Wireless Sensor Network for Forest Fire Detection

Forest fires are one of problems that threaten sustainability of the forest. Early prevention system for indications of forest fires is absolutely necessary. The extent of the forest to be one of the problems encountered in the forest condition monitoring. To overcome the problems of forest extent, designed a system of forest fire detection system by adopting the Wireless Sensor Network (WSN) using multiple sensor nodes. Each sensor node has a microcontroller, transmitter/receiver and three sensors. Measurement method is performed by measuring the temperature, flame, the levels of methane, hydrocarbons, and CO2 in some forest area and the combustion of peat in a simulator. From results of measurements of temperature, levels of methane, a hydrocarbon gas and CO2 in an open area indicates there are no signs of fires due to the value of the temperature, methane, hydrocarbon gas, and CO2 is below the measurement in the space simulator.


Introduction
Forests constitute a source of life for all living things. With the forest, symbiosis, and the chain can run our life. Wildfires are a serious problem facing today. It has been felt by the public and government. Land and forest fires in Riau province has a considerable effect on the occurrence of haze pollution across state boundaries. In general, fires in Riau province in peatlands dominate the region by 60%. Therefore smog constitutes natural phenomenon that commonly occurs during fire season and have an impact on neighboring countries such as Malaysia and Singapore.
To reduce this problem, the government has launched several programs oriented to an appeal or a legal sanction against any action that threatens the sustainability of the forest. However, although the entire work program has been designed, the forests still shows a pretty poor [1].
In the term of the current technological developments, sustainability programs tend to require a system capable of analyzing and monitoring indication of forest fires. Wireless technology that is able to send data without using wires is expected to become one of the development of applied technology that can support forest conservation program. The monitoring system is expected to present data in the form of a fire indication for the vast land even.

Wireless Sensor Network Architecture
Each node WSN systems generally consist of sensing, processing, communication and power that can be illustrated as in Figure 1. How to combine this is the thing to watch out for when we do the designing. The system processor is the most important part of the WSN system that can affect the performance or energy consumption. Several options for the processor may choose include:  Microcontroller  Digital signal processor  Application-specific IC  Field programmable gate array  Figure 2. This sensor will detect UV light in the 185-260 nm interval detected as UV light from the fire. These sensors are not able to detect the size of the fire, because basically the cigarette at a distance up to 5 meters can be detected by these sensors. In realization, flame sensor is activated when the rated voltage of 350 Vdc, so it is necessary to activate C3704 kit. The input voltage required for this kit C3704 9-30 Vdc range, but if we have a fixed voltage of 5 volts, we simply menginputkannya to terminal 'O' on the kit because basically the input voltage of 9-30 Vdc earlier will be diregulator be 5 volts. The working principle of this series of kit C3704 is changing the supply voltage of 5 volts to 350 volts DC at the High Voltage DC to DC converter to activate the sensor. While Signal Processing Circuit serves to regulate the amount of incoming pulses from the UVTron sensor for 2 seconds will be responded by the C3704 a pulse width of 10 ms. Output pulses of 10 ms can be coupling the output capacitor to produce a wider period. In planning, it is used by 1 microfarads capacitor to produce an output of 1 second period on terminal Cx C3704 kit.

B. MQ-2 Gas Sensor
MQ-2 gas sensor is highly sensitive to substances SnO2, which is low conductivity in clean air. When flammable gas is detected, the sensor will be higher conductivity and concentration of the gas also increased. By using simple electrocircuit it will be converted according to the conductivity change of the output signal of gas concentration. MQ-2 sensor also has high sensitivity to LPG, Propane and Hydrogen, and usually areused for methane and flammable vapors and more. MQ2 is suitable for smoke and gas detection, also relatively inexpensive. The physical form of the MQ-2 sensor is shown in Figure 3.

C. LM35 Temperature Sensor
LM35 temperature sensor is a transducer that has a function to convert the temperature scale into electrical quantities in the form of voltage. LM35 has high accuracy and ease of design when compared to the other temperature sensor, LM35 also has a low output impedance and high linearity so that it can be easily connected with a series of special control and does not require further adjustment. The temperature can be measured by the IC is quite wide, between -55 C to 150 C. Figure 4 shows the basic circuit LM35.

D. Node Sensor
System sensor node consists of 3 pieces of sensors, microcontrollers and RF module KYL 500s, as shown in Figure 5. To be able to send serial data over the air at least need a device that can make the process of laying a digital serial data to the carrier frequency to a higher frequency and then emitted into the air. One example of the RF module that can do this is the module KYL 500S In one module can be used as the sender and receiver. Serial data to be transmitted through the RF module KYL 500S fed to the microcontroller serially. Similarly, the data received will be taken by a serial microcontroller. The distance could be gone through about 100 m -5 Km. Baudrate in the air 1200 bps, 2400 bps, 4800 bps, 9600 bps, 19200 bps, or 38400 bps. Modulation GFSK (Gaussian Frequency Shift Keying) is used so it is anti-interference and BER (Bit Error Rate) is low. The physical form of the RF transceiver KYL 500S module is shown in Figure 6. ATMEGA8535 microcontroller serves as the central data processor temperature, smoke and flame which will then be sent to computer through wireless communication channels. At each node there are 3 pieces of sensors are temperature sensors, smoke and flame and RF transceiver module KYL equipped 500s. Circuit at each node is shown in Figure 9.
While the modules that are on the PC as a server are functioning wireless Tx/Rx KYL 500s that function receives the data from the meter and then sent to a computer through a serial cable.

Research Method
To detect forest fires then performed measurements of temperature, flame, levels of methane, gasoline, CO and CO2 from the burning peat. There are several scenarios to perform these measurements, namely: 1. Measurement without burning peat in the space simulator Fire sensing system uses three sensors to detect whether a condition can be considered as early symptom of fire or not. In this fire sensing system used 3 pieces of sensors that will detect fire, smoke and temperature changes. Then the sensing system is put in an enclosed space simulator and temperature measurements, levels of methane, gasoline, CO and CO2 in the space simulator.

Measurement of burned peat in the enclosed space simulator
At this stage of peat burned garbage incorporated into the simulator. The results of measurements from three sensors viewed and stored on a computer application program.

Measurement of levels of air in the city
At this stage, observation and testing in an open area in the town of Duri to take measurements of temperature, levels of methane, hydrocarbons and CO2 that can be used as an indicator of fire.

RF Module Testing
The test aims to determine the performance and signal propagation characteristics (pathloss, power density, power and sensitivity) of KYL 200U in shadowed areas (outdoor), and

Results and Analysis
In Table 2 can be seen results of measurements of temperature, levels of methane, hydrocarbons, CO and CO2 in a simulator without the burning peat. From the results of these measurements the magnitude of temperature, methane, hydrocarbons, CO and CO2 are stable during the process of collecting data and in accordance with normal air conditions.

Measurement of the burning peat in the space simulator
In Figure 8 it can be seen the results of measurements of temperature, and levels of methane for burning peat in the simulator. From the results of temperature measurements, the magnitude of the temperature rose to 34 ° C compared to 31 ° C with no burning peat in the simulator. While methane levels increase dramatically in an average of 7 ppm compared to only 1 ppm without burning peat in the simulator. In Figure 9, it can be seen the results of the measurement of hydrocarbon gas and CO2 levels for burning peat in the simulator. From the measurement of hydrocarbon gases increased dramatically to an average of 95 ppm compared to 41 ppm with no burning peat in the simulator. While the average level of CO2 equal to 97 ppm without burning peat in the simulator, although at one point had increased dramatically to 103 ppm.

Measurement of levels of air in the town of Duri
In Figure 10 it can be seen the results of measurements of temperature, and levels of methane in the open area in the town of Duri. From the results of temperature measurements, the magnitude of the temperature rose to 34 ° C compared to 31 ° C with no burning peat in the simulator. While methane levels rise to a point 4 ppm while then back down to 1 ppm with no burning peat in the simulator. Figure 10. Results of measurements of temperature and methane gas levels in the town of Duri In Figure 11 it can be seen the results of measurements of hydrocarbon gas and CO2 levels in the open area in the town of Duri. From the measurement of hydrocarbon gases increased quite dramatically to 80 ppm compared to 41 ppm with no burning peat in the simulator. While the average level of CO2 equal to 97 ppm without burning peat in the simulator. Figure 11. Results of measurements of hydrocarbon gas and CO2 levels in the city of Duri From the results of measurements of temperature, levels of methane, a hydrocarbon gas and CO2 in an open area in the town of Duri shows no signs of forest fires. It can be seen from the values of temperature, methane, hydrocarbon gas, and CO2 are still below the measurement results in a simulator. Although the levels of methane and other hydrocarbon gases considerably higher than normal, but this is more due to the factor of air pollution from heavy vehicles in the town of Duri.

Power Received Measurement
Measurements were performed in order to determine the trend of received power of the distance. Table 3 is the measurement of average power of each node receiver with 3 different baudrate.  Figure 12, it can be seen that the average power received at each node will be smaller if the distance is farther away, or it can be said that the quality of data received by the receiver will get worse the longer it along with the distance. Data can be viewed on the received power at node baudrate 1200 bps and 19,200 bps is very small, even including the standard noise (<-80dBm), but due to the sensitivity of the module KYL-200U reached 123dBm, then even such a small power receiver can still receive the data well. As for the baudrate 9600 bps, the received considerable power is still in good quality. In this test, the average maximum transmission distance in a shadowed area of the module is 310 m for baudrate 1200 bps, 230 m for baudrate 9600 bps, and 90 m to 19 200 bps baudrate. It is quite different from the maximum transmission distance of the modules on the condition of LOS (Line Of Sight) listed on the datasheet, which is 1000m.

Power Density Measurement
Based on the Fig 13, it can be seen that the power density will decrease with increasing distance between Tx and Rx. This is because the distance is inversely proportional to the power density.

Pathloss Measurement
Based on Figure 14, it can be seen that if the farther anatara Tx and Rx, the greater its pathloss well. This is consistent with the results of measurements of power levels received by Rx can be seen in tables 4.1 and 4.2 as well as 4.6 and 4.7 the previous chart, the greater the distance between Tx and Rx will be smaller then the average power received by the Rx for growing losses that occurred, following this equation 11,727 ln 56,999.  Figure 15 explains that the RSSI value generated will be smaller if the distance between the Tx and Rx farther away. Thus RSSI value is inversely proportional to the value of pathloss. RSSI value is obtained in accordance with the data received by the receiver power measurement before, that the greater the distance between Tx and Rx, the smaller the received power. That way, the data obtained from the results of measurements with data RSSI proportional calculation. Just received power nearing the calculated value of RSSI measurement results baudrate is 9600bps. By doing so, baudrate 9600bps are suitable for use in systems designed.

Conclusion
Temperature measurements, the levels of methane, gasoline, CO and CO2 can be used as a leading indicator for the early detection of forest fires. There are three methods of measurements taken, the measurement in the space simulator with and without burning peat, and measurements in the open air in the town of Duri. From the results of measurements of temperature, levels of methane, a hydrocarbon gas and CO2 in an open area in the town of Duri shows no signs of fires where the value of the temperature, methane, hydrocarbon gas, and CO2 is below the measurement results in a simulator with burning peat.
Average maximum transmission distance RF modules in a shadowed area is 310 m at baudrate 1200 bps, 230 m at baudrate 9600 bps, and 90 m at 19,200 bps baudrate. 5. The sensitivity of the module power is -99 dBm KYL 200U, for minimum power level of the module can still receive the data well is -99 dBm