities and needs for research and development are identified, including the simultaneous optimization of feedstocks, processes and products and a production able to process a variety of feedstocks and to utilize energy when it is cheap.

Decision Support by Multicriteria Optimization in Process Development: An Integrated Approach for Robust Planning and Design of Plant Experiments

Abstract In simulation-based process design, model parameters, like thermodynamic data, are affected by uncertainties. Optimized process designs should, among different other objectives, also be robust to uncertainties of the model parameters. In industrial practise, it is important to know the trade-off between an increase in robustness and the other objectives – like minimizing costs or maximizing product purities. This contribution describes a practical procedure how to incorporate robustness as an objective into a multicriteria optimization framework. The general procedure is illustrated by a concrete example. Finally, we argue that the same approach is useable for an optimal design of plant experiments.

Graybox Models - New Opportunities for the Optimization of Entire Processes

Abstract The use of process optimization is often limited due to missing models for certain process steps. The systematic combination of available operating data with previous knowledge in so called graybox models helps to bridge this gap. Following the concept of incremental model identification it will be shown how the workflow has been integrated into a simulation software environment. It will be applied to a cumene process.

A new scheme for process simulation by optimization: distillation as an example

Abstract Commercial flowsheet simulators suffer from limitations in problem formulation flexibility and numerical instabilities which potentially turn process simulation into a tedious task. In this work, we address these problems and present a new scheme for integrated process simulation and optimization. The new method is developed here for distillation but can be extended to other unit operations. It utilizes numerically stable distillation calculations based on the solution of a sequence of fixed-point problems. Using the ideas of the shooting method the process simulation is embedded into an optimization problem and hence feasible design parameters are found by applying suitable nonlinear optimization algorithms. Systematic studies of typical examples of distillation processes are presented including also flowsheets with several columns. The new method is shown to be flexible in terms of problem formulation and more robust than those presently used in process simulators.

INES – Interface between Experiments and Simulation

Abstract The development of chemical processes is usually based on both experiments (often in pilot plants), and process simulation. Design of experiments, data evaluation and reconciliation, model development and validation are essential steps in this procedure. Different tools and approaches are available for each of these tasks but in the process developer’s workflow, they are usually not supported in an integrated way. Therefore, in the project INES, on which this paper reports, a new interface between experiments and simulation for process design was created, and integrated in a tool box which comprehensively supports process design. It contains modules for data selection and reconciliation, sensitivity analysis, and model validation and -adjustment. Methods from the literature are suitably combined to support the overall goal. The chosen methods, their combination and implementation are described and examples are given which demonstrate the benefits of the new interactive tool in the process development workflow.

Model-based Design of Experiments Using a Flowsheet-Simulator

Abstract The quality of process optimization depends on model parameters. The use of operating data from mini-, pilot or production plants can help to improve model agreement with operation. To support the design of experiments (DoE) in plants with many operability limitations a model-based DoE was implemented into a flowsheet simulator. In this contribution, the implementation of the method is described and its use for the identification of kinetics is demonstrated for a cumene process.