AI Flexibility algorithm in LV /MV Substation Success Story

Enhancing Grid Intelligence through Edge Computing

Distribution System Operators (DSOs) are under growing pressure to integrate an increasing number of Distributed Energy Resources (DERs) into the grid, especially on the Low Voltage (LV) side, where digitization is still lagging behind. This gap directly threatens the resilience and flexibility of the entire distribution network.

The challenge is even greater in Medium and Low Voltage networks, where thousands of geographically dispersed Transformation Centers require real-time data to operate effectively. But traditional Cloud solutions fall short due to latency, cost, and limited scalability.

Meanwhile, SCADA systems, while excellent for automation, rely on proprietary data structures and lack the flexibility and openness needed for modern grid operations.

The answer? Edge Computing, the sweet spot between the Cloud and SCADA. It delivers:
✅ Real-time intelligence at the source
✅ Lower latency and bandwidth use
✅ Better scalability for thousands of nodes
✅ Increased interoperability with existing systems

Deploying AI at the Edge. The Ultimate Goal

‍

Digitizing the Grid: Cuerva’s Journey to a Smarter Distribution Network

Cuerva’s project stems from a clear mission: digitize operations and gain a unified, real-time view of the entire distribution network. This involves integrating a variety of devices typically found in a transformation center, such as concentrators, SABTs, SAMTs, and MV remotes, across multiple voltage levels, all converging at a single point of intelligence.

The goal?

🔍 To better predict grid behavior.

⚡️ Proactively address congestion and over voltage issues.

⏱️ Accelerate decision-making with real-time data events.

One of the most innovative aspects of the project is the development of local flexibility markets for DSOs, powered by AI and predictive algorithms. These tools allow Cuerva to anticipate supply quality issues and respond dynamically, ensuring a more resilient, flexible, and future-ready grid.

‍

‍‍Gridfy has developed and implemented the AI flexibility algorithm located in one of Cuerva´s substation and Barbara has provided the Edge Platform that deploys and orchestrates Artificial Intelligence in the Edge and enables MLOps techniques to facilitate the algorithm retraining in the Edge.

‍

‍

Physical Connections

The Edge Node is connected to the existing communications switch (fiber optic in particular) in the transformation center, with an IP within Cuerva's private subnet. The connection to each equipment is as follows:

1. Concentrator. It communicates through Web Services SOAP, calling the IP that the concentrator has enabled to respond to these Web Services.
2. SABT. There is a physical connection through RJ-45 between the Node and the SABT, calling the Node through Web Services SOAP to the local IP enabled.
3. SAMT. There is a physical connection via RJ-45 between the Node and the SAMT, with the Node calling via IEC 60870-5-104 protocol.
4. MV Remote. There is a physical connection via DB9-RS485 between the Node and the Remote, calling the Node via MODBUS-RTU protocol.

Data is transmitted via an MQTT broker to an InfluxDB database, specifically designed to efficiently handle large volumes of time-series data. For visualization, the data is seamlessly displayed in Grafana, which integrates with InfluxDB to provide real-time dashboards and synoptics for clear, actionable insights.

Installed Applications

The administration of installed applications is carried out by  Barbara Panel designed to manage edge devices deployed in the field. The following image illustrates the array of applications deployed within the Edge Node run by Barbara´s technology.

• General applications: naming configuration, communications, node telemetry, etc.
• Industrial Connectors: Web Services SOAP, IEC 60870-5-104, MODBUS-RTU, Broker MQTT
• Databases: InfluxDB
• Data visualization: Grafana
• Proprietary algorithm: including the flexibility algorithm described above.

‍

Barbara Edge AI Platform for Smart Grid

The proposed solution is composed by: 

• Barbara OS, the Cybersecure by design OS, running on the edge nodes (hardware). A Linux operating system that has been conceived to facilitate the connection of any industrial equipment to any business logic, on the cloud, public or private, and to run Edge Computing applications.

• Barbara API: It is  the communication interface with the Edge that enables the use of all Barbara platform products. It allows Gridfy to manage their fleet of edge nodes, install updates and  manage the lifecycle of Edge Computing applications.

• Barbara Panel:  A web-based management dashboard that provides centralized control over edge deployments.

Key Features include: 

o Device Lifecycle Management: Remotely configure, update, and monitor device health across large fleets.

o Orchestration: Deploy, update, and manage Docker-based applications across distributed edge locations.

o Batch Operations: Execute commands across multiple devices simultaneously.

o Security Controls: Manage roles, access, VPN configurations, and device security policies.

o Telemetry & Alerts: Real-time monitoring of device health metrics, enabling proactive issue management.

‍

‍

‍

‍

Want to know more? Download the success story here‍

About Barbara

Barbara is the Edge AI Orchestration Platform for organisations seeking to overcome the challenges of deploying AI in highly distributed environments.

Barbara empowers OT managers with IT capabilities to digitize processes, enabling seamless data capture, processing, and real-time decision-making.

With Barbara, companies can deploy and monitor AI models in minutes by simply exporting and uploading them to the platform.

Digitization teams need efficient AI orchestration, while infrastructure teams require secure modernization without disrupting existing systems. Barbara enables both, orchestrating AI applications across distributed locations, modernizing infrastructure, and ensuring security, all without replacing existing components.

Our capabilities include:

.– Industrial Connectors for legacy or next-generation equipment.

.– Edge Orchestration to deploy and control docker-based applications across thousands of distributed locations.

.– Edge MLOps, to optimize, deploy, and monitor trained models using standard or GPU enabled hardware.

. – Remote Fleet Management for provisioning, configuration and updates of edge devices.

.– Marketplace of Certified Edge Apps ready to be deployed.

Contact us to discuss how we can help accelerate your Edge AI projects.

‍

‍