Overcoming the Battery Obstacle: 8 Reasons Why Batteries Restrict Cost-Effective IIoT Deployments

The most obvious issue, as the IDTechEx report noted, is that all batteries eventually need to be replaced.  As we pointed out earlier, the cost of accessing and replacing dead batteries—because such processes must still be done manually—is often much greater in resources and man-hours than the cost of the new battery itself.

This need for frequent manual effort immediately defeats the core value of connected sensors.  Indeed, one of the primary reasons manufacturers and industrial-plant operators turn to IoT sensors in the first place is that the automated data coming from those devices can eliminate the need for physical inspections of equipment, machinery, pipes, or other assets.

If all of those sensors themselves need to be manually checked on a regular basis to ensure that their batteries are functioning and/or to ultimately replace those batteries, then the organization has in effect traded one time-consuming manual maintenance schedule for another.

The inevitablility of a dead battery can have consequences beyond the marginal labor and capital resources required to inspect and replace batteries.

Unless the team overseeing a plant’s IoT sensors discovers a dead battery immediately and is able to quickly get out to the sensor and replace it, the plant will permanently lose whatever data the sensor would have been collecting and transmitting in the interim.  To make matters worse, as a Clemson University paper points out: “Batteries wear out quickly in wireless sensor networks, even when carefully managed.”

Because some of an industrial plant’s sensors record and stream data that are mission-critical for safety and compliance, dying batteries can create significant hazards for the business.

Ideally, an IoT device at an industrial plant—say, a sensor positioned near the facility’s chemical operations to continuously monitor the atmosphere for toxic leaks—should be transmitting its data extremely frequently. Updates several times a minute are ideal.

But every data transmission consumes power.  So, to extend battery life, many IoT sensors are unfortunately configured to transmit data far less frequently than would be ideal—sometimes in batches as infrequently as once every 24 hours.

This can give a plant’s operators an inaccurate picture of the data a sensor is capturing.  With updates only once a day, for example, a manufacturing plant’s team might get an erroneous picture about the environmental quality of a given facility or the condition of a given asset.

Over time these inaccuracies, as well as the potential for false positives and false negatives, can increase exponentially, rendering the sensor’s data increasingly misleading and failing to deliver on the IoT’s promise of “real-time” awareness.

As an article in Electronic Specifier magazine explains, batteries are often the largest part of an IoT sensor system, leaving engineers limited choices of which batteries to add to their sensors.8

Moreover, as the Clemson University paper explains, the size, weight, and dimensions of the battery often limit the usefulness of the sensor.  This is because those physical characteristics of the battery can restrict both the types of applications a sensor can perform and which other components the battery can coexist with on the sensor’s board, as well as where it can be deployed (with embedded locations, of course, off limits due to required battery changes).

Finally, as the US National Institutes of Health (NIH) reports, the lithium batteries commonly used in IoT sensors “may contribute substantially to environmental pollution and adverse human health impacts, due to potentially toxic materials.”

You can see why this might be the most concerning aspect of the continued deployment of battery-powered IoT devices around the world—particularly if these devices are rolled out in the coming years by the billions or tens of billions as predicted.

The NIH report explains that lithium batteries pose health risks to humans due to the leaching of cobalt, copper, and other substances, and they can harm the environment through leaching such substances as thallium and nickel.

As you can see from this National Institutes of Health graph, lithium batteries have shown far higher concentrations of all of these metals than the government’s minimum threshold required for classifying a substance as hazardous waste.

Health report, “Potential Environmental and Human Impacts of Rechargeable Lithium Batteries in Electronic Waste”

Another important concern cited in the Clemson University paper is that, unlike other areas of technology, batteries have historically shown slow rates of performance improvement over time.

The paper points out that Moore’s Law (the observation that semiconductor performance doubles roughly every 18 months) does not apply to the chemical and manufacturing processes used in the production and research of battery technology.

For example, after batteries received their most recent “upgrade”—the move to lithium, the lightest substance available for production—performance improvements to batteries have been at most a few percent a year, not nearly enough to keep pace with the advances in computing power in general, or the increasing demands of IoT devices in particular.

Yep, we’re back to the battery problem, although this subset of the problem involves a unique set of challenges.

With existing commercially available parts, the combined energy needed to power all of a sensor’s operations—data sensing, processing, memory, wireless communication—necessitates a battery, particularly if this sensor is built using disparate components from different sources.

Some manufacturers are developing components for wireless IoT sensors billed as “ultra-low-power,” and that’s great.  But when an end-user pulls together a set of disparate components, we’re back to the battery problem.

The microcontroller might be built for low power consumption, but what if the radio, or temperature sensor, or clock chip isn’t?  Ultimately, despite the burdensome integration effort, today’s available technology simply does not allow an end-user to build a useful, fully integrated sensor network that leaves the battery behind.

Another weakness in sensors built from disparate components is that these sensors often can’t effectively respond to the inevitable challenges an IoT network will face.

Imagine, for example, that within some of your wireless sensors the data processing capability is becoming overloaded. If your system is comprised of disparate, standalone hardware and software components that don’t communicate seamlessly with each other, you might have difficulty solving that problem—and you might not even be alerted that the issue exists. The data processing tools will simply continue working as hard as they can to meet the demand, until they fail, and none of the other parts of your system will come to help.

Now imagine that your entire IoT network was built as a unified ecosystem—with all pieces designed to work together, from data acquisition to processing to analysis to transmission.

In this type of cohesive environment, you can much more easily modulate and adjust your system to meet your changing needs. If some of your processors are becoming overloaded, the system can transfer some of the data-processing workload to less-taxed areas of the system.