Advancing an existing project.
Forecast → detect → decide. Forecasting anticipates future operating conditions, anomaly detection identifies faults and abnormal behaviour, and the dispatch layer determines the appropriate response to both.
IoT Integration
Reliable measurement is a prerequisite for any data-driven analysis. We instrument the plant and ensure dependable data acquisition and transmission.
- Edge sensing & metering
- Secure MQTT / TLS uplink
- Gateway to PLC, inverter and BMS
- SQL databases on edge and cloud servers
- Historian & time-series storage
Forecasting
Load, generation and price forecasting using sequence models based on recurrent and transformer architectures.
- Day-ahead generation forecasting
- Short-term load prediction
- Long-term seasonal outlook
- Benchmarked against measured output
Anomaly Detection
Machine-learning models that identify faults, degradation and abnormal behaviour at an early stage, when corrective action is least costly.
- Condition monitoring
- Drift & degradation tracking
- Fault classification
- Prioritised alarms to prevent alarm flooding
Uncertainty Management
A single-point forecast is insufficient for operational planning. We quantify forecast uncertainty and schedule dispatch accordingly.
- Probabilistic & interval forecasts
- Scenario generation & reduction
- Reserve sizing under uncertainty
- Deep reinforcement learning and MILP dispatch
SCADA
Supervisory control and data acquisition for the plant, deployed online or as an offline closed system. The operator interface is based on our solar forecasting dashboard, in which every prediction can be replayed hour by hour against measured outcomes.
- Supervisory monitoring and control of field devices
- Online or offline closed-system deployment
- Live single-line diagram
- Prediction versus actual replay
- Cumulative error as it evolves
- Exportable logs & reports
Custom Hardware
Where the required instrumentation is not available, we develop it, including embedded firmware, sensor integration and custom PCB assemblies designed for the measurements the models require.
- Microcontroller firmware
- Schematic capture & PCB layout
- Design for manufacture
Operator interface demonstration
A working interface running on measured data. Select a plant and an architecture family, then replay the forecast horizon. The error metrics are cumulative over the hours replayed, reflecting how an operator would monitor accuracy through the day.
Edge to closed-loop control
Have operational data and a control challenge?
We will assess which aspects can be learned from the data, which can be optimised, and which functions should run at the edge.
Start a project →