AGCO Power is constructing higher-quality engines with the help of cloud-based analytics and AI platform
Insta is developing a data lake-based analytics and AI platform in cooperation with AGCO Power to process and utilize data from diesel engines. It enables AGCO Power’s R&D team to systematically review vast amounts of data and capture details about the operation of field test engines that previously went unnoticed. This way, engines are more stable when they are moved to serial production and operate more steadily under varying conditions. This is visible to end customers in improved quality. At the same time, it decreases AGCO Power’s warranty-related costs.
AGCO Power develops and builds engines for the machines of the world’s leading agricultural and work machine manufacturers. AGCO Power’s engines are famous for their durability and long service life. For the business and brand, it is critical to ensure that the products function properly under demanding field conditions and in actual work situations around the world.
A field test entails a test-drive period of up to a couple of years under authentic conditions, during which AGCO Power makes sure that a new-generation engine works as planned and meets the requirements set for it. Extensive information on the regular operation, fault situations and customized test drives of the engines being field-tested are needed for the R&D team’s use. Earlier, AGCO Power had to send their own or the manufacturer’s representative to the machine in fault situations to investigate and measure the operation of the engine, which cost money and the specialists’ time. Data loggers that stored data on a memory card were in use, but their data was not continuous or consistent. A systematic process and harmonized procedures were missing.
To solve the challenges of data transfer and processing, the Advanced Analytics team of Insta is developing a data lake and analytics platform for AGCO Power that stores all the relevant data from the engines and makes it available to the product development personnel practically in real time.
“Let’s assume a situation where an engine shuts down due to a fault. Earlier, we did receive a report about it, but we did not get any recorded data on the event. If an engineer was unable to reproduce the situation, it was impossible for us to find out what exactly had happened and why. Now, if an anomaly occurs, all the data is stored in our analytic system and engineers can fetch the data they need from there. We can systematically investigate fault situations and more easily find their causes”, says Jouko Järvinen, project manager at AGCO Power.
“It has come as a complete surprise to me how much our people use the system. Such a tool should have been created a long time ago.” – Jouko Järvinen, AGCO Power
A huge amount of data and metadata about engines is stored to enable data-driven analysis. AGCO Power has dozens of test machines. Hundreds of variables are stored for each of them per second, and a test of a few years is performed with each of them.
“Data enables analyzing engines statistically and using modern machine learning and neural network approaches for spotting anomalies in operations. Validating all data manually would not be possible," says Aki Pajunoja, Analytics Team Lead at AGCO Power.
Use and operation of the system are expanding
At the moment, AGCO Power has several data scientists making use of the analytics platform to create new kinds of models on top of the gathered data. At the same time, more data sources are being integrated on the data lake platform with Insta so that the benefits of the system could be expanded to other parts of the organization.
“Originally, the tool was only used by our product development specialists. We are finding many new ways to utilize the same platform through the entire engine life cycle” explains Aki Pajunoja.
“We picked the right partner for this. Not everyone would have managed it.” – Jouko Järvinen, AGCO Power
Read more about our industrial digitalization services by clicking: