JERA Achieves Secure and Advanced OT/IT Integration Using AWS, Delivering AI-Powered Predictive Monitoring and Operational Optimization Globally

JERA was established in 2015 by integrating the fuel and thermal power divisions of Tokyo Electric Power Company and Chubu Electric Power Company to create a global energy enterprise capable of competing in the international energy market, to provide a stable supply of energy, and to improve corporate value. Under the catchphrase “CO2 emissions–free thermal power,” the company is working on power generation that combines zero-emission thermal power with renewable energy, has set a target of zero CO2 emissions by 2050, and is tackling the global energy issue, which requires a new infrastructure that can stably and economically supply clean energy.

In addition to research and development of power generation technology, the company is actively adopting digital technology and striving to become a data-driven company.

“Since 2018, we have been working on formulating strategies and policies and have driven multiple digital transformation projects company-wide, such as building an integrated data infrastructure and enhancing operation and maintenance (O&M),” says Masatsugu Osawa, assistant manager of the Digital Transformation Unit in the Digital Creation Department at JERA. “Having completed phases such as data visualization and analysis as well as application to business and operations, we are focusing on improving business processes and strengthening data governance.”

The O&M engineering service “JERA O&M Way,” is a solution that uses digital technology and data based on experiences gained through years of power plant operations. It provides power plant operations as a service and incorporates cutting-edge methods, such as remote monitoring with digital technology and predictive monitoring with big data. The monitoring is provided by the Global Data Analyzing Center(G-DAC).

“In 2016, we launched a project to enhance and optimize operations in power plant using digital technology. In 2018, we began providing monitoring services for external customers,” says Michihiro Shimazoe, head of the G-DAC Monitoring Unit in the O&M Engineering Strategy Management Department at JERA. “We then made efforts to enhance and expand services, and G-DAC was established in 2023 to accelerate service provision, both domestically and globally. Currently, G-DAC provides monitoring for 64 power plants in Japan, Southeast Asia, and North and Central America, which generate about 30 GW in total.”

Since the beginning of this project, JERA has been building and operating the monitoring infrastructure using AWS. Osawa led the project as a cloud-first initiative.

“We believed that traditional on-premises systems would be inefficient and hinder the desired transformation,” says Osawa. “We needed a data foundation using a robust cloud with high availability, and we chose AWS because it’s easy to expand overseas and has a wide range of advanced services. At first, we were unfamiliar with cloud technology and AWS, but an AWS Premier service partner and AWS solution architects helped us to adopt and use AWS through close support.”

Among the G-DAC remote monitoring services, the IoT platform plays a particularly important role because it collects and analyzes operating results and equipment data from power generation facilities. Conventionally, data integration systems are installed by OT vendors on power plant premises, and even when data is at hand, manufacturers lead system maintenance and data management. JERA wanted to properly control both data and system design by building and operating its own infrastructure.

“Security is very important in OT/IT integration, and compliance with various regulations and guidelines is required,” says Masayoshi Sato, assistant manager of the Information Security Office at JERA. “We achieved a high level of security between OT and IT by using a closed network divided into multiple subnetworks.” The closed network feature is provided through AWS Direct Connect, which creates a dedicated network connection to AWS, and the subnetworks are created using AWS Site-to-Site VPN, which is a fully managed service that creates a secure connection between data centers or branch offices and AWS.

“The fact that we were able to configure a virtual network with the same requirements as our own network is a major advantage,” says Sato.

The new IoT platform built by JERA on AWS uses AWS IoT Core, which easily and securely connects devices to the cloud; AWS Lambda, which lets you run code without thinking about servers or clusters; and Amazon Kinesis, which cost-effectively processes and analyzes streaming data at any scale as a fully managed service. The platform receives raw data from each power plant and distributes the data to the corresponding websites, which are built for each plant to provide maintenance, monitoring, and management functions using services such as Amazon SageMaker, which is a fully managed service that brings together a broad set of tools for high-performance, low-cost machine learning for any use case. Specifically, the platform provides functions such as predictive monitoring, performance management, and coal fired boiler operation optimization.

The anomaly detection system, used for predictive monitoring, evaluate the deviation between machine learning-calculated predicted values and actual values to detect and alert any abnormalities at an early stage. This allows for a quicker response than conventional alarms that issue after abnormalities have occurred, making it easier to avoid unplanned shutdowns, thus maintaining stable power plant operations. 

The performance management system diagnoses boiler efficiency, turbine efficiency, and auxiliary power used in power plants and analyzes equipment performance deterioration and its causes. It even suggests operational improvements and equipment-cleaning processes, helping on-site operators to plan and implement them promptly.

The coal fired boiler operation optimization system calculates optimal operation parameters with AI trained on past equipment’s operational data. To maximize power plant efficiency, it is necessary to adjust plant settings according to changes in fuel and the external environment. However, because there are many different operation parameters and operation volumes, the operators don’t adjust the operations in detail. An experienced operator could adjust the complex settings based on experience and knowledge, but an inexperienced operator couldn’t. This system lets all operators achieve optimal operation, removing personal differences.

Security is another integral part of working with mission-critical OT. JERA’s IoT platform creates a secure environment by separating plant networks from the internet and enterprise networks in order to meet strict requirements, such as connection point minimization and network isolation, based on security guidelines for electric power companies in Japan.

Due to all these initiatives, the anomaly detection system realized a more than 20 percent reduction in unplanned downtime. JERA is continuing to make effort to further reduce this number, ultimately aiming for zero unplanned downtime. The performance management system and the coal fired boiler operation optimization system improved efficiency, saving more than 100 million yen in fuel costs per year at target units.

JERA intends to further promote the use of AWS in order to build a digital platform for decarbonization. “JERA will closely integrate the maintenance capability of field engineers with the monitoring and analysis by data analysts. We will provide stable and highly efficient operations based on our improvement experiences, which we’ve accumulated at thermal power generation facilities for years,” says Osawa. “Creating new businesses and responding to changing IT risks is not something that we can achieve alone; it can be achieved with cloud-based security technologies and specialized services from AWS. We look forward to receiving continued high-quality support for our pioneering initiatives.”