Spark Project/Blog

Post 2


In my last post, I presented some background on the green building where this project takes place, and I discussed the basic signals that need to be measured for this project. In order to create inputs into my smart home dashboard, I need to measure temperature and humidity both indoors and outdoors, photovoltaic array output, solar thermal usage, and grid-tied energy input. I described how temperature, humidity, and photovoltaic output were going to be measured.

Now I’ll discuss how the remaining signals will be measured. The local utility has already installed devices that measure natural gas usage and grid-tied electricity consumption. Although many of these meters generate signals for automated meter reading by a computer, they don’t provide the homeowner any way to read these measurements in real-time.



As with any dashboard, displaying sensor data in real-time is pretty important. Imagine trying change your driving habits to optimize the fuel efficiency of your car, but you could only look at monthly reports of gas usage. It can be done, but it’s difficult to connect each change you make in the system to the resulting impact on efficiency. In the future, electric and gas meters may have accessible outputs suitable for home computer interface. Since that is not the case now, I have to find a DIY method to measure gas and grid-tied electricity non-invasively. Fortunately, I’m not the only person interested in measuring this information, and several solutions do exist.

I have come across several projects household electricity usage is being measured with a clamp-on current sensor like this project.



That is a good place to start for our current measurement. Generally, these clamp sensors generate a small AC voltage in proportion to the AC current passing through the closed loop formed by the clamp. I can measure this AC voltage output to capture accurate sinusoidal current values. To monitor voltage, I’ll use the Fronius inverter mentioned in my previous post. This is a safe and inexpensive way to measure line voltage since we already have the inverter. Other devices do exist to measure voltage, but connecting a permanent voltage sensing device to the AC mains requires a professional electrician. It is safe to assume that the voltage at the inverter will be relatively close to the voltage at the main circuit breaker panel, so I am happy to use the built-in voltage measurement capability of the inverter to provide a safe, non-intrusive voltage measurement.

Natural gas usage is more difficult to measure. I need a non-intrusive method of finding out how much gas the gas meter is measuring. The most feasible solution is to use some sort of optical sensor to watch the numbers turn over or watch the dials rotate. Once again, others have come before me with optical gas meter monitors

That looks great, but my meter is different:



The meter has dials instead of a reflective pulse wheel. Optical measurement still seemed to be the best option, so I pulled out a reflectance sensor from Pololu:




I want to have decent measurement resolution with my gas measurement, so I pressed the reflectance sensor against the gas meter, facing the 1/2 cubic foot indicator dial. Rather than turn on the gas to spin the dial, I only had to move the sensor around to generate a pulse when the sensor crossed over the dial needle, confirming that I can measure each 1/2 cubic foot of gas usage.  The sensor pulse is an analog signal. Using a simple LM339 comparator circuit, the analog value generates a digital pulse when a reference threshold value is crossed, as shown here:



I’ve set the reference threshold so that I get a digital pulse only when the dial needle passes underneath the sensor. The light reflected by the clear plastic meter face is ignored.

Now I have sensors to capture my energy generation and usage, including natural gas, grid-tied electricity, solar electricity, and solar thermal input. Also, I have sensors which give me indoor and outdoor temperature and humidity. I have set the stage for getting this information into a computer and displaying it to the user, which I will discuss in future posts.

Commentary

Text Comments (2)
desnotes I am looking to do something simliar with a reflectance sensor on my electric meter but the distance from the dial to the outside cover is about an inch. Do you think the Pololu sensor would work on the longer distances?

Thanks,

desNotes@gmail.com
kippbradford Hi desnotes,
It might be challenging to detect the change in reflectance from that distance. The sensor will be impacted by a greater amount of ambient light, so the small change you are looking for will barely be detectable. You could try placing a small diameter, 1/2" long tube around the sensor. Even though this will increase the detection distance, it narrow the angle through which light reaches the sensor, thus reducing the incoming background light.


2010 Challenge
NOW OPEN!
Round 2 ends in:
  Days
  Hours
  Minutes
  Seconds


Key Dates
  • September 22
    embeddedSPARK
    Challenge 2010 Opens
  • January 9
    Round 1 Closes
  • January 15
    Round 1 Winners
    Announced
  • April 2
    Round 2 Closes
  • April 9
    Round 2 Winners
    Announced
  • April 22
    Three Finalists Compete
    LIVE On Stage At
    ESC Silicon Valley
This site hosted for Microsoft by Smooth Fusion, Inc.