SmartSwarm 300 Series
:
11
1.1 WHY ENRICH DATA?
Modbus data is impossible to interpret without very detailed knowledge of the devices producing the data and the
sensors connected to them. Anyone looking at Modbus data can see that the value of unit 31, register 40075 is
2397 – but has no way to interpret this data without prior knowledge of its significance. This is not the way that IoT
systems operate. One of the core concepts of an IoT architecture is that systems can request information based
upon a semantic model – a user can ask for information about temperatures in the rooms of the buildings they
manage, and will receive responses in a form which is self-declaring, for example B+B/Ottawa/Conference
Room{temperature: 72 degF}. This conversion of raw, unintelligible register values into interpretable information is
a fundamental operation in the integration of legacy devices into an IoT architecture.
1.2 WHY AGGREGATE DATA?
Modbus is a poll-response protocol. The master device follows a scan pattern which constantly updates an internal
database with the most recently recovered data in a particular unit, whether that data has changed or not. Once
we start to convert this data into information that is to be sent over, for example, a cellular data link, it becomes
important to regulate to flow of data to that which has value. The fact that the temperature in a room is the same
as it was five seconds ago is of little value, and we can make significant savings in data transmission and upstream
processing costs if we send aggregated data instead -- for example, the max, min and mean temperature each
hour.
1.3 WHY FILTER DATA?
Aggregating data is fine, of course, but there are certain events that we would want to be informed of on an
urgent basis. Examples would include a temperature that has exceeded a threshold, has an excessive rate of
change, or has moved by more than a deadband from the last transmitted value. This is the purpose of filtering. A
series of event triggers may be configured and the recovered data compared against these triggers with any match
resulting in an immediate action.
1.4 SAMPLING THEORY
It is important to bear in mind that the SmartSwarm 351 is only eavesdropping on the Modbus network. It cannot
influence the Modbus Slaves, or the Modbus Master in any way. Effectively, the SmartSwarm device is two levels
removed from the actual process signals in which you may be interested. This is especially important if the original
signal is analog.
The following example shows a “fast” analog signal that has some high frequency components: The signal is first
digitized by a Modbus Slave, at a certain sampling rate, and the quantized values are saved in an Input Register:
This register is then polled by the Modbus Master, at another (slower) sampling rate, and the response values are
used by a SCADA system:
The SmartSwarm 351 eavesdrops on the communication between the Modbus Master and Slave.
It can only observe the same data that the Modbus Master observes. If the Slave sampling rate and/or the Master
sampling rate is not fast enough to capture an event of interest, then that event will be missed.
