Royal Vopak N.V. is a leading independent tank storage provider for the oil and chemical industries, operating 67 terminals in 25 countries. Chris Sheldon, Terminal Manager for the company’s operation in Savannah, GA, recently oversaw the implementation of an Internet of Things deployment that is enabling the terminal to minimize energy usage, which should lead to significant cost savings. Sheldon shared the story with Network World Editor in Chief John Dix.
Let’s start with a thumbnail description of Vopak.
We are a liquid bulk third-party terminalling business. We store liquids for customers -- chemicals, tropical oils, biodiesels, asphalt, many different types of commodities -- and distribute them by truck, rail, pipeline and vessel. We’re headquartered in the Americas out of Houston, Texas and globally out of Rotterdam, the Netherlands.
Here at the Savannah terminal we have 54 tanks in four tank farms. A little over half of those are insulated, meaning that we have product in them that we heat and agitate. The total capacity of the facility is 1.4 million barrels. Our largest tank is 80,000 barrels and our smallest is 5,000 barrels.
What problem were you trying to solve when you went looking at Internet of Things technology?
It started with our global innovation group, which is out of the Rotterdam office. They had contacted IoT company Atomiton and visited the terminal in Savannah last year to see if IoT made sense to help us solve problems. Together we brainstormed ideas on problems we had, like the lack of visibility into and the ability to change the way we utilized electrical energy at the terminal. We have substations throughout the terminal that house different types of equipment and are all fed by the Georgia Power Company.
We had very little real time visibility into what equipment was running when – what tank was being heated or mixed or what pumps were being used to load or unload a truck or railcar – so we had difficulty knowing when we had multiple pieces of equipment running at one time, which would drive up our peak electrical demand and that peak demand determines our billing rate.
Everything that we had tried to do in the past to address energy usage was very after the fact. By the time we were able to get the information from the power company, synthesize it in a spreadsheet and try to determine what path we should take to reduce our consumption, it was too late. We couldn’t make a model that would sustain conservation efforts.
One thing Atomiton and its operating stack of things brought to the table was an ability to talk to these different electrical devices and understand the data that’s coming from them and put it in a format that we could use to make decisions.
Because we had tried a few things in the past, I didn’t believe they could do it at first. But the more I spoke about the details of the equipment in the substation, the more I could see they knew what they were talking about. They were quickly able to mock up and develop a solution that kind of blew our minds.
Did you have any visibility into that equipment before?
The devices come into our Programmable Logic Controller network, but our PLC just reads the data. It wasn’t doing anything with it. We have different field devices that talk in Modbus or Ethernet and they’ve each got a different parameter or bits of information they can send to us. Some of them are very rudimentary, such as on/off. Some are more technical and can give us real time amps. It just depends on the device.
Atomiton was able to take parameters from each field device and slot them into classes -- mixers, pumps, etc. -- and then put that in a format where we could see where our energy was being used in real time.
Now we can not only see the pumps, the mixers, but also the levels in the tanks and the temperature of the tanks. They even take the external temperature from a weather station and use that to predict when a tank is going to reach a given set point. That allows us to shift activities accordingly. We can say, “Okay, this tank is heating right now. Maybe it can wait two hours to heat so let’s turn the mixer off.” Or, “We have a pump that’s going to start in two hours, let’s go ahead and either bring that up or delay that activity in to minimize our demand.”
So staggering operations that consume a lot of electricity adds up to what kind of savings?
We’re still early in the process so we don’t have a hardline number, but we expect around 25% percent savings on our peak and 10% on our usage.
I get how you can reduce peak demand, but how do you realize usage savings?
Because we didn’t have visibility, there were times we were heating pipelines that had little planned activity. We didn’t have the information needed to make smart decisions. Now we have a tool that lets us say, “Hey, we’re not using this piece of equipment for the next two days. Let’s just turn it off.”
How long did it take to roll out this IoT system?
It took about three months, and then about another two to three months of monitoring, making sure the data coming over was the data we expected, looking for interruptions in that data stream. Sometimes interruptions were because of a power blip or because of communication issues between devices.
Are you the first terminal in the Vopak system to adopt this technology and does it scale from here to other operations?
We’re the first. There has been a great deal of interest in this and I’ve given several presentations concerning the technology to organizations in Europe and elsewhere. We’re the proof of concept. And even within our own terminal we’re moving towards expanding this to other electrical substations.
How many terminals in the Vopak network?
Sixty-five terminals in 25 countries.
Wow. Ok, so if I understand it right, IoT is giving you better visibility/monitoring so you can manually intercede. Is there any automation coming down the pipe?
Yes, it is planned. There are certain activities we want to have the software manage and control, such as the heating and agitation of tanks. But there are certain tasks and activities we still want to maintain control of. But yes, in the pilot phase we do have plans for it to predict and also send bits back to the PLC network to tell equipment to turn off or on. There are some noncritical operations that could be managed better by automation with certain criteria and constraints or conditions.
Anything else I didn’t think to ask about that would be important to share?
I’ve had the privilege of working on a lot of projects at Vopak and I can say that working with Atomiton has been very easy. They are very agile and very attentive to our needs. The project went a lot faster than I expected, mainly because their development was quick. They were able to speak to our field devices. They wanted to understand the problem and they actually brought some features to us that we had not even asked for but have since become very relevant.
What kind of stuff?
Just in terms of things we didn’t think we could do. I was happy to get information to my dashboard so I could make decisions, and they said, “Hey, with this information we can look at the variables and make predictions that will help you.” That’s something we didn’t ask for but it’s something they saw once they dove into the problem.