Datta Meghe College of Engineering

Understanding the complex interaction between deep foundations and layered soil profiles is critical for the seismic design of high-rise structures. This study conducts a nonlinear dynamic soil pile structure interaction (DSPSI) analysis of a 20-storey reinforced concrete building supported by layered sandy soil, emphasizing the influence of nonlinear soil pile interface behaviour. A detailed three-dimensional finite element model is developed, incorporating three structural configurations: a regular frame and two vertically irregular frames featuring upper-storey geometric discontinuities. The stratified sand deposit is modelled with depth-varying stiffness, and horizontal damping layers are embedded to capture energy dissipation mechanisms. Two pile configurations, series and 2 × 2 grouped arrangements, are examined across two length-to-diameter (L/D) ratios, maintaining a consistent pile spacing of 3D. The study evaluates key seismic response parameters, including lateral deformation, storey drift, base shear, lateral pile displacement, and settlement. The results demonstrate that nonlinear interface modelling significantly alters seismic response trends and enhances the accuracy of performance predictions. These findings underscore the importance of advanced nonlinear interface modelling and optimized pile configurations in improving the seismic resilience of high-rise buildings founded on sandy soil.

The treatment of cancer is individual for each patient, and can be personalized, as individual reactions to treatment depend on many factors such as tumour characteristics, stage of disease, level of toxicity and clinical history. The research paper is entitled "Reinforcement Learning for Optimizing Personalized Cancer Treatment Protocols" and suggests the use of Safe Offline Reinforcement Learning (RL) approach based on Conservative Q-Learning (CQ) for designing cancer treatment plans for each patient. The model is developed in python and trained using past treatment records of patients without conducting any unsafe experiments in real-time. To minimize over-estimation of the uncertainty of treatment actions and to allow for more conservative recommendations in observed treatment patterns, we chose to use the conservative version of Q-Learning. We chose to use the conservative version of Q-Learning to reduce overestimation of the uncertainty of treatment actions and to allow for more conservative recommendations in observed treatment patterns. The envisioned framework takes into account a number of goals such as tumor reduction, toxicity control, and improvement of survival and minimization of unnecessary burden of drugs. The method has been designed to offer more reliable and patient-specific treatment policies, when compared with conventional methods and other reinforcement learning method. This study shows that safe offline reinforcement learning can be a valuable clinical decision support tool to optimize personalized oncology care.

The separation of 2,2-dimethoxypropane (DMP) from reaction masses is an important process in chemical manufacturing, particularly in the combination of fine chemicals and pharmaceuticals. This review critically evaluates various separation techniques employed to isolate DMP from complex reaction mixtures. It encompasses a comprehensive analysis of traditional approaches such as concentration and liquid–liquid extraction, highlighting their efficiency, limitations, and suitability based on the physical and chemical properties of DMP. The review also addresses the integration of novel technologies, such as membrane separation and advanced adsorption materials, which offer promising solutions to existing challenges in separation efficiency and cost. Comparative analysis of these methods reveals trends towards more sustainable and cost-effective approaches, with a focus on reducing environmental impact and improving process scalability. Special emphasis is placed on recent innovations and improvements in separation methods, such as the use of novel adsorbents and catalysts, which contribute to higher purity and yield of DMP. The review addresses challenges associated with the separation process, including the impact of impurities, reaction conditions, and economic considerations. Comparative performance metrics and case studies from recent literature are discussed to provide insights into the practical application of these techniques. The review aims to offer a detailed understanding of the current state of DMP separation technologies, guiding researchers and industrial practitioners in selecting and optimizing appropriate methods for their specific needs.

The paper presents comprehensive laboratory studies to evaluate the performance of a flexible pavement system incorporating stabilized marine clay (MC) subgrade and a cement-treated sub-base (CTSB). Waste materials used in the study include quarry dust (QD) for subgrade stabilization, and fly ash (FA) and granulated blast furnace slag (GBFS) for the CTSB layer. Laboratory tests on subgrade mixes included the standard Proctor test, California bearing ratio (CBR) test, and direct shear test, while sub-base mixes were evaluated using the standard Proctor test, CBR test, unconfined compression test, and permeability test. The laboratory evaluation identified an optimum mix of 75

The present research aims to investigate the performance of self-compacting geopolymer concrete (SCGC) with sustainable binder materials such as Rice Husk Ash (RHA) and Nano Clay (NC). The main precursors used were fly ash (FA) and ground granulated blast furnace slag (GGBFS), and RHA was substituted for FA in proportions ranging from 0 to 30% to enhance the performance of the geopolymer matrix. The selected RHA mixture was further modified by changing the amounts of NC to obtain the desired results. Fresh, mechanical, durability, and microstructural properties of developed mixtures were assessed. The results indicated M-3 (55% FA+ 30% GGBFS + 15% RHA) as an optimum ratio for RHA and the addition of 2% NC (NC-2) further improved the mechanical performance of SCGC. The results indicated that the use of RHA to enhance the performance of these materials was effective up to a certain replacement rate, after which it gradually decreased. The inclusion of NC enhanced the performance by contributing to pore refinement and creating a more compact matrix. The chosen NC composition exhibited enhanced mechanical strength, reduced water absorption, reduced chloride ion ingress and sulfate exposure, and improved internal quality. SEM observations showed a rather dense and firmly bound geopolymer matrix. The selected NC-2 mix had a 28-day compressive strength of 50.92 MPa and water absorption of 3.20%. The chloride ion penetration and the sulfate strength deterioration at 90 days were 720 C and 5.84%, respectively. The findings demonstrate the potential of combining RHA and NC to produce sustainable and durable SCGC.