Utilización del etanol

monitoring commonly use generic wireless sensor node platforms (i.e.Intel Mote), which limits their operational life due to the high power consumption of these platforms. Therefore, the nodes should be specifically designed based on their intended purpose in order to successfully deploy wireless sensor networks for long-term pipe monitoring. The design of the sensor nodes can be divided into two main categorises of hardware design and software (firmware) design. The main factors affecting the design of the nodes are hardware constraints, fault tolerance, scalability, overall cost, environment and power consumption. These constraints can affect the design of both the hardware and software of the nodes.

The hardware of a WUSN node can be divided into four main subsystems: Microcontroller Unit (MCU), transceiver, power management and signal conditioning (Figure 3.5). Each of these subsystems is responsible for a specific task in the operation of the node.

Figure 3.5 General schematic of the node and its subsystems (Sadeghioon et al., 2014a).

The MCU subsystem is mainly responsible for gathering the data from the sensors, processing them into a usable form, running the leak detection algorithms and buffering them into the transceiver. This unit is usually composed of an ultra-low power microcontroller and its required circuitry. The MCU subsystem is also responsible for time keeping. An internal

Sensor node Power& control& Transceiver& Signal& Condi4oning& Power& regulator& MCU& Power Management

Digital I/O Analog input

Po w er O u tp u t Po w er In p u t

timer (watchdog timer) or a real time clock (RTC) can be used for time keeping. The design of this subsystem can considerably affect the overall performance of the node.

A transceiver subsystem is responsible for connecting the node to the other nodes/master node. The operational characteristics (for example, RF frequency, RF output power and power consumption) of the transceiver are highly dependent on the application of the node (i.e. lower frequency for highly attenuating environments) .

Various types of power supply or energy harvester systems can be used to power the nodes and the power management circuitry is responsible for conditioning and managing the supplied power, in order to provide a usable power supply for different components of the board. The design of the power management system plays a major role in determining the power efficiency of the node as the majority of the losses happen during power conversion.

Depending on the application of the WSN’s nodes, various types of sensors (MEMS accelerometers, temperature sensors, pressure sensors) can be connected to them. The output of these sensors can be in the form of digital output, voltage, change in resistance, etc. In order to interface these outputs with the input of the microcontroller, usually a form of conditioning (for example, amplification, bridge and step change) is required. The signal conditioning subsystem is responsible for this task. An efficient and robust design of this subsystem is crucial for obtaining high quality data from the sensors.

In addition to the above mentioned aspects of the hardware and its design, the size of the node is also important, as the nodes should be small enough to be easily deployed on the pipes without a need for large excavation.

WUSNs are composed of a large number of individual nodes. These nodes can potentially fail due to hardware or software faults. These faults could be caused by various factors, such as individual component failure, software glitches, degradation due to the harsh environment or external damage caused by third parties. Fault tolerance of the sensor network is defined as the capability of remaining functional without any major disruption caused by node failure (Akyildiz and Vuran, 2010). Dense node deployment and higher standard components with a lower failure rate can potentially increase the fault tolerance of a WSN; however, they cause other problems, such as increase in cost of manufacturing and deployment. Other methods, such as systematic recovery algorithms, can be used to maintain the overall fault tolerance of the WSN (Akyildiz and Vuran, 2010). The desired fault tolerance of a network greatly depends on the criticality of the application, acceptable cost of manufacturing and deployment, and ease of deployment. In pipeline monitoring, the cost of dense deployment of the nodes on current pipes is high and infeasible. Therefore a suitable WUSN for pipeline monitoring should achieve high fault tolerance by improved node design and component selection. This is less critical for new pipes installation using conventional trenching techniques, as the nodes can be densely deployed without incurring major extra installation costs.

Pipeline systems such as water distribution systems can extend to thousands of kilometres. This creates a scalability issue for the monitoring of these pipes using WUSN. These networks should be able to handle all the information generated by the nodes effectively and efficiently. This imposes certain constraints on the network and data management design of the WUSN.

As mentioned earlier in this chapter, low overall cost is one of the main requirements for any pipeline monitoring system. The overall cost can be divided into three main parts: manufacturing cost, deployment cost, and maintenance cost. Manufacturing costs are directly related to the hardware design and component selection. Although using high performance components can be beneficial in terms of fault tolerance and system performance, they could significantly increase the production cost of the nodes and make dense deployment of the system economically not feasible. On the other hand, using higher performance components could reduce the failure rates and therefore reduce the maintenance costs. This creates a need for careful hardware design of the nodes in order to minimise the overall cost of the network system. In pipeline monitoring, the deployment cost of the PMS on existing pipes is a major part of the overall cost of the PMS. Pipes are usually buried and deploying the PMS at the pipe level can be very costly; therefore the system should be designed with deployment in mind to reduce the overall cost. This can be achieved by designing the installation of the nodes to be carried out via keyhole vacuum excavation techniques without a need for tappings in the pipes.

The environment of the nodes also greatly affects the design of the nodes. In underground pipeline monitoring, the environment of the nodes can be harsh and therefore the design and packaging of the nodes needs to be robust in order to survive for the desired lifetime. The underground environment also greatly affects the RF transmission of the nodes and they should be designed specifically for the underground environment in order to operate correctly. As mentioned, power consumption is one of the main challenges of any WSN. Power consumption is even more critical for wireless sensor networks in underground pipeline

design are involved in reducing the power consumption. A power consumption goal of 10µW for one transmission per 8 hours was chosen for the purpose of this research based on an average theoretical lifetime of >50 years on 2 AA batteries (although it should be noted that the current shelf life of the batteries is significantly shorter than 50 years). Efforts to minimise the power consumption of the developed node through hardware and software design are further discussed in detail in the next section of this chapter.