Case Study: Siemens achieves real-time, non-invasive transformer health monitoring with MathWorks MATLAB Production Server

Siemens Develops Health Monitoring System for Distribution Transformers

Siemens faced the challenge of monitoring oil-filled distribution transformers that degrade and leak over time while having minimal instrumentation; the key indicators—oil level and top-oil temperature—needed to be measured noninvasively and in real time. To address this, Siemens worked with MathWorks and leveraged MATLAB, MATLAB Production Server™, and the Statistics and Machine Learning Toolbox to develop and validate a physics‑based model that relates housing temperature measurements to oil volume.

Using MathWorks technology, Siemens deployed the model to the cloud via MATLAB Production Server™, with prototype IoT devices streaming temperature data to generate real‑time oil‑level estimates and predictive‑maintenance signals. The solution scaled from a lab proof‑of‑concept to a cloud implementation with minimal effort, enabling non‑invasive, retrofittable monitoring on brownfield transformers, user‑friendly commissioning, online learning of the algorithm, and faster time‑to‑deployment.

Siemens

Simit Pradhan

Siemens