Founded in 1961, Korea Electric Power Corp. (KEPCO) supplies electricity to approximately 20 million customers in South Korea. The company is aiming to become a global, top-five electricity provider by rolling out smart grids and promoting low carbon, power-generating technologies.
KEPCO implemented Oracle Exadata Database Machine and Oracle Business Intelligence Suite, Enterprise Edition to underpin its Substation Operating Results Management System (SOMAS). The system analyzes load status data from transmission and distribution substations across the country. Oracle Exadata has enabled the company to process load information 100x faster and compress data up to 20x. This analysis also allows KEPCO to monitor power usage and take any necessary preemptive action to avoid outages.
Demand for electricity in South Korea is growing by 10% each year, driven by rapid industrialization and an increasingly plugged-in society. KEPCO must ensure consistent power delivery at all times. If demand outstrips supply, outages occur, inconveniencing people and potentially causing serious damage to the grid. On the other hand, oversupply has a detrimental effect on the environment and wastes valuable corporate resources.
To maintain optimum supply, KEPCO has to continually monitor the voltage that flows through transmission and distribution substations across South Korea. The company’s substations produce 74 million kilowatts per day. The company relies on SOMAS to collect data on the load of transformers and powerlines. SOMAS is designed to detect if overloading or defects occur in any of the substations, and to immediately alert support staff to potential issues.
With customer numbers and power consumption increasing, KEPCO needed a robust database that had the power and scalability to process rapidly growing volumes of load data. The company generates 2 gigabytes (GB) of data per day and expects data volumes to increase by 219 GB every year.
KEPCO also wanted the ability to collect and analyze load data simultaneously and in real time, so that it can assess the operational status of substations and take corrective action if it detects abnormal load patterns. Real-time load data would also help KEPCO respond quickly in emergency outage situations.
In 2011, KEPCO implemented Oracle Exadata to collect substation and powerline load data from SOMAS. KEPCO also deployed Oracle Business Intelligence Suite, Enterprise Edition to conduct more sophisticated analysis of the load data.
Oracle Exadata Database Machine enabled significant improvements in processing performance. KEPCO ran data processing speed tests using load data from substations. The results confirmed that processing speed increased by a maximum of 170x when data was uncompressed.
On average, when KEPCO tested load data from substations, the company found data processing speed improved by 100x, regardless of the availability of an index or whether data was compressed.
The company also utilized Oracle Exadata’s Hybrid Columnar Compression and SmartScan features to compress data tables by up to 20x and speed up data transaction speeds by up to 100x. This has improved online transaction processing (OLTP) speed by 126x, as well as reduced storage requirements significantly.
The faster data processing speed also means the company has enhanced its real-time analysis capability by increasing the frequency of data updates to SOMAS from every two minutes to every 30 seconds—a fourfold improvement. This ensures KEPCO can detect and resolve issues and anomalies in the power supply before an outage occurs, improving customer service.
KEPCO incorporated Oracle Exadata into the redesigned SOMAS internal system, reconfiguring it so that its performance remains unaffected, even when real-time data collection (OLTP workload) and real-time data analysis (online analytical processing (OLAP) workload) take place simultaneously. Previously, running these two workloads in parallel slowed down the system’s performance considerably. KEPCO further stabilized performance by improving the hourly CPU usage of database servers, from an average of 95% to an average of 10%.
The faster data processing speed and the ability to collect and analyze load data at the same time resolved all work delay issues caused by slow system performance. This ensures KEPCO has accurate, up-to-date insights into the status of its substations and powerlines. The company can monitor the transmission load of each of these assets, and take proactive action to prevent power outages and ensure electricity supply meets consumer and industry demands.
KEPCO now has a high-performance, next-generation data processing platform with sufficient capacity to deal with the increasing volumes of data and processing demands of the utilities industry.
KEPCO implemented Oracle Business Intelligence Suite, Enterprise Edition to strengthen real-time analysis. The solution allowed the company to analyze load fluctuations in its substations over three-year periods; previously, the scope of analysis was just three to four months.
KEPCO can also draw on data from the past 10 years when conducting hourly queries on power usage and overload rates. A deeper understanding of load fluctuations and power usage helps KEPCO better predict power consumption and increase the reliability of its power operations. It also ensures the company can plan for the future, including building new substations and ensuring the availability of human and technical resources.
KEPCO had four selection criteria for its database. First, it must have a powerful architecture capable of processing large amounts of data. Second, it had to be scalable, as data volumes were increasing at a rapid rate. Third, it should include data compression features. Lastly, the solution had to be easy to migrate; for instance, it had to integrate with existing applications, such as SOMAS, without the need for significant changes. Data also had to be migrated in a stable manner.
After three years of verification and deliberation, KEPCO decided on Oracle Exadata. The company undertook a proof-of-concept to verify the solution’s performance and was impressed with the results. In particular, the SQL test results showed that the speed of almost all SQL executions improved from 137 seconds in the previous system to 1.1 seconds.
The ease of migration was another key reason for choosing Oracle Exadata. The company would not have to make changes to existing applications running on Oracle Database because Oracle Exadata uses the latest Oracle database architecture.
KEPCO kicked off the Oracle Exadata implementation in December 2011. The entire process―from order placement to testing and stabilization―took less than two months. The data migration and environment configuration took only one week to complete.
Oracle Partner KEPCO KDN Co., Ltd is KEPCO’s IT subsidiary. KEPCO KDN helps KEPCO implement IT infrastructures that are tailored to the specific requirements of the power industry; maintains and repairs IT assets; and provides experienced technical assistance to resolve IT issues. The company offers automated control and other systems for power generation and distribution operations and transaction systems to modulate power supply and demand.
“Thanks to KEPCO KDN, we have established a next-generation database without completely overhauling or dismantling the existing system configuration,” said a spokesperson from Korea Electric Power Corp. “This has enabled us to provide higher quality service to our customers within a short period of time.”