Optimization of model parameters

ASCMO-MOCA supports various file formats, allowing you to load and weight multiple datasets for optimization. Channel names in measurement files can be easily mapped to model names, enabling project reuse with new datasets. You can also convert measurement units as needed. The tool offers interactive visualizations like scatter plots and scope views, which can be saved as templates for future use.

In ASCMO-MOCA, you start by providing the relevant models, whether for plant simulations or controller optimizations. It supports formats like Simulink, FMU, ASCET, TSim Plugin, and ASCMO-STATIC/DYNAMIC. Each model is configured to access inputs, outputs, and internal parameters, with Simulink models running directly in Simulink for version independence.

ASCMO-MOCA enables you to link multiple models using customizable formulas, allowing flexible model creation and enhancement. This formula-based approach speeds up optimization, and the integrated formula editor simplifies the process by providing access to existing measurement channels, parameters, and model outputs, along with standard arithmetic operations.

ASCMO-MOCA makes it easy to create, manage, and edit model parameters such as maps, curves, and scalar values. You can import/export parameters in common formats or create new ones. Parameters are displayed in appropriate formats, such as maps, and can be adjusted graphically using the mouse or through tables using the keyboard. The tool also allows upper/lower limits and reference values to be set for each parameter. You can also manage different sets of parameters, including current, reference, and user-defined sets, with all optimization iterations and versions automatically saved. Parameter smoothing and comparison is also possible.

ASCMO-MOCA optimizes model parameters by comparing model outputs with real measurements and calculating the root-mean-square error (RMSE) to gauge correlation. The optimizer minimizes RMSE by varying all the free parameters. You can also define various boundary conditions to be considered in the optimization:

• Adjustment of parameter ranges
• Smoothness, gradients and monotonicity of maps/curves
• Simultaneous consideration of multiple optimization criteria, including weighting
• Exclusion of specific map/curve values from the optimization

ASCMO-MOCA allows you to divide the optimization into sub-steps to achieve goals faster and ensure better parameterization. Key functionalities include:

• Selecting parameters for optimization
• Modifying breakpoints of maps/curves
• Adjusting dataset weightings or criteria
• Adapting limits and smoothness
• Selecting data subsets

ASCMO-MOCA provides several visualization tools to analyze the optimization results in detail. For example, the error distribution before and after an optimization can be displayed in the form of a histogram or any combination of point clouds. The tool also provides sensitivity and correlation analysis.