Fibre Optic DTS Temperature Monitoring System for Power Cables & Tunnels

Rebe's Fibre Optic DTS Temperature Monitoring System serves industrial plant substations and distribution rooms, and is what plant teams searching for cable tunnel temperature monitoring end up deploying. It continuously acquires distance-temperature profiles from a fibre optic DTS device, maps fibre distance onto real cable assets, and returns temperature health verdicts that are explainable, actionable and auditable.

Need product specifications, use cases or references?

Call +86 130 4890 6506 for selection guidance and comparable project types Email us

2. The problem this solves

2.1 Why cable temperature needs continuous monitoring

Inside industrial plant substations and distribution rooms, 0.4 kV high-current main cables, critical-load feeders, parallel cable groups and dense cable tunnels run under sustained load. The recurring failure modes are:

• Joint and cable overheating — rising contact resistance, or uneven current sharing between parallel cables, drives a local temperature rise that spreads.
• Inspection coverage gaps — cable trenches, trays and interlayers are hard to reach for frequent manual temperature checks.
• Late detection — by the time a smell or a fault appears, the incident is already close.

This is the case for continuous coverage rather than a set of point sensors — including the cases where point sensors are the better answer, which the same page sets out.

2.2 The problem is not the sensor — it is the interpretation

A DTS device — distributed temperature sensing, where the fibre cable itself is the sensor — already produces a distance–temperature curve, within the limits of its spatial resolution. What operators actually lack is an answer to five questions:

Is there an anomaly right now? Where is it? How severe is it? What is the evidence? What should I do next?

Most installations leave those five questions to the operator, reading a raw curve. The result is predictable: threshold alarms fire constantly, operators stop trusting them, and when an incident happens there is no retained evidence to reconstruct it.

2.3 What this system does differently

Common outcome	This system
Raw distance-temperature curve handed to the operator	Five-level health verdict with the evidence attached
Distance in metres ("62 °C at 1,243 m")	Distance mapped to a real asset — building, distribution room, feeder, cable section, core
Alarm noise; operators mute it	One hotspot = one alarm. Acknowledge does not clear; silencing does not stop recording
"Normal" shown when data is missing	"Cannot determine" is an explicit verdict — the system never reports "no anomaly" when data is insufficient
Thresholds guessed; conclusions unexplained	Every verdict carries object, time, location, temperature, trend, hotspot, data quality, coverage, limitations and a recommended action
No evidence after an incident	Immutable evidence bundles, retained up to 3 years

---

3. Who it is for

Role	What they get
Substation / distribution room O&M	Which feeder, which cable, how severe, what to do — without reading a curve
Cable tunnel operators	Continuous coverage of dense tunnels and parallel cable groups that manual patrols cannot cover
Plant energy & safety management	Early hotspot detection plus a retained, auditable evidence trail

Applications: industrial plant substations and distribution rooms · underground cable tunnels and utility corridors · parallel cable groups · critical-load feeders.

A note on scope. This is a monitoring and decision-support system. It does not trip breakers, shed load, or execute any switching operation — see section 4.4. If you need automatic protection and tripping, that belongs to your protection relays, not to this system.

---

4. Core capabilities

4.1 Five-level health verdict

Verdict	Meaning
No anomaly found	Data, coverage, mapping and policy all satisfy the conditions for judgement, and no anomaly was seen
Attention advised	A trend or deviation worth watching
Anomaly present	A confirmed anomaly
Serious anomaly	Requires priority handling
Cannot determine	Data, coverage, mapping or policy is insufficient — the system does not fall back to "no anomaly"

This last row is the point. Reporting "normal" because you could not measure is the single most damaging thing a monitoring system can do. Rebe's design rule is: when the data is insufficient, say so.

4.2 Asset mapping

Fibre distance is mapped onto the real asset hierarchy: plant → distribution room → feeder → line section → physical cable → cable core → parallel cable group.

Two separations are maintained deliberately:

• Monitoring topology is separate from physical asset topology.
• The fibre route is modelled separately from the tunnel it runs through. (Which cable a hot spot belongs to is decided by the route model, not by the tunnel's geometry.)

This is what allows a verdict to say which cable rather than which metre.

4.3 Alarm governance

• A hotspot persisting across multiple scans produces one business alarm, not one per scan.
• Condition state and handling state are independent.
• Silencing an alarm does not clear it, does not change its state, and does not stop recording.
• Adjacent anomalous points are merged into a single spatial hotspot.
• Monitored quantities: absolute temperature, rate of temperature rise, spatial hotspot, temperature difference against neighbouring objects, temperature difference within a parallel cable group.

4.4 What the system deliberately will not do

These are capability boundaries, stated because they are part of the value:

• It does not execute breaker tripping, power interruption or automatic load shedding. High-risk operations are denied by default and require authorisation, an impact summary, and approval or two-person review.
• It does not overstate measurement capability. Our rule is simple: capability is never inflated according to sampling density.
• It does not give a definitive cause when current data is absent. Without current data it will not assert overload or connection-fault as a confirmed cause.
• It does not let an algorithm silently change a released policy. Formal policies are immutable and released under control.

4.5 Graceful degradation

The basic temperature monitoring function runs in full without any current data. When current data is available, load-correlated enhancement is enabled. This means a site can start with temperature only and add current correlation later, without replacing the system.

4.6 Evidence and retention

• Immutable evidence bundles: summary, key segments, raw references, quality, mapping, policy, version, limitations, checksum.
• Tiered retention: per-frame profiles 7 days · minute-level profiles 1 year · assessment and alarm evidence 3 years.
• Off-box backup is required — a NAS or independent storage. A directory on the same machine is not accepted as the only backup.

4.7 Unattended recovery

After a Windows restart, acquisition and platform access are designed to recover within 10 minutes without anyone logging in (a design target — see the specification table).

---

Want to know more?

We can provide: selection guidance · integration and network design guidance · product documentation on request

Call +86 130 4890 6506 for selection guidance and comparable project types Email us

Tell us your site conditions and what the data needs to connect to — we reply within one business day.

5. Technical specifications

5.1 Capacity and performance — design targets

⚠️ Every figure in this table is a design target, not a measured result. They describe what the system is engineered to do. For what the resolution and sampling figures mean — and why they are not the same thing — see the glossary entry.

Metric	Design target
Channels per device	4
Points per channel	~25,000
Scan cycle	5 seconds
End-to-end verdict latency	≤ 10 seconds
Loss-of-communication detection	≤ 15 seconds
Recovery after Windows restart	≤ 10 minutes, unattended
Host recommendation	Windows host · 8 cores · 32 GB RAM · 1 TB SSD minimum

5.2 Deployment

A single Docker Compose stack (modular monolith + PostgreSQL 16 + MinIO) on Windows with Hyper-V.

5.3 Device integration

• Primary protocol: Modbus TCP with full function-code support; Modbus RTU as fallback.
• One station address corresponds to one fibre channel.
• Paged reads with before/after refresh-time validation to prevent frame mixing across scans.
• Loss of communication is identified after 3 consecutive failures, or 15 seconds without data (a design target).
• A DTS simulator is provided for integration and testing without hardware.

5.4 Field verification — actually performed

Unlike the design targets above, this is a real, recorded result.

The system has completed a read-only field integration with a fibre optic DTS device. All four channels returned complete temperature profiles.

In one stable acquisition: point spacing 0.20 m, 507 temperature points in a single frame, temperature range 30.4–32.8 °C (mean 31.67 °C).

The evidence is fixed with SHA-256 checksums.

---

6. Delivery, support and what we need from your site

Item	Detail
What you receive	The temperature monitoring software platform (licensed) plus integration into your DTS device and your asset model
Form of delivery	Software deployment on a Windows host you provide, plus engineering configuration
Prerequisites	A fibre optic DTS device accessible over Modbus TCP; the fibre installation and its as-built routing
Integration work	Mapping fibre distance to your asset hierarchy; policy definition; alarm routing
What we need from you	Cable schedule and as-built routing, distribution room layout, DTS device model and register map, current-transformer availability

Commercial note. Because this system is a software platform plus integration rather than a boxed instrument, pricing is quoted per project — software licence plus engineering days. Ask us for a scope-based quotation.

---

7. FAQ

8. Contact