Recourses

• On-Prem Isolated EMS
• Resource-Intensive Model Training
• Maintenance and Management
• Data Privacy

Nikola Power has been offering a proprietary Energy Management System as an offline, on-prem software bundle interacting with local Energy Storage Units (ESUs) to minimize costs. Using AI to optimize energy consumption, the software needed to constantly train its model on massive amounts of data; a powerful computer was a requirement. This. of course, meant a higher initial cost for the customer, who now had to supply an industrial-grade computer dedicated to running the EMS.

Wanting to overcome these problems and willing to change their operational model, Nikola Power contacted us for a preliminary consultation. With feasibility and requirements becoming apparent, the company decided to implement their EMS as an IoT offering, making it a product as well as a service.

Nikola Power wanted to create a fully connected, cloud-enabled edge device capable of connecting to various ESU sensors, processing the data locally, and sending the data to the cloud for training and model updates. The company was also looking toward asset management and predictive maintenance capabilities with regard to the Energy Storage Units. All this had to be implemented as massively scalable and efficient. That meant designing a solution that could balance edge and cloud computing to avoid overburdening either when the company would scale up.

A further challenge presented itself as a question of data privacy and operational autonomy. In commercial and enterprise settings, it would be imperative to define and limit the permissions of each edge device to access specific resources and data.

• Edge Computing
• Model Training in the Cloud
• ML deployed at the Edge
• Analytics
• Web App

IoPIE was given the task of designing the architecture and implementing an end-to-end IoT solution, including analytics and a web app. We substituted the EMS software bundle, dependent on the customer's PC, with a semi-autonomous IoT device with average computational capabilities. Where EMS had to build its model by training on locally acquired data, the new product, EMS2, sends the data gathered from the ESU sensors to the cloud. There the data is anonymized and aggregated with other data to train a predictive model using machine learning and deep neural networks. The model produced at this point is then deployed to the edge device. The device in turn uses the model to control and optimize energy consumption in real-time.

The model in the cloud continues to be trained and refined with ever-increasing data, and the results are regularly deployed to all EMS2 devices. This keeps the model from getting outdated and obsolete, all done remotely and automatically.

IoPIE’s unique multi-tiered access policy was the perfect fit for the data privacy concerns of the client. IoPIE allows the client to define specific data and resource access rights, individually for each device, by device type or location, or by bulk.

• Upgrades deployed over the air
• Continuously Evolving Model
• Smoother Analytics
• Multi-tiered Service Potential
• Predictive Maintenance
• Asset Management

With the implementation of the new IoT solution using IoPIE, Nikla Power succeeded at turning their vision into a market-ready product within months. The new EMS2 system has exceeded initial expectations by offering extra capabilities that can be turned into business value.

EMS2 now allows Nikola Power to send software and firmware upgrades over the air, reducing the need for field service. Without the scheduling hassle and the understandably negative customer experience, EMS2 has even broader appeal and few market barriers.

Unlike its previous iteration, EMS2 features a continuously evolving optimization model drawing on numerous data streams for training. That means a much smarter, more data-driven model that delivers much better optimization, reducing costs and consumption far better than before.

The multi-tiered access policy for edge devices introduced by IoPIE has turned into a core feature of EMS2, its possible applications extending beyond data privacy. Defining rules and access levels for individual edges or edge groups has, of course, solved any concerns about data privacy, but the separate storage and access of data on the cloud has also led to improved analytics.

With the data semi-structured based on the access policy and separate storage of each edge device, aggregating and processing it is faster and easier. Analytics is already underway before data processing has begun and that goes a long way as the system scales up, saving time and money.

The final added value of the multi-tiered access policy involves the possibility of defining different service plans for different types of users. With the separate data gathering and storage mechanisms already in place, offering various levels of access and services to consumer-, commercial-, and enterprise-grade customers is more than simple — it's as simple as IoPIE.