SmlMqttProcessor

Process Smart Message Language (SML) from power smart meters with SML infra-red (IR) interface.

It depends heavily on the great libsml library (GPL3 license), forked from volkszaehler/libsml.

Concept:

This implementation uses a Rasperry Pi Zero with an atatched serial optical TTL reader to capture SML data sent by a power smart meter.
The SML processing is done by libsml.
libsml's textual output is parsed, aggregated for a 1 minute time window and forwared as MQTT statements.
The MQTT messages are used in e.g., Pimatic.org smart home app to display the power consumption history, or using Grafana.

Motivation

I want to monitor the live/current power consumption, and I want the values as MQTT messages.

Know the current power consumption, e.g., to detect peaks, evaluate power production by photovoltaic system, save electricity, etc.
The power grid operator installed a power smart meter which has an optical interface. This interface tells about the actual/current power consumption, encoded as SML messages.
Aggregate the every-minute data for a time window (e.g., 1 minute) to produce mean, median, min/max, etc.
Periodically (e.g., 1 minute) publish that aggregated data as MQTT messages.

Requirements

Python 3.5+
- Python version or code level must be compatible to the Python version on the target system, e.g., Python 3.5 for Raspbian Jessie (05/2020).
Python pipenv
wget
volkszaehler/libsml, version 6609c8117ba

Hardware and runtime requirements:

Raspberry Pi Zero with Raspbian Jessie.
Optical TTL IR reader, directly attached to the Raspi TTL serial TX/RX GPIO pins.

Setup & How-To

git clone --recurse-submodules this repository
- ATTENTION: --recurse-submodules is needed (e.g., for libsml)!
pipenv --site-packages sync
Build sml_server_time, see sml_server_time/README.md.
Run in activated virtualenv and study CLI help: pipenv run python smltextmqttprocessor.py --help
Create local configuration file and adjust settings: cp config.template.ini config.local.ini
- MQTT configuration: server host, port, user/password
  - single_topic boolean switch for all data in one single MQTT topic (default is false)
- Serial port configuration
- Block/Window size (for data aggregation)
Run in activated virtualenv: ./sml_server_time/sml_server_time /dev/ttyAMA0 | pipenv run python smltextmqttprocessor.py --config config.local.ini -

Run at Boot Time

After setting up, use systemd script in ./scripts/systemd/. Basically it runs: ExecStart=/bin/sh -c '/opt/smlmqttprocessor/sml_server_time/sml_server_time /dev/ttyAMA0 | /opt/smlmqttprocessor/.venv/bin/python3 /opt/smlmqttprocessor/smltextmqttprocessor.py --config config.local.ini -q -'

Concept

The processing pipeline is:

Input	Processing	Output
Decoded SML data from IR smart meter reader	smltextmqttprocessor.py	MQTT messages

Sensor Hardware

TTL IR read-write reader, see volkszaehler.org.

Input

My power smart meter model is Iskraemeco MT631. For details see Wiki page of Volkszaehler.org:

"The meter has a bidirectional IR interface in the upper right corner, which can be read by means of an IR read-write head at 9600bd, 8N1. The connection cable of the read head points down."
"The meter outputs the data in SML without request."

The SML data is like (also see libsml examples):

1B1B1B1B010101017605013880316200620072630101760101050068
2ABB0B0A0149534B0004325EC57262016500682793620163F9AB0076
0501388032620062007263070177010B0A0149534B0004325EC50701
00620AFFFF7262016500682793747707010060320101010101010449
534B0177070100600100FF010101010B0A0149534B0004325EC50177
070100010800FF65001C810401621E52FF65001E5A99017707010010
0700FF0101621B5200521A01010163BF270076050138803362006200
726302017101635567000000

The IR interface sends the SML data with a frequency of 1 signal per second.

Processing

sml_server_time/sml_server_time /dev/ttyAMA0 | /python3 smltextmqttprocessor.py -

The output of sml_server_time is piped to smltextmqttprocessor.py.

Step 1: `./sml_server_time/sml_server_time`

Input	Processing	Output
TTL serial data, from smart meter	SML decoding	decoded, textual output

See sml_server_time/README.md. The executable binary sml_server_time is based on the libsml-example sml_server, but modified to also include the act_sensor_time data field.

The output of sml_server_time looks like (decoded SML data):

1-0:96.50.1*1#ISK#
1-0:96.1.0*255#0a 01 49 53 4b 00 04 32 5e c5 #
1-0:1.8.0*255#198927.3#Wh
1-0:16.7.0*255#26#W
act_sensor_time#6825875#

Step 2: `smltextmqttprocessor.py`

Input	Processing	Output
Textutal output from sml_server_time	Parse heuristic, math aggregations (mean, min, max, ...); MQTT message building	MQTT messages sent to broker

Output

The output on MQTT is like this (for multi-topic sending, config setting single_topic=false, which is the default):

tele/smartmeter/time/first 11824115
tele/smartmeter/time/last 11824175
tele/smartmeter/total/value 1189840.5
tele/smartmeter/actual/first 241.0
tele/smartmeter/actual/last 238.0
tele/smartmeter/actual/median 242.0
tele/smartmeter/actual/mean 242
tele/smartmeter/actual/min 235.0
tele/smartmeter/actual/max 254.0

For config setting single_topic=true (without line breaks; here only for readability purposes):

tele/smartmeter {
"time": {"first": 6825875, "last": 6825877},
"total": {"value": 198927.3, "first": 198927.3, "last": 198927.3,
   "median": 198927.3, "mean": 198927, "min": 198927.3, "max": 198927.3},
"actual": {"value": 27, "first": 26, "last": 27,
   "median": 26, "mean": 26, "min": 26, "max": 27}
}

SML Links

License

AGPL3, see LICENSE.md.

the78mole/SmlMqttProcessor