Fuzzy Logic Modelling of a Powder Coating System
D. T. Pham, S.S. Dimov, O. A. Williams, N. B. Zlatov
Manufacturing Engineering Centre, Cardiff University, UK
Outline
- Introduction
- Powder Application Process
- Powder Application Variables
- The Design Methodology
- Fuzzy Modelling
- Response Surface Modelling
- The Fuzzy-logic Based Powder Coating System
- Evaluation of the performance of the fuzzy-logic model
- Conclusion
- Acknowledgement
Introduction
- Electrostatic powder coating is an attractive alternative to solvent based coating
- The technology has both environmental and economic benefits
- The efficiency of a powder application process is indicated by two common terms:
Introduction
- system utilisation efficiency (STE): ratio of powder applied to powder used (including reclaim)
Powder Application Process
Powder Application Process
Variables to considered in order to achieve high FPTE includes:
· Part conveying and orientation
· Canopy design
· Gun-to-part distance
· Applied powder-volume pressure
· Nozzle design
· Powder pump and fluidising
· Gun design
· Applied gun current and voltage
· Booth airflow
Design Methodology for the predictive Fuzzy-Logic Model
-
The predictive fuzzy-logic based powder coating system has four application variables as inputs and one output variable, powder deposition (grams)
-
Design of Experiment (DOE) was used to investigates the effects of different combinations of the application variables on powder deposition
-
Response surface method was used to search for the optimal value of powder deposition and to construct the response surface plots of the four application variables
-
The predictive fuzzy-logic model was designed to replicate the behaviour of the response surface plots given similar experimental conditions
-
Figure 3 illustrates the design methodology
Design Methodology for the predictive Fuzzy-Logic Model
Response Surface Method
- Response Surface Method (RSM) is a composite technique that implements the steepest ascent method to search for optimum points.
- The method can be used to search for the optimal response within ranges of the experimental conditions.
- Using the data collected from the experimental work RSM was used to construct the response surface plots for the four application parameters as shwon in figure 4.
The Fuzzy Based Powder Coating System
-
Figure 5 presents a Mamdani type fuzzy-logic model. Membership functions for the four inputs and one output variable to the fuzzy-logic model were defined used triangular membership functions.
-
The rule based of the fuzzy-logic model were extracted to reflect the behaviour of the process as depicted by the response surface plots.
- A typical rule reads: IF Powder-volume pressure is Medium, workpiece rotation speed is NOT High, Gun-to-part distance is NOT High AND Gun current is NOT Low, THEN powder deposited on the workpiece is NOT Medium.
The Fuzzy Based Powder Coating System
The Fuzzy Based Powder Coating System
The Fuzzy Based Powder Coating System
Evaluation of the Perfomance of the fuzzy-logic model
