Asset Management System

Background
Many energy supply companies have implemented asset management systems, but they experience difficulties in supplying energy stably due to the failures occurring from a large number of equipment and yet they are still pouring huge amounts of time and money to deal with them. Especially if no quick methods exist to identify the failed layers and devices out of all the equipment which belong to those layers, then the costs of facility management for companies will continue to grow and it is difficult to accurately measure the energy consumption of users.

Challenge
Company D has built a data collection infrastructure to secure a stable energy supply, and it is comprised of countless equipment in each layer. Once users start to run the equipment installed where the energy is used, the data is then collected in the servers of the company through the Modem layer which transmits and receives the power usage information data, the Data Control Unit layer installed in public facilities or underground to control the data flow and the Front End Processor layer that already exists in the company, in sequence. Problem with this process is that it takes a long time to find out exactly what the failures are and in which equipment they occur.

This company had an enormous amount of log data collected from every equipment layer, but there was no big data system to store and manage it, and there was a lack of data models and analysis skills to identify the equipment to be repaired promptly through the rapid analysis of the network where the equipment are complicatedly connected to each other. Because all of the log data with equipment network information was stored in a traditional relational database, it was not sufficient to process the data in a large scale network structure.

Solution
Firstly, Bitnine stored all of the log data in the Hadoop echo system (Apache Hadoop), followed by building systems which manage through the graph representation where much equipment are connected to each other by expressing each equipment in the real world as a ‘dot’ and the relationship among that equipment as a ‘line’ in the graph model through Bitnine’s graph database of AgensGraph.
Bitnine also provided a platform for visualizing the equipment network for the company. This platform not only provides the visualization that allows the company to check every single equipment while simultaneously understanding the entire infrastructure structure and its status at a glance but also manages all activities to maintain and manage massive equipment to keep their history.
A system has been built based on these platforms to analyze the failures in full. The equipment whose temperatures exceed the acceptable level or whose network speed is getting too slow can be quickly discovered to analyze when they are using this system. And also, utilization of the failure patterns after learning them using artificial intelligence provides a foundation not simply for “managing” failures but for “preventing” them.

Benefits
The key to the solution that Bitnine has provided to the company is a graph model based platform to manage a complex network of equipment in a database, like the way in the real world. If you use the graph database after escaping from the relational database, no significant differences arise in the time needed to check all of the equipment even if the equipment network becomes complex. Therefore, company D can locate its problematic equipment in the large equipment network and then can quickly check the status of the equipment and track the repair history.
Furthermore, it is possible to learn the failure patterns by applying the various graph algorithms and analysis techniques to the equipment data stored as a graph, and thus it is possible to recognize and prevent the signs of failure in advance.
Our failure analysis system (Asset Management System) that utilizes AgensGraph can be used not only in the energy industry but also in various fields such as communication, smart factory and the road network in terms of providing the relevant capabilities to manage and monitor the failures of large-scale equipment networks easily and centrally.