Predictive Process Model Identification Using Safe Excitation Signals
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
The existing methods for identifying process models for model-based predictive multivariable control in process engineering require significant practical experience and in-depth understanding for planning and implementing active tests to generate adequate measured data, making the process cumbersome and operator-dependent.
Innovation Solution
A computer-implemented method for automated identification of process models using historical data, where manipulated variables are excited in a ramp-shaped and stepped manner to ensure controlled variables depart and return to their operating points within defined tolerance bands, with low-pass filtration to suppress noise, and data stored for least error squares analysis.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional step tests are used to excite the dynamic process for model identification, then measured data with adequate information content can be generated, but the process requires significant practical experience and in-depth control engineering understanding, making it operator-dependent and cumbersome
Solution Approach 1:
The system performs self-identification by automatically generating excitation signals and processing measured data without requiring operator intervention or expertise in control engineering. The automated identification procedure excites the dynamic process and identifies process models independently, transforming an operator-dependent manual process into a self-service automated system
Solution Approach 2:
The manual mechanical process of planning and executing step tests by operators is replaced with an automated computational system. The automated identification procedure uses computer-generated excitation signals and algorithmic data processing to substitute the manual mechanical operations, eliminating the need for operator experience and understanding
2Measurement precision
If strong excitation signals are applied to the process for model identification, then dynamic behavior becomes clearly identifiable in measured data, but the excitation may affect safety and product quality
Solution Approach 1:
The system applies excitation signals that are sufficient (but not excessive) to achieve clear identifiability of dynamic behavior. By using automated optimization, the excitation strength is precisely calibrated to the minimum required level, avoiding both insufficient excitation and excessive excitation that could harm safety or product quality
Solution Approach 2:
The automated identification procedure incorporates feedback mechanisms to monitor the process response and adjust excitation signals accordingly. This feedback control ensures that excitation remains within safe boundaries while maintaining sufficient strength for accurate model identification, preventing harmful effects on safety and product quality
3Reliability
If manual planning of step tests is performed through intensive discussions, then practical experience and understanding can be utilized, but the process requires significant time and resources
Solution Approach 1:
The system performs preliminary automated analysis of historical measured data to determine appropriate excitation parameters and test configurations before actual model identification. This preliminary action eliminates the need for time-consuming manual discussions and planning, while ensuring reliable measured data quality through algorithmic optimization
Solution Approach 2:
The manual mechanical process of intensive discussions between operators and engineers is replaced with automated computational procedures. The system independently analyzes historical data, generates excitation signals, and plans test procedures, substituting human collaborative planning with efficient automated algorithms that achieve the same or better results in less time
Data Source
AI summary
A computer-implemented method for the automated identification of a process model for a model-based, predictive multivariable control of a process installation, wherein reference is made to previously defined controlled variables, manipulated variables and disturbance variables for the model-based, predictive multivariable control of the process installation.