Heuristics guided optimization of a batch autoclave curing process

Effects of stress on the adhesive behavior of photoresist

“…Above the T g , the photoresist may flow and there would be little stress. However, due to the cure of the photoresist, the cross-linking reaction takes place, the flow of the photoresist would stop [4], and the stress would arise. The sources of the internal stress of the photoresist could be the shrinkage of the photoresist during the curing process and the shrinkage difference between the photoresist and the substrate.…”

“…The cure is an exothermic chemical reaction and causes the shrinkage of the resin. The rate of cure is faster when the temperature is higher [4]. Due to the exothermicity of the cure, the temperature of the photoresist during baking is higher than the baking temperature (Fig.…”

Stochastic multi-objective optimisation of the cure process of thick laminates

“…Fig. 7c illustrates the dependence of ΔT max on T 1 and r for constant values for the rest of the parameters (T , dt , h, m, E , α ) 2 1 2 0 as reported in Table 5 (Case 2). The agreement between the surrogate model and the simulation is very good.…”

“…These are parameters of the two-dwell cure profile illustrated in Fig. 3, such as the temperature of first (T ) 1 and second dwell (T ) 2 , the duration of the first dwell (dt ) 1 and the heat ramp rate (r) and stochastic process variables such as the surface heat transfer coefficient (h), the activation energy (E 2 ) and reaction order (m) of the autocatalytic component of the cure kinetics model and the initial degree of cure (α 0 ). The surrogate modelling methodology treats these in the same way, with some of them such as dwell temperature having a dual role both as optimisation parameters and stochastic variables.…”

“…Cure time can be reduced by up to 30% for thick parts [1][2][3] and 50% for ultra-thick parts [4][5][6], whereas targeting the minimisation of residual stresses in a single-objective profile optimisation context can lead to their reduction by about 30% [7][8][9][10][11][12]. The optimal solutions obtained in single optimisation setups merging objectives in a weighted sum [13,14] are dependent on the weights which imply a relative prioritisation between the different objectives.…”