ABSTRACT
Design and implementation of Internet of Things (IoT) systems require
platforms with smart things and components. Two dominant architectural
approaches for developing IoT systems are mashup-based and model-based
approaches. Mashup approaches use existing services and are mainly suitable for
less critical, personalized applications. Web development tools are widely used
in mashup approaches. Model-based techniques describe a system on a higher
level of abstraction, resulting in very expressive modelling of systems. The
article uses Cisco packet tracer 7.2 version, which consists of four
subcategories of smart things—home, smart city, industrial and power grid, to
design an IoT based control system for a fertilizer manufacturing plant. The
packet tracer also consists of boards—microcontrollers (MCU-PT), and single
boarded computers (SBC-PT), as well as actuators and sensors. The model
facilitates flexible communication opportunities among things—machines,
databases, and Human Machine Interfaces (HMIs). Implementation of the IoT
system brings finer process control as the operating conditions are monitored
online and are broadcasted to all stakeholders in real-time for quicker action
on deviations. The model developed focuses on three process plants; steam
raising, nitric acid, and ammonium nitrate plants. Key process parameters are
saturated steam temperature, fuel flowrates, CO and SOx emissions,
converter head temperature, NOx emissions, neutralisation
temperature, solution temperature, and evaporator steam pressure. The
parameters need to be monitored in order to
ensure quality, safety, and efficiency. Through the Cisco packet
tracer platform, a use case, physical layout, network layout, IoT layout,
configuration, and simulation interface were developed.