BreakdownBot

Siemens TIA TDS(Tabular Data Stream) Protocol to use to send data from PLC to database
one of the Visualization in the web appq
Siemens and BCG market research
landing page when data is uploaded
one of the Visualization in the web appq
TIA portal(siemens software) Hardware config where we setup and programmed the PLC

Inspiration

The inspiration for BreakdownBot came from the need for a solution that could streamline complex data analysis and interpretation, particularly for businesses relying on machinery and equipment. In industries where machinery downtime can lead to significant financial losses, BreakdownBot provides a way to analyze machine performance data, identify patterns, and predict potential failures before they occur. This proactive approach can save businesses time and money by preventing unexpected breakdowns, optimizing maintenance schedules, and reducing operational risks.

What it does

BreakdownBot takes in data files from any machine and analyzes them to detect patterns that may indicate impending malfunctions. Using machine learning algorithms, BreakdownBot processes the historical and real-time data from the uploaded file to identify early warning signs of issues, like unusual temperature spikes, vibration changes, or performance drops. Based on these insights, BreakdownBot predicts when a machine component is likely to fail, allowing businesses to schedule maintenance proactively rather than reactively.

How we built it

Using a very unique techstack incorporating siemens TIA portal(To prepare the PLC to send data), from there we utilized Microsoft fabrics onelake to store the data received which is then passed to a synapse Data Science to pre-process and engineer informative features to train our machine learning models for predictions, finally the simple UI built with streamlit affords the user not just the ability to predict the likelihood of a machine failing but the overall performance of these connected devices. The first few steps required not only an understating of the problem we wanted to solve but the device(PLC) as well

Challenges we ran into

Collecting data from the PLC reliably , the biggest issue with the PLC, particularly a siemens PLC is that the device is extremely rigid when it comes to what third party software can connect to it, which is due to the critical nature of the processes it runs. It required extensive research on how best we could achieve the data logging from PLC to onelake, through research we found that the most reliable way would be using bar MQTT and the MODBUS protocol was to log data directly from the PLC with MSSQL server to onelake, a more detailed piece of documentation can be found here link

Accomplishments that we're proud of

The conception and implementation of the project, from mindmaps to microsoft fabrics onelake. When Ideating it did not seam feasible but our perseverance and determination helped us see it through to the end

What we learned

Being able to connect physical, industrial level devices with onelake. It required extensive research and as challenging as it was, it was extremely rewarding. It reminded us that anything can be achieved with enough knowledge and teamwork

What's next for BreakdownBot

Improvement of the overall design and possibly pitch the idea to other companies in the industrial automation space see if it they are interested in using it to reduce their machine downtime and possibly make the entire project open source sharing Ideas with everyone