Allen Bradley PLC + AWS Greengrass (PoC)

Some days ago I decided to start testing AWS IoT stack.

I got Allen-Bradley Micro820 Programmable CCW PLC Trainer Micro800 Training Kit, Raspberry Pi (could be Vagrant box running in the windows machine), a windows machine.

Purposed architecture:

(Note this picture was taken from the book Hands-On Industrial Internet of Things, you must buy this book if you really want to understand IIoT!)

So let's implement the suggested architecture.

Data acquisition steps:

• Download Raspbian
• Install Raspbian in SD Card
• Get windows machine
• Install Kepware KEPServer in the windows machine so it can be used as a device gateway. We will need it to convert Allen-Bradley PLC (Modbus TCP/IP) connection to OPC UA.
• Using the windows machine, connect to Allen-Bradley and assign it static IP through Rockwell software.
• Time to configure Kepware KEPService, create a channel and device. Import the datatags.
• In Kepware configure OPC UA to accept connections.
• Greengrass installation and configuration part. Instead of me telling you step by step how to do it. I am going to share which links I used.

Install and configure Greengrass:

https://docs.aws.amazon.com/greengrass/latest/developerguide/module1.html

https://docs.aws.amazon.com/greengrass/latest/developerguide/gg-config.html

https://docs.aws.amazon.com/greengrass/latest/developerguide/gg-device-start.html

Use OPCUA Lambda:

https://docs.aws.amazon.com/greengrass/latest/developerguide/opcua.html

https://github.com/PacktPublishing/Hands-On-Industrial-Internet-of-Things/tree/master/Chapter10/aws-greengrass

Configure Lambda for Greengrass:

https://docs.aws.amazon.com/greengrass/latest/developerguide/config-lambda.html

Nice YouTube video if you dont want to read:

Data storage steps:

• Create DynamoDB table for hot path and "time series" database

Partition key ts_date

Sort key ts_time

• Create an Act in AWS IoT to insert messages into the previously created DynamoDB table

Table name tablename

Partition key ts_date

Partition key value ${parse_time("yyyy.MM.dd", timestamp())}

Sort key ts_time

Sort key value ${parse_time("HH:mm:ss", timestamp())}

Write message data to this column data

Operation INSERT

Act rule query statement example

select * from '/opcua/kepware-plc/node/input1'

Result:

• Create S3 Bucket for cold storage and data lake
• Create Act in AWS IoT to store messages in the previously created S3 Bucket

S3 bucket bucketname

Key ${parse_time("yyyy/MM/dd", timestamp()) + '/' + newuuid() + '.json'}

Act rule query statement example

select * from '/opcua/kepware-plc/node/input1'

Result

Data processing steps (AWS IoT Analytics):

• Create AWS IoT Analytics Channel based on /opcua/kepware-plc/node/input1 topic
• Create Pipeline
• Create Data store
• Create data set from the data store previously created.
• In the end, you should have a new Act in AWS IoT for AWS IoT Analytics.

Result

Data execution steps (QuickSight):

• Create a new data set, use AWS IoT Analytics as data source.
• Create your dashboards

Result

NEXT STEPS:

• Use SageMaker for ML
• Use Athena to query cold storage (S3)
• Maybe ElasticSearch
• Use Amazon Timestream, AWS Forecast, AWS IoT ThingsGraph, AWS IoT Sitewise, AWS IoT Events