Smart Devices Diagnostics
Device self-diagnostics can be divided into two classes: continuous and on-demand.
Continuous self-diagnostics runs autonomously entirely onboard (inside) the device all the time without user request. Continuous self-diagnostics is non-intrusive, not affecting the process.
Examples of continuous self-diagnostics include:
- Performance diagnostics inside a valve positioner
- Electrode signal fault in magnetic flowmeter
- Thermocouple degradation detection in a temperature transmitter
- Statistical process monitoring in a pressure transmitter
On-demand diagnostics only runs when requested by the technician. On-demand is used if the diagnostics affect the process (requiring process shutdown or bypass), or if the device is not capable of processing the sensor data, but only collect raw data from onboard sensors which require analysis in computer software.
That is, on-demand diagnostics is not continuous. There must be a procedure in place for technicians to schedule and periodically run such diagnostics. On-demand diagnostics include:
- Meter verification in a Coriolis mass flow meter
- Valve signature in a valve positioner
By deploying intelligent devices with centralized Intelligent Device Management (IDM) software, institutionalized in daily work processes plants can reduce downtime and maintenance cost.
Ability to diagnose if the device needs calibration or adjustment
Many transmitters, flowmeters, and analyzers around existing plants are out of calibration but nobody knows which ones. Similarly, some of the control valves may be hunting due to incorrect tuning of the positioner.
Device self-diagnostics can detect when instruments or valves need servicing and notify maintenance technicians centrally in the workshop which can schedule servicing at a convenient time.
For instance detect if pH sensor or flowmeter is likely to have drifted out of calibration, or if the valve is hunting. A temperature transmitter can detect when the temperature sensor has drifted and needs replacing.
That is, device diagnostics when institutionalized in daily work processes can be used to help ensure devices are performing at their best, ensuring on-spec product, more efficient operation, and lower energy consumption etc.
Ability to diagnose a device does not need service
In a typical plant, far too often instrument technicians are called out to inspect instrumentation in the field, only to find there is nothing wrong with it.
By instead checking the device’s self-diagnostics centrally from the IDM software before going to the field, technicians can tell if going to the field is really required or not. For example, it is possible to tell if the transmitter or sensor etc. are healthy or have failed have to be replaced etc.
That is, device diagnostics when institutionalized in daily work processes can be used to reduce the number of times technicians have to go to the field, saving time and resources which is better spent on devices which actually need service.
Ability to diagnose what service the device requires
When a device fails the technician does not know what the problem is, arriving at the device without much preparation, and without the proper tools and spares. The technician may have to return to the workshop to get it, prolonging the downtime, or may replace the whole transmitter.
By instead checking the device’s self-diagnostics centrally from the IDM software before going to the field, technicians can pin-point the probable cause, see recommended action, and bring the correct tools and spares to fix the problem quicker.
That is, device diagnostics when institutionalized in daily work processes can be used to reduce downtime due to instrument failure, and to make the most efficient use of the technician’s time.
Control strategy ability to distinguish a device problem from a process problem
In case of a sensor or device failure a transmitter with 4-20 mA signal will set its output to 20 mA (or 4 mA in some applications), masquerading as a high process variable, causing the loop to shutdown even though there is nothing wrong with the process.
Process variable status, derived from device self-diagnostics, communicated in real-time together with the process variable can be used by the control strategy to distinguish a device problem such as a sensor failure from a process problem such as high pressure or temperature.
Loops can now configured to go to manual mode in this case and centrally alarm the operator. On this alarm the operator can first take action to take care of the process, and then inform maintenance technicians to fix the device. The loop can optionally be configured to shut down on device failure. Moreover, the status gives operators greater confidence in the validity of the measurement.
This functionality requires real-time digital communication of process variable status together with the value. FOUNDATION fieldbus and PROFIBUS-PA meet this requirement. It is also required that the process variable status is supported in the DCS controller function blocks.
That is, device diagnostics when utilized in the control strategy can reduce downtime due to instrument failure.
Ability to diagnose and predict device failure before it actually fails
Usually devices are allowed to run until they fail. The device failure in turn causes the process to shutdown resulting in downtime.
A temperature transmitter can detect not only when a thermocouple has failed, but the early signs when it is thinning before failure. A valve positioner can detect when a valve stem packing needs replacing due to wear and tear etc.
That is, these are examples of predictive alarms before the device actually fails. Device diagnostics alarm monitoring capture these alarms and centrally notifies the maintenance technicians that can order the required spares and schedule maintenance at a convenient time.
That is, device diagnostics when institutionalized in daily work processes can be used to detect degradation early and service before failure materializes thus reducing downtime from a major unplanned shutdown to a minor scheduled outage.
Operator ability to quickly tend process when device fails
A failed device may within hours or minutes cause sequential logic for a batch operation to stop, ruining an entire batch, or may cause a control loop to malfunction making the product out of spec.
Such device failures may go unnoticed until the quality control department detects a problem with the product by which time raw material and energy has been wasted, and product has to be disposed of or reworked.
Device self-diagnostics detect failures almost instantly. Depending on the architecture, the networking can quickly communicate the diagnostics to the system operator consoles where the diagnostics can be viewed centrally by the operators.
Thus operators may have minutes or hours of early warning, enough time make changes before the process is affected by the device failure. The operator will first take care of the process, and then call upon the maintenance technician to fix the device.
That is, device diagnostics when institutionalized in daily work processes can be used to avoid shutdowns and to fix device problems sooner.
Early warning to operators of impending process problems resulting from device failure requires fast device alarm monitoring to quickly communicate the diagnostics from the device to the system.
For plants using 4-20 mA/HART devices, the polling time to report the diagnostics depends on the system architecture. In a system with centralized software polling devices one-by-one, the time to poll all devices can be more than an hour. Newer systems have HART polling decentralized.
A microprocessor in the 4-20 mA/HART I/O card polls the HART devices connected to that card (8, 16, or 32 channels) and use report by exception to communicate device problems to the system software in less than a minute depending on the number of I/O channels in the card. For FOUNDATION fieldbus and PROFIBUS the diagnostics is reported within seconds.
Early warning to operators of impending process problems resulting from device failure also require diagnostics display integrated in the DCS operator console.
Since third-party software is not allowed on a DCS in order to ensure the system remains robust, driver software components from the device manufacturers cannot be used. However, IDM software based on EDDL (a compressed text file, no executable software) is permitted on the DCS.
Ability to diagnose which devices to calibrate and tear-down at turnaround
At periodic plant shutdown in the past, preventive check and calibration was done on all devices, even removing valves and tearing them down. Many of them did not need it. This was a waste of resources.
A flowmeter with meter verification can be checked centrally if calibration is required. A positioner on a control valve be checked centrally if the valve has to be pulled and torn down for repair. This makes it possible to better schedule the plant shutdown.
That is, device diagnostics when institutionalized in daily work processes can be used to reduce preventive maintenance and schedule the plant turnaround shutdown more effectively.