Energy Profiling in Modern Manufacturing Operations using QuickSight | AspireNXT

Our client is at the forefront of the automotive industry, with their presence in automobile design, manufacturing, distribution, and sales. A team of professional from various backgrounds encompassing engineering, product development, sales & marketing and public relations, their key area of expertise in building a competitive platform for motorsports is in chassis, powertrain and electronics. Our client wanted to develop a dashboard to observe the electrical energy consumption in the manufacturing plants.

With a fully integrated engineering service capable of creating new concepts and designs, they have become full-fledged automotive manufacturer from design, development to prototyping, testing and manufacturing. Born out of a passion for high performance vehicles, they specializes in the research, development and production of race and rally cars, limited-edition road cars accessories.

In order to understand the energy consumption trends for each furnace, our client was keen on developing a Business Intelligence (BI)approach by learning energy profiling and researching energy timelines. They wanted to conceive the readings obtained from an IoT sensor installed in their plant, on a BI Reporting dashboard for benefits like energy forecasting, anomaly detection, cost reduction and energy profiling based on data distribution.

AWS QuickSight is remarkably capable of handling several Big Data sources and diligently performing smart visualization on them. It can handle multiple business domains while independently measuring business metrics. It takes advantage of machine learning to identify anomalies in data and make predictions through its ML Insights feature. QuickSight offers a vast library of visualizations, charts, tables, and add interactive features like drill-downs and filters and perform automatic data refreshes to build interactive dashboards.

• The overall flow of the data exploration to final dashboard was carried out in following sequence.
• IoT sensors generate the data, and AWS Glue ingests this information obtained DynamoDB is S3 bucket.
• The information is parsed into parquet and then crawled to record the data catalogue, followed by exposing it to Athena for ad-hoc analysis.
• EDA was conducted using Sagemaker to perform informative and inferential statistics.
• The missing values is imputed using a KNNImputer, which takes the mean value of the nearest neighbors in the data.
• Cleaned data is loaded to QuickSight and a few measured fields are generated with the help of vital visualization scope.
• Dashboard for energy profiling is carried out for the furnaces, anomaly detection and forecasting were performed to understand the consumption trend on QuickSight.
• Inferences based on the visualization is suggested for optimal energy consumption.

As the part of the solution architecture, the following AWS Services were used in the development of the Data Lake project:

• AWS DynamoDB
• AWS Glue
• AWS S3
• Amazon Sagemaker
• Amazon Forecast
• AWS Athena
• AWS QuickSight

Architecture Diagram, Microsoft Excel.

• A 360-degree view of energy profiling is set up which led to a clear visibility to target the top energy offenders eliminating guesswork and delays, subsequently saving costs.
• Data management is rising, with forecasts in place to offer real-time analysis of consumption costs.
• As identifying deviations helped them recognize the needless investment, our client was able to bridge the distance between the data and insights with anomaly detection.
• In comparison to previous BI solutions, over 15% cost saving was accomplished by the consumer with AWS facilities in place.
• Our client saved hundreds of staff hours in order to help proactive decision making with QuickSight as a BI tool in the building process.