Predictive maintenance of remote objects

Sudden changes in  process levels can cause stress and potentially lead to system failures. In example the opening or closing of valves in a water distribution network are sending shock waves through the network and depending of their magnitudes leading to rupture of pipes and fittings. Therefore it is import to monitor such transitions to be able to take preventive measures in time. Capturing transitions requires recording at high speed, causing continuous power consumption and an unmanageable amount of data, consequently very challenging for constraint low power remote monitoring devices to deal with. Our data loggers are provisioned with a low-power sub-system to detect and record such transitions on one primary analog parameter.

Don’t let things derail

One of the analog inputs can be used for digital and/or dead-band triggered 100Hz burst recording to
capture up to 5 seconds of a signal transition in detail.

Burst recorded details can be used to predict the state of the monitored system, allowing for timely repairs or replacements before any catastrophic malfunction occurs. For example, this feature could be used to monitor the current passing through an electric motor that drives a railroad track switch, where the currentprofile could tell something about the state of the switching mechanism. Or monitor the shock waves send thru a water distribution network on opening or closing of valves and starting and stopping of pumps (water hammer).

With this feature enabled a low-power subsystem continuously samples an analog port with 100Hz (even when the data logger is a sleep). A burst recording can be triggered by one of the digital inputs or when the sampled signal surpasses a configured dead-band value. E.g. when you want to record a sudden change in water level you could specify a dead-band value of lets say 0.2m and as soon as the level, since the previous regular interval recording, changes by 0.2m or more a burst recording of up to 5
seconds is started.

During burst recording all other data logger tasks are postponed. When burst-mode is enabled the data loggers quiescent current is increased by about 1mA and therefore not recommend for premium battery powered applications.

Resepecting service intervals

Since FW 3.2B1 we have already introduced a predictive maintenance feature, which allows for the accumulation of assertion time of digital inputs. This feature can, for example, be used to count the running hours of a pump or engine, enabling failure-preventive maintenance based on manufacturer specified service intervals. Our data loggers are equipped with 3 of such digital inputs.