Sensor placement study for online flow monitoring in liquid composite molding

On‐line sensing can play an important part in controlling the quality of the final product in any manufacturing environment, including liquid composite molding (LCM). Having a sensor embedded within the part itself is often the most effective means of monitoring its condition at various stages of manufacturing and even throughout its useful life. However, given their intrusive nature, there are practical limitations imposed upon their size, quantity and trajectory within the part. This study explores the possibility of using a single lineal sensor to monitor the resin flow front during the mold filling stage of LCM, and to detect the onset of void formation and the presence of dry spots within the mold. Experiments were conducted to characterize the response of a fiber optic system previously developed for cure monitoring. Simulations were then performed to determine the optimal placement of just one such sensor in a mold to demonstrate that sufficient information on the mold filling process could be obtained. The purpose of the simulation work was to learn how to interpret the sensor response and, subsequently, use it to control the LCM process.     

Optimizing the production of hydrogen and 1,3-propanediol in anaerobic fermentation of biodiesel glycerol

The conversion of glycerol in biodiesel waste streams to valuable products (e.g. hydrogen and 1,3-propanediol (1,3-PD)) was studied through batch-mode anaerobic fermentation with organic soil as inoculum. The production of hydrogen in headspace and 1,3-PD in liquid phase was examined at different hydrogen retention times (HyRTs), which were controlled by gas-collection intervals (GCIs) and initial gas-collection time points (IGCTs). Two purification stages of biodiesel glycerol (P2 and P3) were tested at three concentrations (3, 5 and 7 g/L). Longer HyRT (longer GCI and longer IGCT) led to lower hydrogen yield but higher 1,3-PD yield. The P3 glycerol at the concentration of 7 g/L had the highest 1,3-PD yield (0.65 mol/mol glycerol_consumed) at the GCI/IGCT of 20 h/65 h and the highest hydrogen yield (0.75 mol/mol glycerol_consumed) at the GCI/IGCT of 2.5 h/20 h), respectively. A mixed-order kinetic model was developed to simulate the effects of GCI/IGCT on the production of hydrogen and 1,3-PD. The results showed that the production of hydrogen and 1,3-PD can be optimized by adjusting HyRT in anaerobic fermentation of glycerol.     

Use of a charged coupled device (CCD) camera for evanescent wave optical fiber cure monitoring of liquid composite molding resins

An optical fiber evanescent wave fluorescence cure sensor described in a previous paper (1) has been interfaced with a fast detector and computer software to provide real-time monitoring of fast-reacting systems. Advantages to the new system include a reduction in data acquisition time (to 200 ms from 3 min), improved time resolution (to 5 sec from 10 min), and improved signal quality. Isothermal cure monitoring measurements were conducted in both distal and evanescent wave sampling modes for an epoxy-amine system. Following a second order background correction, the results for the two sampling modes were identical. Thus the evanescent wave measurement was unable to detect differences between the chemistry in the bulk and at the fiber-resin interface (if there were any differences to observe). The sensor system was also demonstrated for a fast curing (<3 min gel time) polyurethane-isocyanurate system designed for high volume (short cycle time) automotive applications. The system is currently being modified for implementation in a “real world” manufacturing environment.     

Vacuum‐assisted resin transfer molding model

A model of the vacuum‐assisted resin transfer molding (VARTM) process is developed that includes the most important aspects of the processing physics. The model consists of several submodels, such as preform mechanics, Darcy flow, wicking flow, and void formation. The preform mechanics model treats the preform as a linearly elastic, one‐dimensional (1D) solid. However, the key physical process is the lubrication of the preform due to fluid wetting, and this is modeled as a reduction in preform modulus, an easily measurable parameter. Residual stress, three‐dimensional (3D) structural behavior, and nonlinearity are neglected, but can all be included. The fluid flow model of capillary wicking is not tacked onto the Darcy equation as a modified boundary condition, as was previously done. The wicking is treated simply, but more realistically, by performing a force balance on the fluid in a pore. Balancing the capillary pressure and the viscous drag allows the development of a wicking front that precedes the main Darcy flow front to an extent that depends on several easily measurable factors. It is this wicking front that is responsible for the small void formation that reduces the quality of VARTM parts, relative to resin transfer molding (RTM) parts. POLYM. COMPOS. 26:477–485, 2005. © 2005 Society of Plastics Engineers     

Cure monitoring of the liquid composite molding process using fiber optic sensors

Fluorescence has been demonstrated to be an accurate tool for monitoring resin cure. It is measured using an evanescent wave fiber-optic sensor. An economical optical fiber sensor has been developed with a refractive index greater than 1.6, permitting evanescent wave monitoring of epoxy resins. The fluorescence wave-length-shift, which has been correlated with monomer conversion, is monitored during the liquid molding process. Unidirectional glass fabrics with volume fractions from 40% to 60% were injected with epoxy resin at a variety of driving pressures and cured at several temperatures. Several composite parts were fabricated to test the effects of vacuum pressure, injection rate, cure temperature, and fiber fraction on the performance of the sensor. The sensitivity of the evanescent wave fluorescence sensor to the condition of the resin system was also examined. The sets of resin/hardener samples were subjected to rigorous chemical analysis to determine the extent of their differences.     

A permeability database for composites manufacturing

The National Institute of Standards and Technology (NIST) has developed a database containing a set of permeability results from carefully controlled measurements for both saturated and unsaturated flows in glass fabrics. The materials studied are fully documented in the database, and in some cases the material compressibility is also included. The experimental techniques used to generate the permeability data are briefly described and the included permeability data summarized. A sample search is illustrated to demonstrate the database user interface and the extent of the data documentation.     

Influence of processing rate and formulation on the interface strength of vinyl ester/E‐glass composites

The single fiber fragmentation test was used to investigate the effect of gelation time on interfacial shear properties of fast reacting resin systems. We developed a processing system capable of producing single fiber fragmentation samples with gelation times that ranged from 2 min to 45 min. The interfacial properties of E‐glass fibers in vinyl ester resin were measured with single fiber fragmentation tests using a manual and an automated testing machine. We found that vinyl ester resins catalyzed with methyl ethyl ketone peroxide and promoted with cobalt naphthenate and dimethylaniline gelled in less than two minutes and had an estimated interfacial shear strength of 105 MPa. Specimens cured without the promoter gelled in 45 min and had an interfacial shear strength of 72 MPa. Further curing of the unpromoted specimens resulted in an increase in shear strength to 96 MPa. We have demonstrated the ability to make and test rapidly cured specimens, thus expanding the range of materials that can be tested using the single fiber fragmentation testing technique.     

Computer systems availability evaluation using a segregated failures model

This paper presents the segregated failures model (SFM) of availability of fault‐tolerant computer systems with several recovery procedures. This model is compared with a Markov chain model and its advantages are explained. The basic model is then extended for the situation when the coverage factor is unknown and the failure escalation rates must be used instead. A simple practical analytical approach to availability evaluation is provided and illustrated in detail by estimating the availability of two versions of a reliable clustered computing architecture. For these examples, numeric values of availability indexes are computed and the contribution of each recovery procedure to total system availability is analysed. Copyright © 2008 John Wiley & Sons, Ltd.