Nonlinear Parameter Varying Model Identification via Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current model-based control algorithms for nonlinear systems face limitations due to the need for accurate and reliable nonlinear models, particularly in industrial processes where both parameter varying and input-output nonlinearity exist, often requiring extensive data for identification.
Innovation Solution
A method for identifying nonlinear parameter varying models that uses a system comprising an input module, identified object, output module, and system identification module, applying excitation signals and nonlinear model identification to capture both local nonlinear models and their transitions, converting these into multi-input single-output nonlinear parameter varying models using interpolation philosophy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If linear parameter varying models are used to describe the system, then computation time is reduced and model structure is simplified, but the nonlinear relationship between system input and output at a given operating point is ignored
Solution Approach 1:
The patent segments the nonlinear system into multiple local linear models at different operating points. Each local model captures the linear behavior around a specific operating point, while the collection of these segmented models represents the overall nonlinear system. This is achieved by dividing the operating range into multiple regions and identifying a linear model for each region.
Solution Approach 2:
The patent makes the model parameters dynamic by allowing them to vary with operating conditions. Instead of using fixed linear parameters, the model parameters are made to change dynamically based on the operating point, enabling the model to adapt to different system states and capture nonlinear behavior through parameter variation.
2Measurement precision
If comprehensive nonlinear identification is performed to capture both input-output nonlinearity and parameter varying nonlinearity, then model accuracy is improved, but data requirements and identification complexity increase
Solution Approach 1:
The identification process is segmented into multiple stages: first identifying local linear models at each operating point, then determining parameter variation patterns. This segmentation allows the use of limited data at each local point rather than requiring comprehensive data for the entire operating range simultaneously.
Solution Approach 2:
The patent performs preliminary identification of local linear models at each operating point before synthesizing the overall nonlinear parameter varying model. This preliminary action at each local point reduces the overall data requirement by breaking down the complex identification task into manageable local tasks that can be performed with limited local data.
Data Source
AI summary
The invention discloses an identification method of nonlinear parameter varying models (NPV) and belongs to the industrial identification field. The invention carries out identification tests and model identification for an identified object with nonlinear parameter varying characteristics. Firstly, the multi-input single-output nonlinear parameter varying model is identified through the steps of local nonlinear model tests, local nonlinear models identification, and operating point variable transition tests; after completing the identification of all the multi-input single-output nonlinear parameter varying models with respect to all the controlled variables, the completed multi-input multi-output nonlinear parameter varying models are built. The nonlinear parameter varying models of an identified object can be obtained by the identification method of the present invention with limited input/output data without detailed mechanism knowledge of the identified object. The nonlinear parameter varying models obtained can be used in model-based control algorithm design and process simulation, as well as in product quality prediction reasoning models and soft sensors.