Towards an automated masking process: A model-based approach

Abstract

The masking of aircraft engine parts, such as turbine blades, is a major bottleneck for the aerospace industry. The process is often carried out manually in multiple stages of coating and curing, which requires extensive time and introduces variations in the masking quality. This paper investigates the automation of the masking process utilising the wellestablished time-pressure (T/P) dispensing process for controlled maskant dispensing, and a robotic manipulator for accurate part handling. A mathematical model for the T/P dispensing process was derived, extending previous models from the literature by incorporating the robot velocity for controlled masking line width. An experiment was designed, based on the theoretical analysis of the dispensing process, to derive an empirical model from the generated data that incorporates the losses that are otherwise difficult to model mathematically. The model was validated under new input conditions to demonstrate the feasibility of the proposed approach and the masking accuracy using the derived model.

A novel modelling of glue allowance prediction for time-pressure dispensing system based on gated recurrent unit and fully connected neural network

“…Previous researches presented various modeling for time-pressure dispensing system. Such modelling applied automatic control algorithm to calculate the current glue output to ensure the stability of the glue output (Elgeneidy et al, 2019;Liu et al, 2007;Chen et al, 2007). However, the time-pressure dispensing system involves the rheological properties of the glue and the compressibility of the gas, which puts the dispensing process in complicated environment, ultimately the model is often over-simplified in the modeling process practice.…”

Detection and Classification of Aircraft Fixation Elements during Manufacturing Processes Using a Convolutional Neural Network

“…One fundamental part of referencing these processes is the detection and measurement of the parts in their real positions by artificial vision systems, which have been used in different industrial sectors [6,[11][12][13][14][15]. Non-contact methods are being increasingly employed during aeronautical manufacturing processes as enabling technology in the fields of verification [6], metrology [7,10,13] and quality analysis [11] are implemented, both for their excellent inherent performance and the saved cost of accelerating production rates to meet production demands.…”

“…Images obtained by vision systems can be identified and classified by disruptive technology, such as artificial intelligence [14,[19][20][21], and machine-learning techniques specifically have been used widely in the industry to classify objects [6][7][8]13].…”

Towards a Decision-support Framework for Reducing Ramp-up Effort in Plug-and-Produce Systems

“…. Exemplary outcomes of the glueing process[24]. While outcome 3 represents the desired outcome, 1 and 2 show unacceptable quality results (discontinuity of line; dispensed line too thick).…”