In this research paper, the effect of seismic load for overhead water tank, the water is the most essential element to a life on the earth. There is two different types of tank, of same capacity but having different height and different diameter of the overhead water tank in seismic zone (zone IV and V), this study on displacement of tank, time period, base shear and base moment cofficient of different height and diameter of overhead water tank, seismic plays an important in design of tank structure because of dynamic nature. Effect of seismic is predominant on tank structure, height of the struture in this paper the compression of the tank their height and its different diameter on tank for analysis of seismic loads on tanks

1PG Student, Dept of EEE (PS), GCET, Kadapa, AP, India. 2Associate Professor, Dept of EEE, GCET, Kadapa, AP, India. A Cuk converter based EV (Electric vehicle) battery charger is designed and developed in this work. It supplies affordable as well as high-power density-based billing service for EV. This charger includes much less variety of tools running over one switching cycle, which minimizes the additional transmission loss incurred by a diode bridge rectifier of conventional battery charger and hence, enhances the charger effectiveness. Throughout constant present and also consistent voltage regions, the commands for battery charging are synchronized by a flyback converter. The included benefit of recommended geography is that the undesirable capacitive combining loophole is eliminated, as well as unwanted conduction via the body diode of non-active button in formerly created BL Cuk converter is prevented. This substantially boosts the charger effectiveness. For the consistent present and also continuous voltage (Curriculum Vitae) charging, the commands are synchronized by a flyback converter. The suggested charger is examined to demonstrate the improved power quality. Test results validate the better performance of the recommended battery charger.

This paper describes the control of single-phase grid-coupled solar photovoltaic (PV) power generating system with universal active power filter (UAPF) capacities. The SPVUAPF system includes series and shunt voltage converters. The shunt VSC exports the actual electricity removed from the PV panels to the grid similarly to furthermore near-by way of entire loads. Along with searching after the real electric strength, the shunt VSC offers charge of responsive in addition to harmonic currents created the use of the masses. The recommendation caution symptoms and signs required for the control of the shunt further to additionally series VSCs of the SPV-UAPF device are approximated utilizing least advice consolidated modern-day (LMMN) adaptive acknowledgment components. The overall performance of the system with series shunt recompense abilities are demonstrated using MATLAB/Simulink software in different conditions like irradiance variation, voltage droop and swell together with present harmonics.

Community detection is a fundamental task and an important area of research, aimed at partitioning a network into multiple substructures. This process helps uncover the underlying functions or causes behind large, complex networks, which has garnered significant attention within the scientific community. Numerous studies have been published on this topic. The primary goal of community detection is to identify subgraphsreferred to as "communities"in which nodes are more densely interconnected with each other than with other parts of the network. Nodes within these communities are expected to share similar features, indicating a common role or property. Classic approaches to community detection often use probabilistic graphical models that statistically infer the most likely community structures. These methods also integrate prior knowledge to guide algorithms and enhance the accuracy of the detected partitions. By using these techniques, researchers and data scientists can transform a convoluted web of connections into a clear map of meaningful communities, providing critical insights across various fields, including social media analysis and biological network studies. Community detection has vital applications in various areas of social networking, such as sociology, business, criminal detection systems, and recommendation systems. As network data becomes increasingly complex, addressing challenges related to computational efficiency due to the large size and dynamic nature of social networks is essential. Despite substantial advancements in the field, there remains a significant gap: a comprehensive understanding of the theoretical and methodological foundations of community detection. Bridging this gap is crucial for unlocking the next generation of breakthroughs in network analysis. This article provides a comprehensive overview of existing community detection methods and introduces a new taxonomy that categorizes these approaches based on network substructures, mining attributes, and interaction systems. We classify the methods into several fundamental categories: probabilistic graphical models, deep learning-based approaches, node content-based methods, edge content-based methods, and hybrid methods that incorporate both node and edge content. Each method is examined in detail, highlighting its core principles and applicability to social network data. The discussion emphasizes the strengths and limitations of various techniques, offering readers a clear understanding of the current landscape in community detection. Finally, we address key challenges in the field and provide actionable suggestions for future research, aiming to inspire advancements in scalable, accurate, and interpretable community discovery. This survey serves as a valuable resource for researchers and practitioners working with complex network data.

a near-ring which is both left and right weakly regular is called weakly regular. Kovace [1956] proved that, for commutative weak regularity and regularity are equivalent condition. This paper attempts to generalize the properties of regular near-ring taking the condition of weakly regular near-rings.

Improvement of the fact at which patients survival and treatment results break is that which we see out of very early detection of breast cancer. Presently we see that traditional diagnostic methods do in to play to that of human subjectivity, tiredness as a factor in performance and also variable interpretation which is a large issue because they mostly use radiology and pathology for evaluation. These issues put into light the need for data based, automated approaches which in turn will give out very reliable and unbiased results. To have an objective performance evaluation we pre-processed the breast mass image data which included features from fine-needle aspirates. We used the same settings for training and testing each model to ensure evaluation consistency. We found that ensemble models which include Random Forest and Gradient Boosting performed the best out of the studied algorithms with over 98.8% accuracy. Also, these models had better recall and AUC which in turn showed their performance in differentiation between benign and malignant cases. Also, we saw that which statistical analysis, confusion matrices, and descriptive visualizations. What we found is that ensemble-based machine learning does in fact out perform traditional models in the case of breast cancer which we put forth as a reliable and scalable solution for early detection. In health care settings this type of predictive modelling may improve the accuracy of cancer screen out comes, reduce in diagnostic errors, and support clinical decision making.

Problem to detect the Breast Cancer timely and with accuracy is one of the most important diseases in modern health care industry. We have a lot of algorithms related to machine learning shows accurate results but due to data imbalance and other problems in data there is fluctuation in the results. In this paper, the combination of different ml algorithms has been used as a hybrid assembly for better results, enhancing the separate results of these algorithms and correcting the drawbacks of such procedures. The important factor is to increase the robustness and generalization of model so that the consistency in prediction remains. So the hybrid model is used in Biomedical Data Analysis. I compared the hybrid with classical ml models. And I got 99.3% accuracy that shows the better performance than others. Visualization charts prove the diagnostic capacity of hybrid model. The results shows that when we work on different methods together then prediction and detection of breast cancer can effectively complete This research provide a feasible solution for the applications of diagnosis.

In this paper, we introduce the notion of an element -primary to another ele- ment in a multiplicative lattice L. We obtain its various characterizations. We investigate many of its properties. 2020 Mathematics Subject Classification: 06B99

PDFs are widely used for sharing academic, techni- cal, and professional content, but extracting relevant information from lengthy documents remains difficult. Traditional keyword searches miss the semantic depth of queries, resulting in irrel- evant information. This paper presents DocMind, a Retrieval- Augmented Generation (RAG) framework for conversational question answering on PDFs. The system integrates semantic document parsing and chunking, embedding generation, and vector database retrieval (FAISS/Chroma) for efficient semantic searching.For each users query, we embed it within the same vector space as the users query. The most relevant chunks of text are identified, and the information is supplied to a Large Language Model (LLM) to generate a coherent answer. Unlike traditional RAG systems which produce static answers, DocMind is capable of multi-turn conversational memory and provides fallback clarity by informing the user when an answer is provided without supporting PDF data. In our experimental evaluation, we showed that DocMind significantly increased the baseline models using BM25 and TF-IDF in recall, F1 score, and ROUGE-L metrics due to improvement in contextual accuracy and hallucination mitigation. By merging semantic retrieval with generative reasoning, DocMind offers dependable, interactive, and domain-agnostic QA on PDF documents, greatly improving information accessibility and reliability. Index TermsRetrieval Augmented Generation, Vector Database, FAISS, Chroma, Conversational QA, PDF Processing, Large Language Model

The 4D drug delivery system is a revolutionary approach to pharmaceuticals, leveraging advanced technologies like 3D printing and smart materials to create devices that change shape or release drugs in response to specific stimuli. This system enables personalized and targeted therapy, improving treatment outcomes and patient compliance. The design and development of 4D printed devices involve careful consideration of materials, shape, and size, as well as simulation and modeling to predict behavior. The 4D drug delivery system has various applications, including cancer treatment, diabetes management, and tissue engineering. Stimuli-responsive materials, such as pH-responsive, temperature-responsive, and light- responsive materials, play a crucial role in these devices. The system also raises regulatory challenges, scalability concerns, and biocompatibility issues, which must be addressed. The future of 4D drug delivery systems looks promising, with potential integration with artificial intelligence, nanotechnology, and other technologies. Personalized medicine and tissue engineering are key areas where 4D printed devices can make a significant impact. As research advances, 4D drug delivery systems are expected to transform the pharmaceutical industry, enabling more effective and efficient treatment options for patients.

Smart farming systems increasingly depend on Internet of Things (IoT) technologies to monitor field conditions and support timely decision-making. However, when communication channels are not properly secured, sensitive agricultural data can be intercepted, altered, or exposed, undermining both privacy and system reliability. In this study, a secure channel encryption model is designed, implemented, and experimentally evaluated for an IoT-enabled smart yam farming environment. A real- world smart farm testbed was deployed at the Federal University of Technology, Minna, Nigeria, where microcontroller-based sensor nodes were used to collect soil and environmental data and transmit them to cloud infrastructure. The proposed security framework applies AES-128 encryption with Cipher Block Chaining (CBC) mode and incorporates initialization vectors, salting, padding, and message authentication to safeguard device-to-cloud communication. Performance evaluation compared CBC and ECB cipher modes using different key generation mechanisms. The results show that AES-128/CBC increased the data size from 110 bytes to 176 bytes (approximately 60% overhead), while producing more randomized ciphertext and eliminating the pattern leakage observed with ECB mode. Although CBC required about twice the encryption time of ECB (average encryption time of 3.2 10 ns), decryption remained efficient, with times below 1.3 10 ns. Overall, the findings confirm that the proposed approach enhances data confidentiality and integrity with a level of computational overhead that is acceptable for periodic smart farming data transmission. The key contribution of this work is the delivery of a practical, deployable, and security-focused channel encryption solution validated using live farm data rather than simulations, demonstrating that strong cryptographic protection can be realistically achieved in resource-constrained agricultural IoT environments.