Predictive maintenance for power plants
How can a power plant move its fans, pumps and crushers from reactive repair to predictive maintenance?
DHMSCare monitors blowers, pumps, air compressors and coal crushers.
Wireless vibration and temperature sensors sit on each machine.
A power plant typically needs about 100–300 sensors, depending on the number of machines.
Its industrial AI agent watches every machine around the clock.
It warns your team of a developing fault before it forces an outage.
Industry challenges
Blowers, pumps, air compressors and coal crushers run continuously for long periods. That raises their failure risk year after year.
A sudden equipment failure forces an unplanned outage. Generation efficiency and supply reliability both drop as a result.
Traditional maintenance relies heavily on emergency repair after a failure. There are few tools to predict failures in advance.
High temperatures and heavy loads accelerate the degradation of critical components. With limited monitoring, the early signs go unnoticed.
Quantifiable value
Monitoring solution
DHMSCare places sensors on the motors and bearing seats of blowers, pumps, air compressors and coal crushers across the plant.
Data reaches the platform through DHMSCare gateways, with DMG311 (4G) for areas without Wi-Fi.
From the day of installation, DHMSCare gives 24/7 condition monitoring, fault alerts and component-level diagnosis, with no months-long learning period.
Platform
DHMSCare uses AI to identify early fault signs such as gear wear and bearing anomalies, assess equipment health and provide maintenance recommendations.
Recommended measuring points and plan
What a typical production line needs, so you can see the whole plan before we talk.
| Asset | Suggested sensors | Measuring points (bearing seats) |
|---|---|---|
| Blower / fan | 4 | Drive motor DE and NDE, fan bearings DE and NDE |
| Pump | 4 | Motor DE and NDE, pump DE and NDE |
| Air compressor | 4 | Motor DE and NDE, compressor bearing seats |
| Coal crusher | 4 | Drive motor DE and NDE, rotor bearings on both sides |
| Gateway | DMG310 | DMG311 (4G) where there is no Wi-Fi |
| Per power plant | approx. 100–300 |
Actual quantities are confirmed on site
AI Care
The first plan with AI fault diagnostics, the AI that finds the gear wear and bearing anomalies behind these results.
Choose AI Care Plus for unlimited automatic diagnostics, remaining-life forecasts with Mina Ultra and 24 months data retention.
Smaller teams can start on Care: AI condition monitoring and daily reports with 10 user accounts, then move to AI Care when you need AI fault diagnostics.
Start with one line
The smallest sensible start, then grow line by line or roll out plant-wide.
Watch the machines you can least afford to stop.
You learn about faults days ahead.
All your lines, all sites, on one dashboard.
Tailored to your project's requirements.
Runs on your own servers or private cloud.
Contracted and delivered as a project.
Can DHMSCare monitor boiler feed pumps, fans and cooling-tower gearboxes?
Yes. DHMSCare monitors power-plant fans, pumps such as boiler feed pumps, and gearboxes; at a coal-fired plant it found early misalignment on a secondary air fan before it forced an unplanned shutdown. Pumps and gearboxes are on the supported list. Cooling-tower gearboxes need their tooth counts entered, and for slow fan shafts, tell us the shaft speed and we confirm the right sensor first.
Is it a fit?
Is DHMSCare suitable for turbines that need proximity probes and API 670 protection?
No. DHMSCare does not replace an API 670 machinery-protection system or its proximity probes; keep that system on the turbine and use DHMSCare on the plant’s pumps, fans and motors. API 670 is the industry standard for machinery protection systems, and proximity probes measure shaft movement directly. DHMSCare is a monitoring system, not a machine-protection system, and does not trip machines; its sensors are MEMS accelerometers bonded to the machine surface, and no proximity probes are offered.
Questions
We are a small or mid-sized company. Is this for us?
Yes. Start with one line and one gateway, subscribe per sensor per year, and begin on AI Care; Care if you only need condition monitoring.
How long does a one-line pilot take?
Installation follows the seven steps in our quick installation guide: create the equipment, mount the gateway, attach the sensors and bind the measuring points. We agree the pilot period with you by line size.
We have no vibration engineer. Can we still use it?
Yes. The AI identifies early fault signs, assesses equipment health and gives maintenance recommendations, and Mina tells you which machine to handle first and why.
Which equipment should a power plant monitor first?
Start with the blowers, pumps, air compressors and coal crushers that run continuously. Their failures cause most unplanned outages.
How many sensors does a power plant need?
About 100–300 per plant, depending on how many machines you cover. We size the count with you, machine by machine.
Do we have to collect months of data before the first alert?
No. DHMSCare gives its first diagnostic verdicts on the day the sensors are installed, with no months-long learning period.
View all questions → Help center
Related case studies
60 days from sensor to save: catching fan misalignment at a coal-fired power station
At a coal-fired power station of Fujian Energy & Petrochemical Group, DHMSCare caught early misalignment, when two connected shafts are out of line, on the secondary fan of Furnace No. 3. DHMSCare flagged it on 7 January 2025, eight days before it turned severe, and named the fault only after two independent signs agreed: rising vibration and the vibration pattern typical of misalignment. The plant acted in time and avoided an unplanned shutdown. About 60 days passed from installing the sensors in November 2024 to this first verified save.
- Unplanned shutdowns: down 65%, from 5.6 to 2 per year
- Mean time to repair: from 4.5 hours to 1.2 hours
- Manual meter reading and inspection work: down 70%
- Environmental compliance reporting: about 20 person-days saved per month
- About 60 days from sensor install to the first verified save
Tell us the line and its critical assets; we size the sensors and the plan with you.