Process simulation is frequently adopted to facilitate the optimization of the resin transfer molding process. However, it is computationally costly to simulate the multi-physical, multi-scale process, making it infeasible for applications involving huge datasets. In this study, the application of K-nearest neighbors and artificial neural network metamodels is proposed to build predictive surrogate models capable of relating the mold-filling process input-output correlations to assist mold designing. The input features considered are the resin injection location and resin viscosity. The corresponding output features investigated are the number of vents required and the resultant maximum injection pressure. Upon training, both investigated metamodels demonstrated desirable prediction accuracies, with a low prediction error range of 5.0% to 15.7% for KNN metamodels and 6.7% to 17.5% for ANN metamodels. The good prediction results convincingly indicate that metamodeling is a promising option for composite molding applications, with encouraging prospects for data-intensive applications such as process digital twinning.

The article presents the results of modeling various modes of vacuum infusion molding of thin-walled polymer-composite structures of arbitrary geometry. The small thickness of the manufactured structures and the fixation of their back surface on the rigid surface of the mold made it possible to significantly simplify the process model, which takes into account the propagation of a thermosetting resin with changing rheology in a compressible porous preform of complex 3D geometry, as well as changes in boundary conditions at the injection and vacuum ports during the post-infusion molding stage. In the four modes of vacuum-infusion molding studied at the post-infusion stage, the start time, duration and magnitude of additional pressure on the open surface of the preform and in its vacuum port, as well as the state of the injection gates, were controlled (open–closed). The target parameters of the processes were the magnitude and uniformity of the distribution of the fiber volume fraction, wall thickness, filling of the preform with resin and the duration of the process. A comparative analysis of the results obtained made it possible to identify the most promising process modes and determine ways to eliminate undesirable situations that worsen the quality of manufactured composite structures. The abilities of the developed simulation tool, demonstrated by its application to the molding process of a thin-walled aircraft structure, allow one to reasonably select a process control strategy to obtain the best achievable quality objectives.

This paper considers a new approach to the modeling of the vacuum infusion process at the manufacturing of three-dimensional composite parts of complex shape. The developed approach and numerical methods focus on reliable prediction with needed accuracy and elimination of the unrecoverable defect of composite structure such as the dry spots. The paper presents some experimental results, which demonstrate two cases of dry spot formation in large aircraft composite panels, and analyzes the reasons for these defects arising. Our numerical technique is based on the vacuum infusion of the liquid resin into porous preform as the two-phase flow, which is described by the phase field equation coupled with the Richards equation describing the fluid motion in unsaturated soils with spatially varied pressure dependent porosity and saturation. This problem statement allowed correctly reconstructing the resin front motion and formation of inner and outer dry spots depending on its movement. For the rapid detection of preform zones that are suspicious for defect formation, two indicators calculated during process simulation are proposed and tested at the numerical experiments. The auxiliary program tool has been developed in the MATLAB environment to correctly detect the times of formation, localization and dimensions of the arising dry spots by using the results of the finite element model simulation.

High performance composite structures produced by the processes at which the consolidation of the fibres and matrix is done at the same time as the component is shaped. Full curing schedule include a pre-warming for resin liquefaction, next apply of pressure to remove the gas bubbles, and finally consolidation of resin at elevated temperature to its full polymerization. The change in the state of the composite should be made as possible uniformly across the thick-walled products. The complexity of process control is due to unobservability of the rheological state of material in a closed volume of a mould. In this paper we propose a mathematical model of epoxy-based thick-walled composite structure curing. PDE system linking a kinetic equation of the resin cure with heat transfer equation, take into account a phase transition from liquid to gel and further to the solid state. On the basis of transient analysis of the developed model we optimize the temperature control law.

This paper highlights the electric energy harvesting system that consists of piezoelectric (PZT) stack transducer which is excited by random mechanical oscillating force and charges the Li-Ion battery. This battery is a power supply for LED lamp lighting. A very big electric capacitance of the multilayered PZT stack affects on the electric energy flow from PZT energy source to the electric sink. The reason for this influence is because the electric capacitance of the device, its mechanical losses and leakage increase along with the increased number of PZT layers. An electronic circuitry of the studied energy harvesting system provides the interfaces between the piezoelectric device and electric load. Such the circuits should match the input impedance of electric load (battery) and output impedance of voltage source (PZT stack). Basing on the experimental data and the finite element (FE) investigation we derive and tune the lumped model of multilayered PZT stack, which transforms the energy of low frequency mechanical excitation with random amplitudes and discretely distributed frequencies to the energy of electric current. In order to optimize an efficiency of harvester we find the number of PZT layers at constrained excitation parameters, stack's dimensions, and generated voltage required by the charging battery. The lumped model of the whole harvesting system has been built in SimElectronics / Simulink MATLAB© environment. It was also simulated to find the best harvesting system parameters