I.B OWUNNA

This paper critically examines the emerging field of blind computation in AI machine operations, challenging conventional approaches to data privacy and security in artificial intelligence systems. As AI continues to permeate various sectors, from healthcare to finance, the need for robust privacy-preserving techniques has become paramount. Blind computation offers a promising solution by enabling AI models to process encrypted data without decryption, thus maintaining data confidentiality throughout the computational pipeline. This research synthesizes cutting-edge developments in homomorphic encryption, secure multiparty computation, and federated learning, presenting a comprehensive framework for implementing blind computation in AI systems. We propose novel architectures that significantly enhance data protection without compromising computational efficiency. Our analysis reveals that while blind computation techniques offer unprecedented levels of privacy, they also introduce new challenges in terms of computational overhead and model accuracy. We present empirical evidence demonstrating the trade-offs between privacy, performance, and precision, and propose innovative strategies to optimize these competing factors. Furthermore, we critically assess the ethical implications of blind computation, examining its potential to either mitigate or exacerbate existing biases in AI systems. This paper concludes by outlining a roadmap for future research, emphasizing the need for interdisciplinary collaboration to address the technical, ethical, and regulatory challenges associated with blind computation in AI. Our findings have significant implications for the design and deployment of privacy-preserving AI systems across various domains, potentially revolutionizing the way sensitive data is processed in the age of artificial intelligence

The propeller shaft is a critical component in marine propulsion systems, transmitting power from the engine to the propellers. In twin-screw passenger ferries, failure of the propeller shaft can lead to severe operational disruptions, safety hazards, and costly repairs. This study presents a comprehensive failure analysis of a propeller shaft from a twin-screw passenger ferry to determine the root cause of failure and suggest mitigation strategies. The investigation includes visual inspection, non- destructive testing (NDT), metallurgical analysis, and finite element analysis (FEA) to evaluate material properties, fatigue characteristics, and stress distribution. The findings indicate that fatigue failure, corrosion-assisted cracking, misalignment, and improper lubrication are potential contributing factors. Based on the results, recommendations for improved maintenance, material selection, and design modifications are proposed to enhance the reliability and longevity of propeller shafts in marine vessels. This study provides valuable insights into preventing similar failures in future maritime applications.

In Nigeria today and in the world at large, PET bottle waste has grown to become hazardous as it constitutes part of the non-biodegradable waste. Hence, recycling becomes necessary to curb its menace. This project work is centered on designing and fabricating a PET bottle crushing machine from locally sourced materials for both home and industrial use in an attempt to proffer solution to the PET waste problem in Nigeria. Preliminary tests and mechanical factors were extensively evaluated on the conceptual designs to ensure that the design that most suits the purpose was selected and detail design was carried out. Experiment to determine the power required to overcome the shear resistance of the PET bottles was carried out and it was discovered that 10hp at 450N was the power required. Finite element analysis was also performed on the cutting blade to inspect
the materials response to stresses and the corresponding deformation. Furthermore, a design study was carried out in order to ascertain the minimum and maximum loads that can be handled. Tests carried out on the machine showed its efficiency to be 82.2% which is only 6% less than the efficiency of foreign counterparts