Energy harvesting offers significant advantages to the development and improvement of the Internet of Things (IoT). It is a critical component for creating an enhanced class of autonomous and mobile applications that can operate for much longer periods of time without the need for battery charges.
It also drives cost savings by significantly delaying battery replacement, which often costs more than the battery itself.
And energy harvesting is a key element in bringing intelligence to the edge and the IoT to a world of new places and applications. Analog Devices can help entrepreneurs with energy harvesting solutions through the Arrow Certification Program (visit https://www.arrow.com/en/indiegogo).
Enabling power at the node
To begin, let’s explore the setup of an energy harvesting system. First, an energy harvesting transducer converts ambient energy to harvested power. Then, a power management unit, or PMU, converts that harvested power into usable power. The energy storage is the third pillar in the energy harvesting power generation design. Storage is used as an energy buffer to collect energy for the power requirements of the sensor node including sensing, data processing and radio transmission.
In order to most effectively design a self-powered system, it is best to start with an estimate of how much power can be harvested from the available energy. In the case of photovoltaic systems, this is the light energy available. If using a thermoelectric harvester, a measurement of the temperature gradient that can be developed is needed, whereas in the vibration harvesting case, both the acceleration level and vibration frequency are needed to develop an effective self-powered system. A harvester type is selected based on the available ambient energy.
Next, the energy balance equation comes into play. The sensor node is designed with the energy budget in mind. Ultra low-power system design techniques are used to trim the power requirements of the sensor node. The sizing of the harvester and storage and the design of the sensor node tend to be an iterative process.
There are still some barriers to overcome when employing this new form of technology. Intermittent and low-level power sources can affect when and how much energy can be harvested. At the same time, the demand for more features at the node and the transmission of high volumes of data create a substantial drain on the available power.
Analog Devices offers ultra low-power solutions that remove these barriers, to help engineers design and create a new class of high-performance, low-power energy harvesting applications. The company has a suite of PMUs that offer highly efficient power conversion and highly optimised storage usage – all while consuming very little current when dormant.
The ADP5091 ultralow boost regulator (PMU) effectively harnesses and converts ‘indoor’ ambient light. Delivering efficient conversion (as low as 10 μW) with sub-μW losses, it also leverages open-circuit voltage sensing for optimised extraction.
Another such PMU is the LTC3109 ultralow voltage step-up converter and power manager. The LTC3109 is designed for applications with intermittent and low-level power sources, where the low input voltage requirement is critical, such as thermoelectric generators, thermopiles or small solar cells.
The LTC3331 is engineered to enable multimode harvesting from two sources, which allows for additional energy to be harvested in a particular environment. In addition, DC sources such as solar cells or thermoelectric generators can be paired with a high-voltage piezoelectric harvester to help maximise harvested energy.
Beyond harvesting: enabling intelligence at the node
Currently, the industry is experiencing a paradigm shift in edge sensor devices: low-power edge computing devices are advancing faster than cloud devices with the capability to bring intelligence as well as connectivity to the edge. As a result, data collected from multiple inputs can be fused into a single processor. For example, an outdoor air quality monitoring system could have multiple gas sensors and particulate matter sensors deployed in a common sensor node.
The advantages to leveraging intelligent edge nodes are significant. First, with intelligent filtering and decision making at the edge, less power is used because there’s little need for power-intensive data transmission to the cloud. There’s also reduced latency. When information is processed at the node, it can produce meaningful information sooner than sending large amounts of data to the cloud where more time is required to sort, analyse and deliver results.
Smart city applications
By reducing the need for battery maintenance and replacement, energy harvesting can help bring intelligent sensing to remote or hard-to-access locations within municipal infrastructures. For example, guardrails can be equipped with ADXL372 sensors that can detect an impact. That information can be processed locally and then communicated to an aggregator to alert first responders. This can bring needed help sooner and perhaps save more lives.
Additionally, with energy harvesting technology, multiple sensors and processors can be deployed over miles of road without requiring frequent and costly battery maintenance.
Utility metering is another area where intelligent nodes can play an advantageous role. Using a mesh network equipped with ultra low-power radio equipped meters, a single truck can gather data from the whole network of meters simply by passing by one. This can greatly reduce the man-hours needed for gathering billing information.
Smart factory applications
Ultra low-power sensing can also deliver benefits for industrial condition monitoring. Combining the ADXL357 ultra low-noise accelerometer with the ADuCM3029 for edge node processing, processes vibration data via an onboard FFT to aid in preventive maintenance and detect warning signs. This information can help plant and factory personnel take early corrective action to avert failures and costly downtime.
While cloud computing gets much of the focus in IoT, an increasing number of entrepreneurs and product developers are recognising the power and value of gathering and processing information at the edge. It is important to remember, however, that ubiquitous sensing in data-rich environments is only possible when the cost to maintain these sensing solutions is low compared to the value of the resulting data. Central to that is energy harvesting and management.
Analog Devices’ ultra low-power microcontroller and sensing portfolio enables efficient processing at the node, which in turn allows for more sensing systems to be self-powered. By employing self-powered sensor nodes in these environments, it is possible to achieve a low total cost of ownership solution, making sensing possible in areas previously considered impractical and ultimately solving challenges that couldn’t be solved before.