The Most Dangerous Instrument Failures Are Not the Ones That Trigger Alarms
That night, every instrument was showing normal values.
But the company lost more than USD 100,000 in one day.
This was one of the most memorable incidents I encountered in a fine chemical plant.
It happened around 1:00 a.m.
The operator performed the routine inspection as usual.
Pressure was normal.
Temperature was normal.
Level was normal.
Flow rate was normal.
The DCS screen was completely green.
No alarms.
No interlocks.
No abnormal signals.
Everything looked exactly like a normal night shift.
But the next morning, the quality department called production:
“The entire batch has failed the quality inspection.”
The whole plant immediately started searching for the cause.
The process team suspected raw materials.
The mechanical team suspected equipment problems.
The instrumentation team was called to investigate.
The first question from the instrument engineer was:
“Before blaming the instrument, let’s check whether the data is telling the truth.”
After a full day of investigation, the real problem was discovered.
The temperature transmitter was not damaged.
The PLC was working normally.
The DCS system had no problem.
The actual failure was much simpler — but much more dangerous:
The temperature sensor installation position had changed.
After maintenance work, the temperature probe had not been fully inserted into the correct measurement position.
The actual process temperature was already 190°C.
However, the instrument was only measuring 176°C.
There was:
- No over-range signal
- No broken sensor
- No communication failure
- No alarm
Every value looked reasonable.
But the process had already moved away from the optimal operating condition.
As a result, dozens of tons of products were scrapped.
Later, the plant manager said:
“The most dangerous failures are not the ones that trigger alarms. They are the ones that look completely normal.”
Why Do the Most Dangerous Failures Often Have No Alarm?
Many new engineers have a common misunderstanding:
“If there is no alarm, there is no problem.”
However, experienced engineers know that many serious process failures happen without any alarm.
Why?
Because an alarm can only tell you:
“The measured value has exceeded a preset limit.”
But it cannot tell you:
“Is this measurement value actually correct?”
A normal-looking signal does not always mean a healthy process.
Hidden Failure Type 1: The Instrument Works, but the Measurement Is Wrong
This is one of the most common problems in industrial plants.
Examples:
Pressure transmitter
The zero point slowly drifts.
The transmitter still outputs a signal.
The display still works.
But the pressure value may already have a significant error.
Flow meter
The flow meter may gradually lose accuracy.
A 5% measurement deviation may not trigger any alarm.
However, in a precise production process, this error can directly affect product quality and energy consumption.
pH analyzer
The electrode becomes aged.
The response becomes slower.
The instrument still displays a value.
But the chemical process may already be out of control.
Oxygen analyzer
The response time increases due to sensor aging or contamination.
The reading appears stable, but the actual oxygen concentration may have changed.
The most dangerous part is:
Operators continue trusting the data because the display looks normal.
Hidden Failure Type 2: The Valve Position Is Correct, but the Process Output Is Wrong
A control valve once showed:
Valve opening: 100%
The DCS command was 100%.
The positioner feedback was also 100%.
Everyone believed the valve was fully open.
However, after dismantling the valve, technicians discovered:
The valve plug was heavily scaled.
The actual flow passage opening was less than 60%.
The control system believed:
“I have completed the command.”
But the process was still suffering from insufficient flow.
This is why:
A correct feedback signal does not always mean the equipment is performing correctly.
Hidden Failure Type 3: Every Individual Data Point Looks Normal, but the Process Is Changing
Many process problems do not happen because instruments are wrong.
They happen because engineers only look at individual values instead of analyzing relationships between data.
For example:
Pressure remains normal.
Flow remains normal.
Level remains normal.
Everyone thinks the system is healthy.
But at the same time:
- Temperature slowly decreases
- Motor current gradually increases
- Steam consumption rises
Each parameter alone looks acceptable.
But together, they reveal the real problem:
The equipment may already be developing fouling, blockage, or efficiency loss.
Experienced engineers do not only watch numbers.
They watch trends and relationships.
Hidden Failure Type 4: Calibration Passed Does Not Mean the Field Measurement Is Correct
Many companies calibrate instruments once a year.
Calibration certificates are all qualified.
Everyone feels confident.
But field problems can still exist:
- Incorrect installation direction
- Blocked impulse lines
- Liquid accumulation in pressure tapping points
- Failed heat tracing
- Incorrect sensor location
The instrument itself may be highly accurate.
But the measurement system can still provide incorrect information.
Experienced engineers often say:
“Calibration checks the instrument. Field operation checks the entire measurement system.”
Hidden Failure Type 5: The Most Dangerous Problem Is “It Has Always Been Like This”
There is a very dangerous sentence often heard in plants:
“It has always been like this.”
Many accidents start with this sentence.
Examples:
“The pressure has always been slightly low.”
“No problem.”
“The temperature has always been two degrees higher.”
“No problem.”
“The flow has always been slightly different.”
“No problem.”
Until one day:
All these small deviations accumulate.
And they become a major failure.
The biggest enemy in industrial operation is not abnormal conditions.
It is:
People becoming accustomed to abnormal conditions.
What Do Experienced Instrument Engineers Really Look For?
A new operator checks numbers.
An experienced engineer checks changes.
They ask:
- Why is steam consumption higher than yesterday?
- Why does motor current increase at the same production rate?
- Why does level change differently even when flow remains constant?
- Why does the pump sound different while pressure is normal?
These questions are not answered by DCS.
They are not answered by PLC.
They are not answered by alarms.
They require engineering experience.
A Good Engineer Never Trusts Only One Measurement
Professional engineers always perform cross-checking.
Pressure normal?
→ Check whether flow matches.
Flow normal?
→ Check pump current.
Level normal?
→ Check material balance.
Temperature normal?
→ Check product quality.
One measurement can be wrong.
But multiple related parameters rarely lie at the same time.
The Future of Instrumentation: From Alarm Response to Predictive Maintenance
Over the years, many engineers realize one important fact:
The biggest enemy in industrial plants is not equipment failure.
It is:
A failure that has already started but has not been detected.
Traditional maintenance focuses on:
Alarm → Response → Repair
Modern industrial reliability focuses on:
Trend analysis → Early warning → Preventive action
The most valuable engineers are not those who repair failures fastest.
They are the ones who can identify problems before failures happen.
Final Thoughts
Do not only watch alarms.
Learn to analyze:
- Trends
- Relationships between measurements
- Process behavior
- Equipment performance
Because in industrial plants:
The most expensive failures are often the ones that look completely normal.