“No one needs transport monitoring. We want complex solutions applicable in the industry as well as in the agricultural sector. ” Leading European companies now regard predictive analytics as a valuable tool for saving money and avoiding million-dollar spending. World manufacturers also see great potential in such systems and it is highly possible that integrators will soon start receiving orders for predictive analytics projects implementation. We’ve put together several examples of smart predictions to warm up your interest in this topic.
Forecasting systems are of particular importance, due to the high error cost in mechanical engineering, power generation, or metallurgy. Based on historical data and its comparison with normal indicators, tools allow us to build a forecast of equipment condition, create a model of its behavior, and assess the possibility of failure or accident. So, this information will help companies to move away from the troubleshooting model to preventing potential problems.
1K vs 4M$
At one of the American power plants, system sensors recorded a slight vibration on the steam turbine shaft. The forecasting system in the monitoring department provided a list of possible causes of abnormal vibration. Among them, there was a defect in the turbine blade, which collapsed and caused an imbalance in all rotating parts of one axis. Its price doesn’t exceed one thousand dollars. The check-up confirmed the systemic prognosis and the defected item was fixed on time. According to experts, early response saved up to $ 4 million.
According to McKinsey international consulting company, such a preventive maintenance model reduces equipment downtime by 30-50% and increases its service life span by 20-40%.
How could Galileosky equipment be integrated into such a project? The main element in the system is the GPS unit, which will transmit vibration level data to the server. Moreover, measurement accuracy will depend on the vibration sensor. An obvious question is: will the system feel such a slight imbalance? If you choose a sensor with high sensitivity, then the unit will definitely be able to indicate the anomaly and warn the staff about a potential malfunction.
A couple of minutes instead of an hour
Emergency equipment repair is one of the most important, but not the only reason for financial losses. Overall efficiency is directly affected by the product quality and conveyor performance. Predictive analytics could be applied in these processes as well.
Several years ago, Mercedes-AMG used a predictive system to cut engine test times by 94%. Test runs typically reveal most engine defects and malfunctions within several minutes after starting. However, specialists had to wait for tests to finish (more than an hour) to analyze all the results.
The solution, which included a set of sensors and software, monitored engine behavior in real-time and allowed testing to be completed well ahead of schedule. As a result, quick malfunction diagnostics and generated reports appeared 3-5 minutes after the test started. This approach has reduced time resources and increased the accuracy of fault identification.
Galileosky’s specialists could suggest the following scenario: first of all, connect to engine CAN-bus, use CAN Scanner tool for engine quick diagnostics. In order to check additional parameters, use vibration sensors that can record abnormal vibrations during misfire and temperature sensors connected to the unit through thermal grease. An Easy Logic algorithm can be created up for the described peripherals, where the reaction to deviation from normal parameters will be malfunction recording in the final report.
Temperature and pressure are under control
In 2019, a well-known energy company in the United States conducted an experiment where additional sensors were connected to several heat exchangers. These are larger analogs of car radiators, made to cool oil and other fluids as they circulate through them. They are widely used in oil and gas processing, so temperature and pressure in such systems are high and the accidents can be fatal.
Four wireless sensors were installed to provide a broader set of data, including temperature level and oil consumption. Thanks to special software, predictive analytics have been configured to predict when the heat exchanger becomes dirty and there is a need to clean it before it is completely clogged.
There is an alternative solution: use Easy Logic technology as a predictive element. Create an algorithm that will generate a report and give the necessary reactions. Conditions such as operating temperature, the pressure within acceptable parameters, and the oil flow rate can be processed into the algorithm and response can be based on historical data and the current situation.
Measure the level in time and pump out water
While extracting raw materials from quarries, depending on the terrain and season, there is a possibility of flooding. So, water is being pumped regularly and monitoring systems control the volume of water pumped. However, during floods or heavy rains, there is a real danger of rapid flooding in a few hours.
To avoid such situations, use water level sensors in the surrounding water bodies, soil moisture sensors, measure the groundwater level in the wells, analyze temperature and weather forecast, set up an early warning system about an impending water level rise.
How to put this into practice? Connect Galileosky device to soil moisture or water level sensor. In order not to overload the GSM network for regular data transfer, put the device in the Hub to Hub mode. Then all devices outside the GSM network will be able to transmit data to one device, which regularly visits the coverage area.
Now predictive analytics is a necessary tool that helps to maintain equipment, increase its efficiency, and forecast demand. All of the described solutions can be implemented on Galileosky equipment. The main issue is in the software and thorough peripherals selection, as the reliability of the database and calculations credibility depend on sensors.