Efficient and accurate fuel management systems are becoming increasingly important as energy needs and operational complexity increase in various industrial sectors. This study aims to design and implement a fuel flow monitoring and control system using a variable control valve. This system allows real-time monitoring of the level and flow of fuel in the service tank on the ship, as well as automated control. In its implementation, ultrasonic sensors, control valves are combined with LM2596 connected to Arduino nano for data collection, analysis and visualization. The test results show that this system is able to provide accurate data on the flow and level of fuel in the service tank on the ship by displaying the measurement results on the LCD and providing warnings via a buzzer. The LCD display shows the reading of the fuel level and valve opening in the tank. At a fuel level of 0 - 25% the valve will open 100% and requires an average filling time of 25 seconds, At a fuel level of 25% - 50% the valve will open 75% and requires an average filling time of 27 seconds, At a fuel level of 50 - 75% the valve will open 50% and requires an average filling time of 30 seconds, At a fuel level of 75 - 99% the valve will open 25% and requires an average filling time of 32 seconds.

The storage space for CO2 gas cylinders is inside the ship's accommodation, so there is a risk of danger if a leak occurs because the ship's accommodation has poor air circulation. This research is devoted to removing dangerous gases from the room. This research designs and modifies a tool that can detect CO2 levels and can provide a danger signal to the surroundings. This modification uses a Raspberry Pi Pico W microcontroller. This research method uses system design, a series of tools with wiring design and uses 2 test plans, namely: testing static and dynamic testing.  Testing the precision of the sensor by reading the CO2 leak detector using a CO2 measuring instrument obtained an average error value of 3.5%. The error value is still categorized as a safe difference value and the prototype works according to the expected function. Testing the function of the tool results in values ??for the MQ-135 buzzer and exhaust fan sensors which will turn on if the CO2 level is above 600 ppm in accordance with the safe threshold for CO2 levels in the air which has been set in the program. Telegram request bot testing can work well as expected. Testing the Telegram Receive bot can work well as expected. However, for testing Telegram bots, requests and receives are influenced by the internet network.