An Automated Drip Irrigation System Based on Soil Electrical ...

PHILIPP AGRIC SCIENTIST Vol. 94 No. 4, 343-349 December 2011

ISSN 0031-7454

An Automated Drip Irrigation System Based on Soil Electrical Conductivity Murat Yildirim1,* and Mehmet Demirel2 1

Canakkale Onsekiz Mart University, Agricultural Faculty, Irrigation and Farm Structure Department, Terzioglu Campus, 17020 Canakkale, Turkey 2 Canakkale Vocational College, Canakkale 17020, Turkey * Author for correspondence; e-mail: [email protected] We assessed the irrigation performance of the automated irrigation controller. In the study, a drip irrigation system automatically governed irrigation in accordance with water consumption of the substrate-plant system. Data acquisition was performed by an electronic circuit, which processed data and then sent the data to the microcontroller (programmable integrated circuit-pic16f84). The pic16f84 functioned as a controller, which decided when and how much water to apply; hence, the pumps ran and stopped according to the irrigation strategy defined by the microcontroller. The required time to pump water according to the irrigation programs corresponded to the time to increase soil moisture up to field capacity in the full treatment whenever 30% of the available water in the substrate was depleted by the pepper plant (Capsicum annuum L.) in the experiment. Therefore, once defined, the microcontroller utilized the data and controlled the relays connected to the pumps. Soil moisture content was monitored by only one sensor installed in a representative pot throughout the experiment. The automated system applied four different water applications; one treatment was full and the other three were deficit treatments. These were compared with the control treatment. The automated system maintained the soil moisture level at the desired level for the full treatment and took over irrigation events, started and stopped the irrigations throughout the entire growing season. Even though yield value was high in treatment I1.0, the best quality parameters were obtained from I0.75. In the deficit treatments I0.50 and I0.25, yield and quality parameters decreased since plants in those treatments were under stress. The performance of the automated system can be increased as the time in the software is adjusted according to full irrigation application.

Key Words: automated irrigation, irrigation controller, soil moisture sensor Abbreviations: EC – electrical conductivity, ET – evapotranspiration, FC – field capacity, IWUE – irrigation water use efficiency, MCU – microcontroller unit, WP – wilting point, WUE – water use efficiency

INTRODUCTION Irrigation water is applied to the root areas of a crop by using different irrigation methods, one of which is drip irrigation. The current trend is toward switching from a manual system to automatic operations in a pressurized system. Energy savings, reduced labor cost and control in fertilizer application are among some of the major advantages in adopting automated techniques in drip irrigation systems. Automated irrigation systems provide high crop yield, save water usage compared with conventional systems (Mulas 1986), facilitate high frequency and low volume irrigation (Abraham et al. 2000) and also reduce human error (Castanon 1992). In order to increase irrigation efficiency in automated irrigation systems, a number of research studies have been carried out. Shull and Dylla (1980) used gypsum resistance block as a soil moisture sensor to activate the The Philippine Agricultural Scientist Vol. 94 No.4 (December 2011)

sprinkler irrigation system. Frankovitch and Sarich (1991) developed an electronic switching system to control pumping time for the irrigation system. Cuming (1990) controlled the common lines of various irrigation systems by using a soil moisture sensor. Many methods have been described and sensors developed to manage irrigation systems objectively (Weinstein and Avishai 1995; Biernbaum and Versluys 1998; Salas and Urrestarazu 2001; Yildirim 2010). Also, recent irrigation technologies used sophisticated equipment to supply water to the root area of plants as they need it. However, use of these sophisticated methods is not possible for all growers. Therefore, when developing an automatic irrigation system, it needs to be designed so it can be adapted by most growers. Abraham et al. (2000) developed an automated irrigation system based on soil electrical conductivity. The irrigation control tray is a commonly used method for automatically activating an 343

Automated Drip Irrigation System

irrigation system (Salas and Urrestarazu 2001; Moreno 2003; Gallardo 2005). Also, Caceres et al. (2007) tested and evaluated an irrigation control tray which activated the irrigation system with the aid of a level-control relay. The objective of this study was to test a prototype of the irrigation controller and to modify the irrigation controller, which works based on soil electrical conductivity, and thereby develop a simple and economical irrigation system appropriate for greenhouse growing of high-value crops.

MATERIALS AND METHODS The experiment was conducted outdoors from May to August, 2008 at Canakkale Onsekiz Mart University, Turkey. The geographical location of the experimental area is 40°06'32.64'' N latitude, 26°24'45.31'' E longtitude, and has a 5-m elevation. The site is influenced by the climates of the Mediterranean and the Black Sea. Normally, almost no rainfall occurs during summer. Pepper (Capsicum annuum L.) plants were grown in the nursery and transplanted in pots. The substrate was a mixture of peat (1:4, v/v) and soil (3:4, v/v). Peat was used for its excellent aeration of nurseries, and the soil consisted of sand (58%), silt (23%) and clay (19%); each pot contained 4 L of substrate. The chemical characteristics of the substrate and the quality of the irrigation water are given in Tables 1 and 2, respectively. A standard soil must have a pH between 6.5 and 7.2 and electrical conductivity (EC) of less than 4 mS cm-1 (Ayers and Westcot 1994). Based on these values, the salinity level of the substrate was in the normal range. The irrigation water, however, was in the moderately tolerable range; it had already been used for irrigation at the site. Each pot in the experiment was applied with the same amount of fertilizers: triple super phosphate (3 g per pot), potassium sulfate (3 g per pot) and urea (3 g per pot). Urea was applied again at 15- and 20-d intervals after planting at the same dosage. Components of the Automated Irrigation System In each treatment, the components of the irrigation system were as follows: water storage tank (250 L), submersible pump operating at 12 volts direct current (vdc) and 2.05A, power supply (12 vdc), pots (250 x 210 mm, 9 L) with four drainage holes, drainage canal, drainage tank (30 L), Ø16 pipes with drippers (4 L h-1) at a spacing of 33 cm, with one dripper serving each pot. Valves and connection apparatus were used to integrate all items of the irrigation system. The layout of the experiment is given in Figure 1. Each irrigation treatment included all the above listed items separately. The most important and basic component of the automated irrigation system was the sensor, which reacted to the changes in resistance depending on the soil 344

Murat Yildirim and Mehmet Demirel

moisture level. It was made of brass rods 7 cm long and 0.5 cm in diameter. The distance between the rods was 2.5 cm. They were placed in a plastic box (width 3 x 3 cm, height 1 cm), then filled with silicone. At the end of the rods, the cable was connected to provide an electrical communication between the rods and the circuit (Fig. 2). The electrodes were made of brass. Abraham et al. (2000) found brass to have the best performance in sensing moisture content compared with stainless steel, and copper. The electrodes were placed at a depth of 5 cm, parallel to the soil surface. Only one sensor was used and was installed within the root zone of a representative plant in the 100% treatment (full irrigation treatment I1.0), that is, irrigation events were achieved according to a signal produced by only one soil-moisture sensor. To determine the value of resistance corresponding to the critical soil moisture level, a calibration curve was established. The substrate was saturated, then the electrical resistance corresponding to different soil moisture levels was monitored up to the wilting point (WP) so that the corresponding values of resistance and soil moisture content could be plotted on a graph. The threshold value of soil moisture level to begin irrigation was determined as the point when 30% of the available moisture content in the substrate was depleted by a representative plant in the I1.0 treatment. The electrical resistance corresponding to that moisture content level was found to be 15 kΩ for the sensor. The resistance value was set to the pot (100K) in the circuit (Fig. 3), in which CD4011 (NAND gates provide NAND function) processed the signal and NE 555 (monolithic timing circuit) provided a time delay to prevent fluctuations. Therefore, inputs such as high or low voltage sent to the Microcontroller unit (MCU-Pic16f84) were in a stable form. The circuit also included a buzzer which sounded when irrigation began and switched from one relay to another. The MCU unit is a device that has programmable capability, reads sensor, and controls the devices such as relays connected to the pumps. In this experiment, the MCU was actually a controller which decided the time when the pump had to be operated and when it had to be stopped. The MCU had a 20 Mhz pic processor with 18pin Dual In-line Package (DIP) and 13 programmable I/O pins. Each pin was capable of carrying a 25 mA sink current and 20 mA source current. The MCU runs at a relatively low voltage value of 5 vdc (Altınbasak 2004). One pin of the MCU was assigned as an input to monitor the soil moisture sensor every second for the whole day and the entire experiment. Even though the circuit had 10 relays, four pins of the MCU were assigned as output pins to pump water to the root area of the plants, since there were four irrigation treatments in the experiment. Therefore, it is possible to increase the treatments by using this circuit. A signal was produced by the electrodes embedded in the pot whenever 30% of The Philippine Agricultural Scientist Vol. 94 No.4 (December 2011)



Table 1. Chemical characteristics of substrate used in the experiment. EC P K Ca Mg Cu Zn pH (mS cm-1) (kg da-1 ) (kg da-1) (ppm) (ppm) (ppm) (ppm) 6.0

1.37

2.26

613.8

8111

2500

3.12

5.20

Fe (ppm)

Mn (ppm)

Ca CO3 (%)

5.00

13.80

5.23

mS-milli Siemens da-1,000 square meters

Table 2. Quality of irrigation water used in the experiment. Cation (me L-1) Anion (me L-1) Na EC SAR pH -1 -1 1/2 RSC % (mS cm ) (me L ) Na K Ca Mg Total HCO3 CO3 Cl SO4 7.32 0.14

0.98

0. 67

None

1.37

0. 17

3. 7

4.6

9. 84

3.8

-

2.8

3.24

Total 9.84

me-milliequivalents RSC-residual sodium carbonate SAR-sodium adsorption ratio

Fig. 1. Layout of the experiment‟s components communication cables.

Fig. 2. View of the soil moisture sensor made of brass. available water capacity was consumed, which resulted in a high voltage at the input pin of the MCU. The MCU then ran the algorithm in its memory.

The Philippine Agricultural Scientist Vol. 94 No.4 (December 2011)

Controller Software The irrigation controller program was written using the PicBasic Pro software program and the general strategy for the automated irrigation defined in the logic was loaded into the memory of the MCU. Hence, the logic of the irrigation strategy was defined in the MCU, having a memory of 1K, which then took over and made detailed decisions on when to apply water and how much water to apply. The dosage of water to be applied was determined according to the pumping time of water to refill the root zone as moisture content dropped 30% below field capacity in the substrate. In the system, the feedback and control were done constantly, depending on the feedback from the sensor. Whenever a high input was sent to the MCU, the irrigation actions were carried out during the whole experiment period. The logic in the software was simple and maintained the desired state, that is, as the voltage rose at the input pin of the MCU, the output pins triggered the relays in turn (Fig. 4). Irrigation Applications The irrigation treatments were arranged as follows: the required time (10 min) for pumping water to the root area of the plant in the full irrigation treatment (I1.0) as 30% of the available soil moisture was depleted was the time required to raise moisture content of the substrate up to field capacity (FC) again. In the deficit treatments, water was applied in the range of 75% (I.75), 50% (I.50) and 25% (I.25) of the full irrigation. The required times for deficit treatments were the fraction of the time assigned to full irrigation. In the control treatment, whenever an irrigation event in the automated system took place, all pots were irrigated manually and weighed individually, then the water quantities were regulated by the weight of the pots intended to refill the root zone up to field capacity. Therefore, there was no leaching water in the control treatment. Daily evapotranspiration (ET) was estimated by using the water balance method between the two irrigations (Yurtseven et al. 2005).

345



12V 7805

POT2

1

R2 27k

470k

4

VCC

R

5

D1

5

LED R3

7

27k CV

1

4011

2

TR

GND

2

6

TH

BC237

555

6

TH

C6

1

TR

GND

2

100nF

555

C4

C2

0.01uF

4 MHz

C8

U5 16 15

Şase

PIC16F628A

RA7/OSC1/CLKIN RA6/OSC2/CLKOUT

4

RA5/MCLR

22nf

RA0/AN0 RA1/AN1 RA2/AN2/VREF RA3/AN3/CMP1 RA4/T 0CKI/CMP2 RB0/INT RB1/RX/DT RB2/T X/CK RB3/CCP1 RB4 RB5 RB6/T 1OSO/T 1CKI RB7/T 1OSI

Proplar

R7

10k

330

R17 KRIST AL

R6

R16

10k

10uF

C7 22nf

R15

C3

0.01uF

220uF

10k

Q11

100uF 3

C1

R5

BUZER

3

Q DC

CV

3

C5

1N4148

2k

R

U1:A 1

R1

4

7

DC

U3

8

3

Q

VCC

8

U2

VO

2

100k

U4

VI GND

POT1

R8

10k

R4

5V

10k

6 7 8 9 10 11 12 13

R9

10k

D12

10k

17 18 1 2 3

R10

10k

R11

10k

R12

10k

R13

10k

R14

10k

10k

Q1

Q2

Q3

Q4

Q5

Q6

Q7

Q8

Q9

Q10

BC237

BC237

BC237

BC237

BC237

BC237

BC237

BC237

BC237

BC237

1N4001

1N4001

1N4001

1N4001

1N4001

1N4001

1N4001

1N4001

1N4001

D2

D3

D4

D5

D6

D7

D8

D9

D10

1N4001

D11

RL1

RL2

RL3

RL4

RL5

RL6

RL7

RL8

RL9

RL10

12V

12V

12V

12V

12V

12V

12V

12V

12V

12V

Fig. 3. Circuit of irrigation controller. ET=[((Wi-1–Wi) + I–D)/A] i=1,2,3,……….n

(1)

where ET is the evapotranspiration (mm), Wi-1 and Wi are the mass (kg) of the pot at day i-1 and i, respectively, I is the amount of irrigation water (kg), D is the quantity of the drainage water if available (kg) and A is the pot surface area (m²). Water use efficiency (WUE) (kg m-3) and irrigation water use efficiency (IWUE) (kg m-3) were estimated as follows: WUE=Y/ET (2) IWUE=Y/I

(3)

where Y is yield (kg) and I is irrigation depth (mm). Yield was measured in terms of weight of fruits per plant; there were five plants in each treatment. Weights in gram for stem, leaf, etc. were determined by using sensitive weighing (0.01 g) and fruit diameters were measured by a digital caliper (to 0.01 mm).

RESULTS AND DISCUSSION

Fig. 4. Flow diagram of controller program in the microcontroller unit (MCU). 346

Electrical Resistance and Critical Moisture Level The response of the electrodes in resistance to the variation in soil moisture level is given in Figure 5. The resistance values indicated a slight decline as the level of soil moisture increased. The electrical resistance corresponding to the critical level was found to be about




Table 3. Measured irrigation depth (I), evapotranspiration (ET), leaching water (LW), water use efficiency (WUE) and irrigation water use efficiency (IWUE). Treatments I1.0 I.75 I.50 I.25 IControl

I (mm) 832.7 664.3 412.1 287.5 428.1

ET LW (mm) (mm) 483.3 349.4 469.2 195.1 383.3 28.8 287.5 not 428.1 not

WUE (kg m -3) 84.94 63.33 28.00 22.49 50.04

IWUE (kg m -3) 49.30 44.73 26.04 22.49 50.44

not – there was no leaching water

Fig. 5. Electrical resistance differential depending on soil moisture level.

Table 4. Plant yield and yield components of pepper under different water applications. Parameters

I1.0 1289 ± 205 6.54 ± 1.35

Irrigation Treatments I.75 I.50 I.25 IControl 933 ± 337 ± 203 ± 678 ± 97 75 55 110 7.47 ± 5.36 ± 4.35 ± 4.25 ± 1.08 0.5 1.0 0.8

1.20 ± 0.1

1.30 ± 0.2

1.10 ± 0.12

1.02 ± 0.06

1.06 ± 0.05

12.5 ± 0.4

12.6 ± 1.6

11.0 ± 0.4

9.9 ± 0.2

10.6 ± 0.2

15kΩ, which was set to the pot having the value of 100KΩ in the circuit. Hence, the irrigation event commenced as soon as moisture content decreased 30% of the available water. Therefore, whenever moisture content dropped by 30% in the full treatment (I1.0), it was refilled up to FC by the irrigation controller. The excess water above FC, if available, was drained to the drainage canals and then collected in the storage tanks and measured.

Yield (g plant-1) Mean fruit weight (g) Mean fruit diameter (cm) Mean fruit length (cm)

Irrigation Applications The highest quantities of applied water and ET, 832.7 and 483.3 mm, respectively, were obtained from the I1.0 treatment, and the lowest were 287.5 and 287.5 mm, respectively, in the I.25 treatment (Table 3). Water use efficiency declined as the applied water decreased in the irrigation treatments. In this case, the lowest was obtained from the treatment where the least water was applied (I.25), while the highest came from I1.0. Although it was expected that WUE in the control treatment would be very close to that of the full treatment (I1.0), it placed third after I.75, since there was no leaching water, which might be the deposition of some ions in the root area.

in the I0.25 treatment.

Crop Yield and Yield Parameters The highest yield was obtained from full irrigation since full treatment provided the evapotranspiration demand and leaching water requirements of the plants (Table 4). In the control treatment, the quantity of yield and the value of WUE were low, even less than that of treatment I.75. The reason for the low yield obtained from Icontrol treatment may be attributed to ion deposition in the root area of the plants. No leaching of water occurred in the Icontrol treatment, resulting in 47% reduction in yield, compared with I1.0. Parameters related to fruit qualities such as mean fruit weight, diameter and length were highest in I0.75. Severe water shortage in the deficit treatments (I0.50 and I0.25) reduced both the yield and quality parameters seriously; more severe stress occurred


Five samples were measured for all parameters.

Irrigation Applications of the Automated System The action of root zone depletion and the timing of the irrigation events throughout the calendar days are shown in Figure 6. Water was applied according to the pre-set strategy by the automated system whenever soil moisture was depleted in the root zone up to a critical level. Toward the end of the irrigation process, full irrigation (Fig. 6) and the control treatment (Fig. 7) had similar trends, but on some calendar days, they did not match since irrigation events on the 187th, 199th, 204th, 218th 222nd and 232nd calendar days in I1.0 were generated too late according to the timing criteria (set at resistance value of 15 kΩ corresponding to 30% soil moisture depletion from field capacity). The system underestimated the amount of soil moisture by almost 10%. It is possible that as the representative plant had been maturing, it had started shading the surrounding area of the sensor, which caused the soil moisture around the sensor zone to hold a high moisture level for a longer period compared with the less shaded area. It therefore delayed the trigger time more than expected. Water was generally pumped to the root zone as soil moisture approached nearly 10% below the critical level. However, the fluctuations of the substrate moisture were between field capacity and wilting point for the I1.0 347



Fig. 6. Changes in moisture content of different irrigation treatments implemented by irrigation controller.

Fig. 7. Changes in soil moisture content from the control treatment. treatment, hence, the irrigation controller achieved the moisture content in the container to be held between the two critical points. In deficit treatments, the soil moisture level had to be at a certain level below field capacity throughout the experiment. In the I.75 treatment, however, moisture content in the container on the 187th, 189th, 199th, 204th and 227th calendar days was above field

348

capacity, since the system triggered the relays two or three times more, which pumped more water than was required for the treatments. The controller ran the pumps two or three times more on those days since a slight change in the distance between the dripper serving the representative pot and the soil moisture sensor occurred after weighing the pots, therefore, the automated system pumped water till water came in contact with the sensor. The water content above field capacity, however, was not held against the forces of gravity and was drained out of the root zone, that is, into the drainage channel and was collected in the storage tank (30 L). Also, these actions took place in the I.50 treatment on the 187th, 189th, 199th and 204th calendar days. The excess water increased the drainage water, while dropping irrigation water use efficiency in the I1.0 treatment, followed by I.75, I.50 and I.25, respectively. The applied water in the full treatment was almost twice that of the control treatment so that the peak of WUE was obtained from the full treatment while IWUE was lower than that of the control treatment, since no leaching of water occurred in Icontrol. Hence, ions deposited around the root area may put the plants under stress.




CONCLUSION

REFERENCES CITED

The prototype of the irrigation controller was tested to determine its performance. Especially in the full treatment, once the general strategy was defined by the MCU, it took over and made decisions about when to apply water and how much water to apply. In the automated system, depending on the feedback of the sensor, the irrigation decision was made and actions were carried out throughout the entire experiment. Also, the software used in the experiment was simple and maintained the desired state for the full application. The irrigation controller performance can be improved if adequate attention is given to the following points: In the experiment, only one sensor was used to sense the moisture content. However, it should be noted that it is better to use at least three soil moisture sensors since it will give a clearer result. If three sensors are used, signals from the sensors need to be compared, then the decision can be made by the MCU. Another important point is the installation of the soil moisture sensor in the pot, since if placed in a shady area, the circuit may delay the trigger time. The most critical point ensuring the continuation of irrigation is to provide a contact between water and the soil moisture sensor, which will reduce the value of resistance between the electrodes. Therefore, the distance between the dripper and the soil moisture sensor constitutes a critical point in the experiment. The accuracy of the calibration curve between the moisture content and the resistance is another important point. The automated system gives good performance as the software defined to the microcontroller was adjusted according to the full irrigation time.

ABRAHAM N, HEMA PS, SARITHA EK, SUBRAMANNIAN S. 2000. Irrigation automation based on soil electrical conductivity and leaf temperature. Agric Water Manage 45: 145-147.

ACKNOWLEDGMENT The authors would like to thank the Scientific and Technical Research Council of Turkey, Research Project Reference No. 108O276, for financing the study and the Canakkale Onsekiz Mart Agricultural Experiment Station for assistance in the conduct of the research.

ALTINBASAK O. 2004. PicBasic Pro ile PIC Programlama. Atlas Yayincilik. AYERS RS, WESTCOT DW. 1994. Water Quality for Agriculture. Rome: Food and Agriculture Organization. 174 p. BIERNBAUM JA, VERSLUYS NB. 1998. Water management. Hort Technol 8(4): 504-509. CACERES R, CASADESUS J, MARFA O. 2007. Adaptation of an automatic irrigation-control tray system for outdoor nurseries. Biosys Eng 96(3): 419-425. CASTANON G. 1992. The automation irrigation. Maquinas Tractors Agricolas 3 (2): 45-49. CUMING KD. 1990. Irrigation control system. United States Patent US4, 943,400. 5 p. FRANKOVITCH DJ, SARICH JI. 1991. Automation plant watering system. Canadian Patent application. 16 p. GALLARDO M. 2005. Metodos de control de riegoen cultivos en sustrato (Irrigation methods in soil-less cultures). Horticultura 183: 26-31. MORENO JF. 2003. El riego a la demanda en cultivos hidroponicos (Irrigation taking into consideration water needs in hydroponics). Horticultura 168: 42-45. MULAS P. 1986. Developments in the automation of irrigation. Colture Protelle 15(6): 17-19. SALAS MC, URRESTARAZU M. 2001. Objectivos, controles y metodos de fertirrigation en cultivos sin suelo (Purposes, controls and fertigation methods in soil-less cultures). Horticultura 157: 38-48. SHULL H, DYLLA AS. 1980. Irrigation automation with a soil moisture sensing system. Trans ASAE 23: 649-652. WEINSTEIN J, AVISHAI S. 1995. Computerized water control systems. Int Water Irrig Rev 15 (1) January 1995. p. 39-42. YILDIRIM M. 2010. The performance of irrigation automation system based on soil moisture level for pepper (Capsicum annuum L.) growth. J Food Agric Environ 8 (3&4): 629634. YURTSEVEN E, KESMEZ DG, UNLUKARA A. 2005. The effects of water salinity and potassium levels on yield, fruit quality and water consumption of a native central Anatolian tomato species (Lycopersicon esculentum). Agric Water Manage 78 (issues 1-2). p. 128-135.


349