Surface roughness parameter prediction and evaluation are important factors in determining the satisfactory performance of machined surfaces in many fields. The recent trend towards the measurement and evaluation of surface roughness has led to renewed interest in the use of newly developed non-contact sensors. In the present work, an attempt has been made to measure the surface roughness parameter of different machined surfaces using a high sensitivity capacitive sensor. A capacitive response model is proposed to predict theoretical average capacitive surface roughness and compare it with the capacitive sensor measurement results. The measurements were carried out for 18 specimens using the proposed capacitive-sensor-based non-contact measurement setup. The results show that surface roughness values measured using a sensor well agree with the model output. For ground and milled surfaces, the correlation coefficients obtained are high, while for the surfaces generated by shaping, the correlation coefficient is low. It is observed that the sensor can effectively assess the fine and moderate rough-machined surfaces compared to rough surfaces generated by a shaping process. Furthermore, a linear regression model is proposed to predict the surface roughness from the measured average capacitive roughness. It can be further used in on-machine measurement, on-line monitoring and control of surface roughness in the machine tool environment.

Freeform surfaces have wider engineering applications. Designers use B-splines, Non-Uniform Rational B-splines, etc. to represent the freeform surfaces in CAD, while the manufacturers employ machines with controllers based on approximating functions or splines. Different errors also creep in during machining operations. Therefore the manufactured freeform surfaces have to be verified for conformance to design specification. Different points on the surface are probed using a coordinate measuring machine and substitute geometry of surface established from the measured points is compared with the design surface. The sampling points are distributed according to different strategies. In the present work, two new strategies of distributing the points on the basis of uniform surface area and dominant points are proposed, considering the geometrical nature of the surfaces. Metrological aspects such as probe contact and margins to be provided along the sides have also been included. The results are discussed in terms of deviation between measured points and substitute surface as well as between design and substitute surfaces, and compared with those obtained with the methods reported in the literature.

Today’s electricity management mainly focuses on smart grid implementation for better power utilization. Supply-demand balancing, and high operating costs are still considered the most challenging factors in the smart grid. To overcome this drawback, a Markov fuzzy real-time demand-side manager (MARKOV FRDSM) is proposed to reduce the operating cost of the smart grid system and maintain a supply-demand balance in an uncertain environment. In addition, a non-linear model predictive controller (NMPC) is designed to give a global solution to the non-linear optimization problem with real-time requirements based on the uncertainties over the forecasted load demands and current load status. The proposed MARKOV FRDSM provides a faster scale power allocation concerning fuzzy optimization and deals with uncertainties and imprecision. The implemented results show the proposed MARKOV FRDSM model reduces the cost of operation of the microgrid by 1.95%, 1.16%, and 1.09% than the existing method such as differential evolution and real coded genetic algorithm and maintains the supply-demand balance in the microgrid.