5 min
Typical installation target per monitored switch machine
Switch Machine Monitoring
AI-assisted switch machine monitoring for rail operators that need fewer failures, fewer delays and better maintenance decisions.
PointGuard learns how each point machine behaves, detects abnormal movement patterns, and connects every warning to Sepolo service planning, spare parts forecasting and asset cost analytics.
Current signature analysis
Inspect the latest switch movement and compare the shape of the current curve with the learned baseline for that machine.
PointGuard
The business case is operational uptime, not another dashboard.
PointGuard is built for the conversation between engineering, operations and finance. Engineers need early technical evidence. The CTO needs predictable maintenance execution. The CFO needs fewer delay events, fewer emergency interventions, controlled spare parts and a clear annual cost picture per point machine.
24/7
Automatic monitoring of movement, current signature and abnormal behaviour
1 asset
One history for movements, anomalies, service visits, parts and cost
12 mo
Usage, cost and spare parts planning horizon in Sepolo
Customer footprint
Used by rail operators and infrastructure teams around the world.
PointGuard and RailGuard are already supporting real railway environments, giving operators a practical path from connected measurements to better service decisions.
How it fits into Sepolo
Switch machines are small assets with huge operational consequences. A failed point machine can stop trains, trigger emergency call-outs, damage passenger confidence and create revenue loss that is far larger than the cost of monitoring. PointGuard connects switch machine telemetry to Sepolo so rail teams can move from reactive maintenance to evidence-based intervention.
What PointGuard monitors on a switch machine
Movement and cycle behaviour
Track whether the switch machine completes each movement inside its learned operating pattern, and whether cycle duration begins to drift over time.
Current and RMS signature
Measure the current profile during the throw and calculate RMS/load indicators so engineers can see how hard the machine worked, not only how long it moved.
AI anomaly learning
Sepolo compares the latest movement with the learned model for that exact point machine and suggests anomalies for review, tagging and follow-up.
Usage and operational demand
Monitor low-frequency depot switches and high-demand commuter switches with the same asset-specific learning model, from a few movements per day to one movement per minute.
Asset-linked events
Keep movements, anomalies, incidents, service visits, replaced parts and technician reports connected to the same asset history in Sepolo.
PointGuard turns each switch movement into measurable engineering evidence.
Cycle time
Sepolo records when the movement started, when it stopped and how many seconds the switch machine needed to complete the throw.
Current curve
PointGuard samples current during the movement. Sepolo visualizes the shape so engineers can compare the latest movement with previous movements from the same machine.
Load indicators
RMS current, current area, P90, high-load duration and peak-hold time help show whether the machine is working harder than before.
Learned baseline
Sepolo builds a movement profile for each switch machine, so anomaly review is based on the behaviour of that specific machine rather than a generic threshold.
Usage count
Every movement contributes to asset usage history, so service planning can be based on real activity instead of calendar time alone.
Context signals
Signal quality, battery and environmental readings help separate device issues, weather effects and real mechanical changes.
PointGuard
A monitoring window engineers can actually use.
PointGuard is not only a sensor. It is a full operating view where rail teams can inspect the latest current profile, RMS behaviour, cycle duration trends, AI suggested anomalies and the service actions that follow.
Current signature analysis
Inspect the latest switch movement and compare the shape of the current curve with the learned baseline for that machine.
RMS and load calculations
Use RMS current, P90, current area and high-load indicators to understand whether the drive is working harder than normal.
Cycle duration and prediction
Follow cycle time changes over time and use predicted future duration to plan inspection before the machine becomes a disruption risk.
AI suggested anomalies
Review suggested anomalies, tag what matters, and teach the model which changes represent maintenance, failure risk or normal operation.
Copilot-style engineering support
Summarize risk, observations, recommended actions and follow-up questions so less experienced teams can review the evidence faster.
From switch movement to actionable maintenance signal.
Fast site installation
PointGuard is installed as a compact IIoT monitoring device at the switch machine and starts collecting movement evidence as soon as the machine operates.
The switch moves
PointGuard wakes when the switch machine draws current and records the movement with high-frequency samples.
Sepolo detects the cycle
Raw telemetry is grouped into a movement window and stored with duration, current values, current signature and usage information.
Patterns are compared
Sepolo compares the latest cycle with the learned history for that device to suggest anomalies, trends and unusual movement behaviour.
Maintenance is planned
Confirmed issues can become inspections, service visits or scheduled preventive maintenance in Sepolo, with asset history and spare parts connected.
PointGuard
Partner delivery
PointGuard projects are delivered with local rail and infrastructure partners where installation, validation, operator training and operational rollout require local railway experience.
PointGuard
How PointGuard helps railroad operators
Find deterioration earlier
Gradual increases in cycle time, current area or high-load duration can point to friction, adjustment issues, wear or mechanical load changes before the switch machine becomes unavailable.
Reduce delay risk and lost revenue
When abnormal behaviour is found early, maintenance teams can inspect the machine before a failure creates traffic disruption, knock-on delays, passenger impact and avoidable revenue loss.
Give engineers real evidence
Engineers can inspect actual current curves, RMS calculations, movement history and anomaly decisions instead of relying only on isolated alarms or subjective reports.
Schedule service from usage
Because every movement is counted, Sepolo can support maintenance plans based on actual switch activity, not only calendar intervals.
Forecast spare parts demand
Replaced parts can be linked to the asset and service visit, making it easier to predict which parts may be needed next quarter, in six months and across the year.
Measure maintenance cost per point machine
Sepolo can connect asset usage, service visits, engineer time, parts and vendor cost to show the annual cost profile for each monitored point machine.
See engineer utilization in real time
Service planning, dispatch and resource calendars help maintenance leaders understand who is available, who is overloaded and where specialist skills are needed.
PointGuard
A new category for rail teams ready to move beyond reactive point-machine maintenance.
Switch machine monitoring is still a young market. Many rail organizations know monitoring is the future, but they need a practical way to justify it, install it and connect it to daily maintenance work. PointGuard combines IIoT telemetry, machine learning, AI-assisted anomaly review and Sepolo service execution so operators can move from detection to documented action.
- Monitor switch machine cycle time and current signature over time
- Review anomalies and teach the system which patterns represent maintenance, failure or normal behaviour
- Support low-usage depot switches and high-demand commuter switches with asset-specific learning
- Link every signal to the asset, location, service visit, incident, replaced parts and technician report
- Support predictive maintenance decisions without separating monitoring from field execution or spare parts planning
Business value
Earlier warning when switch machine behaviour changes
Lower risk of switch-machine failures causing operational disruption and rail delays
Better maintenance decisions based on AI-assisted movement evidence and asset history
Usage-based service scheduling and a stronger spare parts planning foundation
Clearer cost, utilization and delay-risk arguments for CFO, CTO and engineering teams