Novel adhesive systems based on neoprene-phenolic blends

The aim of the investigation is to develop new high performance adhesive systems based on neoprene-phenolic blends. Initially the effect of addition of all possible ingredients like fillers, adhesion promoters, curing agents and their optimum compositions to neoprene solution is investigated. The phenolic resin used is a copolymer of phenol-cardanolformaldehyde prepared in the laboratory. The optimum ratio between phenol and cardanol that gives the maximum bond strength in metal-metal, rubber-rubber and rubber-metal specimens has been identified. Further the ratio between total phenols and formaldehyde is also optimised. The above adhesive system is further modified by the addition of epoxidized phenolic novolacs. For this purpose, phenolic novolac resins are prepared in different stoichiometric ratios and are subsequently epoxidized. The effectiveness of the adhesive for bonding different metal and rubber substrates is another part of the study. To study the ageing behaviour, different bonded specimens are exposed to high temperature, hot water and salt water and adhesive properties have been evaluated. The synthesized resins have been characterized by FTIR , HNMR spectroscopy. The molecular weights of the resins have been obtained by GPC. Thermogravimetric analysis and differential scanning calorimetry are used to study the thermal properties. The fractured surface analysis is studied by scanning electron microscopy. The study has brought to light the influence of phenol/ formaldehyde stoichiometric ratio, addition of cardanol (a renewable resource), adhesion promoters and suitability of the adhesive for different substrates and the age resistance of adhesive joints among other things.

A Reinforcement Learning Approach to Economic Dispatch using Neural Networks

This paper presents a Reinforcement Learning (RL) approach to economic dispatch (ED) using Radial Basis Function neural network. We formulate the ED as an N stage decision making problem. We propose a novel architecture to store Qvalues and present a learning algorithm to learn the weights of the neural network. Even though many stochastic search techniques like simulated annealing, genetic algorithm and evolutionary programming have been applied to ED, they require searching for the optimal solution for each load demand. Also they find limitation in handling stochastic cost functions. In our approach once we learn the Q-values, we can find the dispatch for any load demand. We have recently proposed a RL approach to ED. In that approach, we could find only the optimum dispatch for a set of specified discrete values of power demand. The performance of the proposed algorithm is validated by taking IEEE 6 bus system, considering transmission losses