Solar Plant IoT Gateway Architecture

Problem Statement.

Design an IoT gateway for Solar Plant that does the following:

1. Gathers status data from trackers(motors) controlling the solar panels, once every 10 minutes.

Status Data:

Angle of the panel.

Battery Voltage and Current (of the battery controlling the motor).

2. Gathers data from following sensors once every 10 seconds.

·Wind Sensor: to measure wind speed

·Snow Sensor: to measure snow level

3.Generates Alert if something goes wrong with trackers or sensors.

4.Stores all the data from sensors and trackers locally(with backup of 15 days of data).

5.When Wind Sensor detects wind with more than 10m/s, the gateway puts all the trackers to 0 degrees angle.

6.When Snow Sensor detects Snow with more than 500mm, the gateway puts all the trackers to 55 degrees angle.

7.Pushes all the data to a cloud storage periodically.

8.User should be able to connect to the gateway, when they are at the field, and see the status of the tracker and control the trackers.

Constraints:

Gateway should be able to work in areas with intermittent internet connection.

Gateway should not consume more than 40% of memory (1 GB RAM).

The SW should implement good security features.

Gateway SW should be easily upgradable.

The SW should work 24x7 and should be able to recover from any issue.

Proposed solution

Architecture diagram:

Main points:

• The main protocol and underlying technology will be MQTT
• Another protocol and underlying technology will be modbus to control all serial bus connected IoT sensors and devices
• RaspberryPi will be used as MQTT broker as well as the module gateway besides it will run a small application for report dashboard with matplotlib
• Cloud server we can use any however here we depicted as AWS with its IoT core resources
• We will use several IoT things say Arduino or RaspberryPi Zero or other smaller things which will be used for sensors or motors
• Highly secure
• dashboard with the main broker can access only with VPN and login
• all IoT devices will be attached with a certificate key
• can create IP boundary to access the message
• even local NoSQL database will keep the .csv or similar file with encrypted way

Features:

Already described in first page

Solution details:

• All IoT devices will be connected with Modbus and local network
• RaspberryPi as a broker and gateway will interact with all IoT devices for pub/sub
• Small dashboard application with the local data will be using matplotlib and related lib
• Every one hour or so it can run a cron job to push the data to the cloud or can use a MQTT pub/sub model where AWS IoT core and Lambda will be as a subscriber to get the message and process
• Customer and admin can access via a REST API based application from either mobile or web to see analytics or else
• There will be a fail save raspberry pi as redundant/automatic back for main broker
• Any emergency field operator [who is working in same network] get subscribed message of any kind of failure of system
• From mail application always subscribe a heart beat check of main broker from cloud. In case of failure it will send the sns or/and ses to responsible team
• There will be a facility to automatic update the broker or gateway application via OTA
• We can keep a read replica of main database [sql/no-sql] so that it will be helpful for any kind of read access including analytical process
• We will use a bridge for MQTT [mosquito] broker and AWS IoT core. So that bi-directional exchange is possible from field to the cloud and vice versa.
• The gateway application process is in an asynchronous way. So that it can wait/sleep for some time say 10 min to get the information and command IoT devices what to do, eg. Manoeuvre the solar panel in a particular angel based on some conditions. Which input comes from these sensors.
• Every IoT device or cloud IoT core lambda will be subscribed or/and publish some messages in time intervals or conditions to a specific MQTT topic.