Terabyte-scale data platform enabling Valmet to provide data-driven services to its manufacturing customers
"I would recommend working with Kaito to whoever expects the consultants to learn their industry fast. "
My name is Arttu Matinlauri and I'm the director of analytics and applications here at Valmet.
We at Valmet Industrial Internet develop data platforms as well as analytical applications for the manufacturing industries. Our customer segments include pulp, paper, board, tissue, energy, and marine. Some other industries as well, like oil and gas, are coming from our automation business background.
When Valmet Industrial Internet was started some years ago, we were one of the world's first to start leveraging IoT and big data in this scale in the field of manufacturing.
Actually, it has been an educational journey. We have tried solutions already existing on the market, tested them out, identified how their performance is or actually is not adequate enough for our purposes, and based on those gathered matters we have developed the leading system on the market for leveraging process data coming off of our customers equipment.
Our journey together with Kaito started roughly five years ago.
They've been part of our first proof of concepts, but also testing out the different technologies. Together with Kaito we've been able to reflect continuously upon what technologies exist on the market, how do they fit our process data purposes. Also together with Kaito we've been able to identify technologies that we have to remove from our technology stack to really make it more data and cost efficient. I really find that the collaboration has been truly a partnership, I actually feel that Kaito is a sub team within our whole unit and not a partner consultant per say.
Over the time we've had to grow both our own team as well as our subcontracting team.
What expectations that has set for our partners, is that they also have had to educate our own Valmet people. Kaito has been very supportive in growing our competencies for example in the database side, in efficient data warehousing and data processing.
We are running our cloud platform in Amazon Web Services to receive the data files that we're getting. After that we are processing the data into our Snowflake data warehouse where we are managing the overall context and semantics of the data. From Snowflake we take it to our own database management systems which we have created ourselves, and our user interfaces that we also program. That allows us to really bring up the data fast and efficient toward customers’ purposes.
On a daily basis, we are already connected to hundreds of customer mills and plants across our customers’ industries. That means we are collecting hundreds of billions of rows of data on a daily basis. We're not only collecting it, obviously that data is being then calculated through our AI and machine learning algorithms to provide further insights for our customers on how to adjust how they're operating, or for example to bring up health scores on their equipment condition.
For example one of the solutions that we have developed here at Valmet, is looking at optimizing the way a paper and board machine is being run. We are able to identify energy savings up to 15 - 20% in just adjusting the way the machine is being operated.
Scaling that up is something we are looking to do next. Think about it, if on a single machine you get 15 to 20% energy savings, it translates directly into emission savings, really making a global impact, really making the whole world a greener place to live in.
When I started in this position, Valmet Industrial Internet was having over 10 partners working with them. We decided to reduce the number of partners to improve our efficiency and collaboration. In that process Kaito was selected specifically to be our partner in the domain of data engineering, as well as analytics and visualizations.
As we run a multiple vendor partnering scheme, Kaito has had to collaborate also with our other partnering companies. We have found that to be very efficient. Kaito works as they would be people actually working for Valmet. That is something we appreciate and something we look for in all of our partners, now and in the future.
