Using IOT + MQTT + Node-Red to Monitor my leaky wall


#1

I have been an an infrastructure guy more so than a developer my whole career. Last month, I discovered Node-Red and was instantly hooked. I work with developers and databases all day, but my programming skills have never been A+. Discovering Node-Red has rapidly allowed me build applications and connect them to various IOT devices.

Here is the project I used to really cut my teeth on Node-Red.

Problem: I have a leaky wall on the outside of my building. I had thought that putting a new roof on the building and patching the stucco on the outside of the building would have solved the issue, but it did not.

Solution:
The only way to solve a problem is to know what is causing it. Here is the application I built to monitor the moisture coming into my house from the wall:

  1. Created some IOT sensors that were able to detect variations in Humidity and Moisture in a physical substance. I built two of these devices, and coded them in Ardunio. These devices use the NodeMCU chip which is an awesome little ESP wifi chip. One I put between my drywall and the old brick and emended the Moisture sensor in the insulation. The other I put in my room under the baseboard and stuck the moisture sensor in the exposed insulation.

  2. I spun up a Ubuntu VM and installed Mosquito as my MQTT broker, NGINX to reroute web requests, and MySQL as a database. I then created a simple DB to write the requests from the MQTT broker to a table, and a few report tables to build my charts off of.

  3. To monitor humidity variations from the outside world, I connected to a weather API and also write that data to a table. I also wrote water events and their duration to a table so that can be compared to sensor readings

There is much more to do on this project. I need to make more sensors that focus on measuring moisture alone and deploying them in multiple locations. Making them battery powered is also preferred (so thats a bye bye to wifi, hello to LoRa or Zigbe). Once I have more data points I can start answering questions like: Are rain events triggering Moisture increases? And if so in what areas?

Let me know your thoughts, ill be sure to post updates as I make progress on this!


#2

Wow, great project!

Living in a Victorian house in the North of the UK, we also regularly suffer from "damp" problems (hint, they are more often condensation not damp for us). So this might be of considerable interest.

Would love to know the details of the sensor.


#3

So far I havent seen any solid correlation between Humidity and dampness, the moisture output is going to be a lot more useful in detecting the issue. I have these guys attached to NodeMCU sensors:

I believe the best way to really get to the bottom of this is to, build 5 of these sensors and deploy them along the suspected issue sports. Then make sure the areas dry out and watch how they react to rain events.


#4

Even if the problem were condensation related, you might not see that correlation since the real issue is the temperature of the wall compared to the humidity. If the wall temperature is above the dewpoint, you shouldn't get any condensation.

Of course, once the plaster on the wall is damaged by condensation or damp, you are also likely to get a salt buildup. Once that happens, you will continually get problems anyway since the salt absorbs water.

Do please update use once you get some measurements. I'm really interested to see how you get on.


#5

I certainly will! I am in a tight spot with this wall and it has been a frustrating past year. I tried to make the most out of a bad situation by cutting my teeth on Node-Red and building a platform where I can use data to test results.

A good next step would be to start monitoring temperature as well and comparing it to dewpoint.


#6

I posted a more in depth update on this on my Medium. I discuss some of my findings. I'm pretty certain I have discovered the issue. I'm throwing out one more sensor tonight... I live in Kentucky but we are still going to get a TON of rainfall here from the hurricane. I will post my findings!
Here is the link!: