Overview
Maintaining the consumption of energy is an important task to obtain the sustainable environment. Increasing population, incomes and industrialization has lead to massive scarcity of natural resources.
Therefore, there is a need to maintain the natural resources such as energy consumption.
Prediction of energy consumption helps to make the decisions for energy purchase and generation which is done using various algorithms.
Business Challenge
-
Collecting real data for machine & pipelines is a tough challenge.
-
The collection of power consumption is a difficult task.
-
The data for the machine and its features are required to predict the power consumed by them.
Solution Offered
Multiple IoT devices installed on different machines will emit the data to the Apache Nifi, and Apache Nifi will then route the data to different sources. Now, data pipeline is built Using Apache Kafka, and then data will be processed using Apache Spark, and for Data Storage Hbase or Cassandra will be used
Using Apache Nifi to route data to S3 for creating historical data.
Spark SQL and Graph Database for Pattern and Link Analysis.
Process
Calculation of Energy Consumption and Demand
Energy consumption is measured in Watt per Hour whereas demand is measured regarding work done in 15-30 minutes.
Calculation of Energy Consumption Index
PUE is the proportion of the amount of power required to operate and cool the data station vs the volume of power extracted by the IT equipment in the data hub. The equation is:
PUE = (Total Facility Energy) / (IT Equipment Energy)
Prediction of Energy Consumption, Demand and Change of Energy Consumption
According to the need, predictions can be made on an hourly, monthly or daily basis. The models that can be used are mentioned below:
-
Gaussian process regression
-
Linear regression