ABSTRACT: 

This research represents an evaluation of the performance of multi-stage solar distiller water designed and manufactured in Baghdad, Iraq. All experiments were conducted at the Scientific Research Commission in Baghdad. Solar radiation and temperatures were recorded in specific locations in the solar distiller, whose glass cover was tilted 20 degrees from the horizon. All experiments were completed under the climatic conditions of Baghdad for three days in September and the other days in October and November 2024. The solar distiller basin was filled with 4 liters of salty water. The water produced from the solar distiller was collected in water-scaled containers. The highest productivity of distilled water was obtained on 18 September with a volume of (1890 ml), with a connecting solar spiral collector increasing by 92.86 % when it was (980 ml) without a connecting solar spiral collector which adds energy 770.292 KJ. In these experiments, the distilled water produced from the solar distiller was analyzed. Chemical analyses showed Total Dissolved Solids (TDS) changed from 900 to 25.

Keywords: Solar distillation; pre-heating; solar spiral collector; Solar Thermal Insulation; evaporation.

ABSTRACT: 

This study develops and evaluates an intelligent Insulin Dosage Calculator powered by a Random Forest machine-learning model to deliver personalized insulin dosage recommendations for individuals with Type 1 and insulin- dependent Type 2 diabetes. The system addresses the limitations of traditional insulin dosing methods, which rely on complex insulin-to-carbohydrate ratio (ICR) and insulin sensitivity factor (ISF) calculations prone to errors. A comprehensive dataset, including blood glucose levels, carbohydrate intake, insulin doses, physical activity, and other health metrics, was preprocessed and used to train the Random Forest Regresses. A parallel group design compared the machine learning approach against traditional methods across control (n=50), experimental (n=50), and crossover (n=50) groups. The experimental group achieved a significant improvement in Time in Range (TIR) (+9.3% vs. +2.7% in the control group) and a greater reduction in average blood glucose levels (-23.1 mg/dL vs. -5.5 mg/dL). The model’s Root Mean Square Error (RMSE) was 0.78 ± 0.14, compared to 1.35 ± 0.22 for traditional methods, indicating superior accuracy. Participants rated the system highly for usability (4.4 ± 0.5 vs. 3.1 ± 0.6). Despite limitations such as a small sample size and short study duration, the findings highlight the potential of machine learning to revolutionize diabetes management. Future research should focus on long-term efficacy and integration with wearable technologies.

Keywords:  Insulin Dosage, Random Forest, Machine Learning, Diabetes Management.

ABSTRACT: 

Radioactive leaks are a serious threat to human life and the environment. They may occur as a result of wars, natural disasters such as earthquakes or tsunamis, or human errors inside nuclear facilities. These incidents can lead to dangerous health effects, long-term contamination, and forced displacement of communities. The current monitoring systems often rely on fixed sensors and manual response, which may be too slow or limited during emergencies. This study presents a theoretical vision for using artificial intelligence (AI) to enhance the early detection of radioactive leaks. The idea is to design a smart monitoring system that uses AI algorithms to analyze radiation sensor data in real time, detect unusual patterns, and provide early warnings before major damage occurs. The suggested approach includes combining machine learning models with environmental sensors, drones, or satellite data to create an intelligent and adaptive system capable of functioning even in difficult or high-risk conditions. This theoretical model highlights the potential of AI to improve nuclear safety and reduce the impact of radiation on human health and ecosystems. Although this study does not apply the model to real-world data, it lays the foundation for future research and development in this area. The integration of AI in radiation monitoring represents a promising step toward better preparedness and faster response in the face of nuclear accidents, especially in times of war or natural disasters.

Keywords:   Artificial Intelligence, Detection of Radioactive.