Acknowledgments

This project leverages the great work done by Douglas Hall (https://github.com/bemasher) on rtlamr, a rtl-sdr receiver for Itron ERT compatible smart meters operating in the 900MHz ISM band. Project link for rtlamr: https://github.com/bemasher/rtlamr. He has also some work with Docker (see https://hub.docker.com/u/bemasher).

Background

I've been using rtlamr plus some custom collection and graphing code on my general purpose Linux server for several years to track electric usage and solar panel output. In the middle of 2019 I decided to move it to a Raspberry Pi. At that time I decided to see if using Docker and Docker Compose could help with managing deployment and fault tolerance.

So far the results are very promising, and I have decided to open source the result. Docker-ized RTLAMR Meter Scanner was the first release, designed to help potential users test rtlamr and see if their meter data can be collected (https://github.com/val123456/Docker-ized-RTLAMR-Meter-Scanner).

This project, Docker-ized rtlamr to SQLite is the second release. Three Docker containers are deployed using docker-compose to run rtl_tcp, rtlamr, and a custom Python app to read the JSON output from rtlamr and store the meter reading and current daily use in an SQLite database. Diagram and further explanation TBS.

Storing the data in SQLite form allows for many uses. The graph shown here https://val123456.github.io/index.html was generated with Plotly and allows zooming, etc. My third release will be the code used to generate graphs from the SQLite database.

Note: This project is hosted on Gitlab at https://gitlab.com/colinv/docker-ized-rtlamr-to-sqlite and mirrored to GitHub here https://github.com/val123456/Dockerized-rtlamr-to-SQLite

Caveats

I have developed this based on the meters I "own": two SCM type electric meters, one SCM type gas meter, and one R900 water meter. All of these meters provide a Consumption field that reflects the actual meter reading. For information on other meter types, see the rtlamr documentation here: https://github.com/bemasher/rtlamr/wiki.

Requirements

Linux OS with Docker and Docker Compose installed and an RTL2832U SDR. I have had good luck with the dongles from RTL-SDR (https://www.rtl-sdr.com/buy-rtl-sdr-dvb-t-dongles/).

This project has been tested on the following using my SCM electric and gas meters and R900 water meter:

OS	Docker	Docker Compose	Hardware
CentOS 7.7 (x86_64)	19.03.5	1.18.0	Dell PowerEdge T20
Ubuntu 18.04.3 LTS (aarch64)	18.09.7	1.17.1	Raspberry Pi 3 B+
Raspbian 10 (buster)	19.03.5	1.25.0	Raspberry Pi 3 B+/Pi 4 B

Note: This will not work on Mac OS X due to the way the USB is handled when using Docker. I will release a version that works on Mac "soon".

Building/Using

Download/clone source. Open a terminal, cd into the top-level directory and run docker-compose build to build the containers.

Edit the text file variables.txt to include your meter ids. It will work with gas, water, or electric meters. See more info by reading the file and docs/variables.

Then run docker-compose up -d to run it. Output is placed in a directory called data at the same level (e. g., ../data/. To stop, use docker-compose down.

Note: On some systems, you may have to run these as commands as root using sudo: sudo docker-compose build and sudo docker-compose up -d

On most systems you can avoid using sudo by adding your account to the docker group like this:

sudo gpasswd -a $USER docker

Expected Output

The data directory will look like this while the program is running. raw_output.json is the raw output from rtlamr that is saved in case you want to re-process the data at some point in the future (application for that will be released in the future after it is documented).

pi@pi4:~/development/data $ ls -lh
total 13M
-rw-r--r-- 1 pi pi 5.8M Jan 25 11:31 meters.db
-rw-r--r-- 1 pi pi  32K Jan 25 11:31 meters.db-shm
-rw-r--r-- 1 pi pi    0 Jan 25 11:31 meters.db-wal
-rw-r--r-- 1 pi pi 6.6M Jan 25 11:31 raw_output.json

Data is stored in SQLite. The schema in SQL form is contained in docs/schema. Note the meter IDs are used in the table names. The documentation here shows my meter IDs.