Collecting and analyzing IoT data with respect to time can yield some very interesting findings.  Within the main circuit breaker panel of our house, I have a small device that monitors how much power each circuit is consuming.  One of the circuits is connected to the circulation pump for our outdoor pool.  In Florida, February is a pretty dry month.  Water evaporates from our pool, and the water level drops.  As the water level drops, there is not enough water to drive the circulation pump, causing the pump to start gulping air.

The chart above shows the watts that are being consumed by our circulation pump.  We have it set on a timer so that it comes on at 7:00 a.m. and goes off at 8:00 a.m., then it comes back on later in the day.  In the diagram, you can see that the power started fluctuating from between 1000 and 2000 watts around 7:05 a.m.  This was due to the pump not getting enough water to continually run.  Once I noticed this, I started adding water to the pool.  You can see that at roughly 7:30 a.m., there was plenty of water to drive the pump and the amount of power supplied was consistent.

What can we take away from this?  First off, this likely is not good for the long-term life of the pump.  If you manufactured pumps with warranties, tracking this kind of misuse would help in warranty claims.  More importantly, safeguards could be put in place to minimize damage.  In our case the fix was simple, we added more water to the pool.

What if the water was at a consistent level?  This might indicate a different problem, such as a bearing failure.  With this knowledge, one option would be to fix the pump prior to total failure, resulting in reduced costs. Or, if this pump was part of a critical process, it could be replaced prior to that critical process going offline for extended periods of time.

What else can we look at here?  This pump runs on a daily schedule for a total of about 4 hours a day.  Our electric company charges us $0.11 per kW hour.  So, the math is simple.  I know my pool pump consumes 1.8 kW per hour, so at a cost of $0.11 per kW hour, the hourly cost is $.198.  At 4 hours per day, that brings us to a daily cost of $0.792.  Not terrible, you say.  But multiply that by 30 days per month and that adds up to a monthly cost of $23.76.  Let’s say you have 20 of these pumps doing something in your factory, that now amounts to $475.20 a month and the numbers are starting to get interesting (or maybe not). At least now you have more insight into your costs.

Another topic that we will touch on in a later post is how Machine Learning can be used to look at this kind of data.  Machine Learning is a tool that can help us identify trends and patterns for either predictive maintenance or reducing expenses.

To learn how Software Logistics and NuvIoT can work with your organization to better understand all the data produced by your sensors, please Contact Us.

Pin It on Pinterest