Architecting For Energy

It’s one thing to architect a system for quick discharge of energy, but what if you want the energy discharged over a longer timeframe?


Most of the time, electrical design engineers create their designs by putting in the functions first, then going back later to figure out what they can afford to shut down. But with energy harvesting, this is flipped, and the systems must be built to normally be ‘off,’ and with a clear understanding of the minimum power requirements.

Also, most of the time, energy harvesting systems are designed for a quick discharge of energy — but what if you want the energy discharged over a longer timeframe? Is it a matter of storage?

Jeff Miller, product marketing manager, in the deep submicron division at Mentor Graphics explained it’s not really a matter of storage, as it turns out. “If you want a longer discharge less frequently, then storage is your answer. If I wanted a system that could use a large quantity of energy, then I could say, I’ve got a harvester that can get this amount of power over time, and I can integrate that over a longer interval in order to get to the point where I have enough stored energy to do something large; for that you would need more storage but usually people think of these things in terms of wanting to be able to do something with more energy, more frequently, or something like that. And for that, you have to harvest more energy in order to consume more energy.”

It turns out, what it comes down to are the three fundamental pieces within the energy harvesting system:

—The harvester. This has a couple of important characteristics such as the power output (both the maximum power output, and the average power output over time). For example, a solar panel only gets full sun at certain times of the day; a vibration harvester isn’t necessarily going to get consistent kinetic energy applied to it. As such, there is a a concept of the average energy harvested, then it needs to be converted and stored.
—The power management system. This charges the battery, and wakes up the system when there is enough energy stored to do what it needs to do.
—The system itself. This is the sensor, the communication package, the data collection device, etc.

Miller explained you want to either work forwards or backwards when created the system. “Say you know the amount of energy that a single operation takes, then you work backwards to set the size of the storage system. Then the question is, as you have a harvester that can do so many watts, how long will it take to store up that storage system to the point that it can do an operation, and that will set the duty cycle. How many times a day I can take a measurement is determined by the energy required to take a measurement divided by the power harvested.”

Interestingly, there are other constraints that come into designing energy harvesting systems because they are often IoT kinds of systems that contain lots of little sensors, he said. This then leads to other constraints such as physical size, durability and handling. “This is especially true with MEMS energy harvesters because by their nature they could tend to be fragile because there are little moving cantilevers that harvest the energy through a piezoelectric system. As such, there are a lot of interesting things that go into the design especially the MEMS energy harvester because they have to be able to tolerate handling yet still generate useful amounts of power.”

Not surprisingly, the packaging turns out to be one of the most interesting parts of these energy harvesting systems because the packages must be designed so that when it gets dropped or shaken, that the cantilever collides with the roof of the package and rests in a safe way. “A lot of MEMS have interesting packaging considerations but energy harvesters in particular have that consideration as the package provides the robustness of the system,” Miller added.

As a design engineer, what are your experiences designing energy harvesting systems? Chime in below.