International Commission of Agricultural Engineering

Commission Internationale du Génie RUI'al

13th INTERNATIONAL CONGRESSON AGRICULTURAL ENGINEERING

Rabat (Morocco)2-6 February 1998

13ème CONGRES INTERNATIONALDU GENIE RURAL

Rabat (Maroc)2-6 Février 1998

Proceedings / Actes

VOLUME 3

AGRICUL TURAL MECHANIZA TION

MECANISATION AGRICOLE

CIGR accepts no responsibility for the statements made. opinions exprcssed and the maps included inthese proceedings.

La CIGR se dégage de toute responsabilité pour les déclarations faitcs, les opinions fonnulées et lescartes reproduitcs dans ces actes.

Edited by:

Professor EI Houssine BART ALI. Agricultural Engineering Department of theHassan II Institute of Agronomy and \ eterinary Medicine of Rabat - Morocco

Mohammed DAOUDL Agricultural Engineer. Consultant, Rabat - Marocco

Published by ANAFID 2, Rue Haroun Errachid. Rabat - Marocco

For the complete set of six volumes ISBN <)981-18~n-()-()For volumc l: ISBN 9<)81-1887-1-9For volume 2: ISBN 9981-1887-2-7Far volume 3: ISBN <)<)81-1887-3-5Far volume-l: ISBN <)981-1887-4-3For volume 5: ISBN <)981-1887-5-1For volume 6: ISBN <)<)81-1887-O-X

Printed in Morocco / Imprime au MarceDépòt légal : 90/ l<)<)8

CONTENTS / SOMMAlRE

SECTION 3 : AGRlCULTURAL MECHANIZA TION

MECANISA TION AGRICOLE

3.1. National trategies For Agricultural Mechanization

Stratégi _-ationafes de Mécanisation

Agriculrural mechanization strategy formulation : Concepts and rnethodologv and therol of the private sector and the governmentL J Clarke (FA.O)

lmernational standards for agricultural equipment 17H. Ru eli, Ha/m. & Barrie L. Smith (USA)

Evaluation methods to assess the benefits of precision agriculture techniques in the 39Italian situation.\1. Lazzari, F Mazzetto & M. Vaccaroni (Italv)

111eresearch and inquiry of mechanized technology for dry agriculture 51Z. Shukuan, Z. Jingzhe, L. Jushuang & L. Shenggong (China)

3.2. Mechanization Techniques Adapted to Mediterranean and Tropical Rcgions 57

Tcchniqucs de Mécanisation Adaptées aux Régions Méditerranécnnes et Tropicales

Studies on use of improved single and double animal harnesses and Yokes 59M. P. Singh & Jayant Singh (India)

Le travail du sol dans les oliveraies en milieu aride 67B. Ben Rouina, M. Yousfi, _M. Mlaouah & A. Omri (Tunisia)

Design and trial of a chisel plough fitted with parabolic shanks to work as rigid or 77flexible shanksF Plegrin , A. Madueno.M. Lopez, D.L. l/a/era, 1. Aguera, .J.R. Jiménez & GL Bianco(.pain)

Tillag and erosion 85E Cerruto , S. Failla & G. Schillaci (Italv)

Soil compa tion management by vari-width subsoiler 931.1. JOr/ & . Salamon (Hungary)

Axle load and ground pressure responsibility 011the compaction of tilled soils. Part I : 101Root bedD. .Iorajurta. .F Botta & t.s: Draghi (Argentina)

Axle load and ground pressurc rcsponsibility 011 thc ompaction or tilled soils. Pan Il : IO]Secdbedc.F Bouo. D. Jorajuria, & LII Droghi f.·lrgenllno,

Comparison or zero. minimum and convcutional ullag ~ tCI11S011 whcat vicld inpaddy-wheat rotauonTV .)/ngl1. & B Singh (Indio)

11~

Seedbed characteristics effects on whcat emcrgcncc in convemional aud conscrvation 121tillage systems.·r flemma! .·1 (Iran)

Des outils intelligents pour les lits de semenccs I:nR. Rouveure & .f-F Billot (France)

Contròle de l'uniformité de I'enrouleur l-n.'11.. ·lzlIuggogh (.Iforocc:o)

Design or a dccision-support systcm for thc icchnical-cconomical analysis or I.j.l)agricult ural contractors.1/. l 'accaroni, F. Mazzetto. D. Piro la & (;. Gastelli (Italv)

Artificial neural network for Ù1C indircct cvaluation or tractor cugine torquc 1St),)' Sorcinelli & G. Peri (Italv)

A studv or sugar bcet root strength I()7L·1. Gemtos (Greecei

Fabrication locale de matériel agricole au Marce 175j':.H Boati, A. Bahri (Marocco) & P. Schulze Lanuners (Gerntanv)

Preliminary trial of selected agriculturaJ machinery Ior small farms or Morocco IXIT Takesono, l. Jenane & A. Ezzahouani f.l!oroCCO)

Straw-grain separation with screw conveyor in conveniional combincs IX7A. Ince & E Guezel (Turkev)

Ccrcal harvesting in developing countries: Evaluation or different t hnological 105solutions.'1/ Cicoria, Ci. Ottaviani, M. Vieri & .11. /.oli (ltalv)

Design. Dcvelopment and testing a date harvcsting machine 205.1I. Shanisi, .J. Kilgour, R.J. Godwin & S Blackutore (i 'm/ed .ingdom)

Development of a new combinc harverster for multi-crop in Japan 2 I IT Sugtyama, T Ichikawa. Y. Hidaka. ,,1. Yamamoto, A.'. Hamada, T Odaliara &M, Mizumoto (Japan)

Perforrnances d' une arracheuse de pommes de terre 2 19EH Baali (Morocco) & HJ. Heege (Gennanvt

-'II

3.3. Equiprnent and Technical Itincrarics Rcspcctuous of thc Envirunmcnt ll5

Matériel ct Itinéraires Techniqucs Rcspectueux de l'Environncment

Les effets de trois itinéraires techniques sur le comportement du tracteur/outil et sur 227létat structural du solT. Xlansouri, .ILI. Ben Abdellah & .II.E Hamza (Tunisie)

Evolution or dry bulk density in a vertisol after passes with tracted and wheeled vchicles 237f)'!', l'o/era,.I Gil, .I I. Galvez, .'1. Xladueno. .''L Pena & .I.,t Salinas (Spain)

Exploitation or Diesel internaI combustion engines in ihe buildings with lirnited 2.J.5airexchange.LX Kartashevich. 1>1.Belousov & .Lt Sushnev (Belarus)

Design and testing of a zero emission tractor 253L Bodria, R. Guidetti & C. Bisaglia (Italv)

Work quality or a fcrtilizer spreader equipped with a DPAE control system 2(l'\.11.Fio/a & R. Oberti (lta/vi

Development of a robotic sprinkler hcad [or precision irrigation 275C. Turker (Turkey), B.S: Blackmore & F. t; 1I'eatherhead (Unhed Kingclotu)

Spray volume rate errors in intcnnittent opcration of hydraulic nozzles 283st. Sa/vani ([j.~>lJ

Implementation of a spray distribution modcl for the evaluation or field sprayer boom 29 IbehaviorL l,ardo ux, C Sinfort, A. Miralles, H Bonicelli & F Sévila (Francei

Autornatisation d'un banc de répartition destiné éÌ tcster les buses de pulvérisation 305E. Hamza (Tunisia), F Lebeau & .1 l-I" Destain (Belgium)

praycrs used in greenhouses: Parameters or distribution J 13E Cerruto, R. D 'Antico, S Failla & G Maneuo (lta/vi

'ind tunnel simulation of chemical antidrift adjuvants for pesticide rrcauncnts 325C. De Zanche, D. Friso, C Baldoin ce- .1. Zelante (ltalv)

or conservation tillage on watershed hvdrology in semi arid Kcnya: An 335tion of AGNPS. SCS-CN and rational formula runoff models

_ - Btantah, (Kenva), L Stroosnijder {i\'elher/andsi, Te. Sharnia & R.K.J.:. Cherogonv'61.'0)

III

Work Quality or'a: Fertilizer Spreader Equipped witha DPAE Control System

Marco Fiala and Roberto Oberti •

ABSTRACT

The field performance of a spreader depends on both transversal and longitudinal uniformityof the fertilizers distribution. On the basis of the experimental tests on a DPAE control system(distribution proportional to work speed), carried out with 3 different fertilizers, at 3 application ratesand at 8 speeds (6 to 13 kmIh), a model to calculate the total Coefficient ofVariation ofthe spreaderwas defined, determining a global index of its work quality. Moreover, knowing the fertilizerresponses to the different arabi e crops and quantifying the tota! distribution error, it has beenobtained a preliminary evaluation of the economie benefit related to the use of a spreader equippedbya DPAE system.Keywords: fertilization, rate control, distribution uniformity

INTRODUCTION

Within the more advanced agricultural systems, Precision Farming (PF) is the topic on whichmuch of the current scientific research in Agricultural Engineering has been focused (Stafford ed.,1997). This is due to the development of technologies able of localising machines in tbe field,controlling their operations and processing information, thereby enabling site-specific contro l of fieldoperations.

However, the widespread adoption of PF is subordinated to several technical and economiefactors; this points out the insufficient maturity ofPF and hinders its rapid commerci al diffusion.

There are also particular situations - such as those existing in lta!ian agriculture - whosestructural characteristics (fragmentation of plots, shape and layout of fields, disformity of crops, farmorganisation, etc.) consti tute further causes against the adoption of this new system of "agriculturalmanagement" (Mazzetto et al., 1997).

Consequently, it is stili a general interest to analyse the performance offarm equipments usedin the traditional agriculture, for which the smallest unit of reference is the whole plot. In thisconnection, one of the most interesting aspects is the evaluation of the work quality improvementswhich can be obtained through the adoption of innovations (sensors, control systems, etc.), notnecessarily integrated into a PF system.

This paper analyses the main causes of errors in the application rate of fertilizers, and theimpact of such errors on the economie performance of a particular crop. However, this approachdisregards the systematical andlor instrumental errors connected to the measurement of the variables.

FERTILIZER CROP RESPONSE, FERTILIZER RATE AND CROP PROFIT ABILITY

For certain types of crops (cereals, roots and tubers), several studies (KIing, 1986; England,1986; Shiles, 1989; Morrison et al., 1980; Sequi ed., 1997) correlate - for nitrogen only - the cropyield Y (t/ha) with the fertilizer rate D (kg/ha) (Fig. 1).

Authors are Agronomist and Physicist, respectively. University ofMilan, Institute ofAgricultural Engineering, ltaly

263

12.0 'O.

il-,I

corn19.0 t f · , · ·..··; · ·,

jI

ro 1~ 6.0 t

corn2

--~------+------~300100 200

Applicatian rate (N, kg/ha)

Fig. l - Yield response vs. nitrogen fertilization

The various Y(O) functions, notwithstanding common characteristics (concavity towards thebottom, a single maxirnum YM at a rate OM, a value Y(O) > O), exhibit marked numerical variabilityeven within the sarne crop; this clearly limits the applicability of these functions to the site andconditions ofthe experimental measurements.

Nevertheless it is possible to make a few general considerations. In fact, isolating thecontribution of fertilizer rate from the other production factors, the protit per unit of surface (speciticprotit Ps; ECU/ha) obtainable for a specific crop is given by:

Ps(O)=Y(O)·pu -O-cu [ECU/ha] (I)

where Pu (ECU/t) is the sale price ofthe product and Cu (ECU/t) is the cost of each unit offertilizer.dP

The optimal fertilizer rate Do (kglha) must verify the condition ~ = O; differentiating

(Eq.l) the opti mal fertilizer rate is obtained when:

, CuY(O) =--Pu

(2)

Do is therefore the fertilizer rate for .which the derivative of the yield response function isequal to the ratio between fertilizer cost and product price (Fig. 2).

264

Y'(D} <

C.

P,~I--__

Ps (D) i

~

Y'(D}

c.P,+---+--="""-'~~

P,; (D)

DL----:!Q,:----->D

Fig. 2 - Qualitative yield response functions to nitrogen fertilization for cereals (A) and roots/tubers(B); identification of the optimal fertilizer rate Do and the maximurn yield fertilizer rate DM;effect ofthe applied fertilizer rate on the specific profit P, obtainable from the crops.

CAUSES OF ERROR IN THE FERTILIZER RATE

Assurning that the fertilization requirements of a plot are homogeneous - i.e. that the fertilizerapplication rate DI (kg/ha) must be the same over the whole plot area A (ha) - by operating with acentrifugai spreader with applicationldistribution components properly functioning (constantfertilizer flow rate Q, kg/min; constant spinner rotation speed co, rmp), the possible causes which canlead to an error in the fertilizer application rate DI are:

i. irregular spreader pattern;ii. incorrect forward travel-line (trajectory) along the field;iii. variations in the work speed.

For evaluating the cost/benefit ratio of a technological innovation aiming to minimise sucherrors is essential to quantify their effect on the economie performance of'the crop.

NON-UNIFORM FERTlLIZER DISTRIBUTION AND CROP PROFITABILITY

Subdividing the plot surface A (ha) into elementary areas - whose dimensions (éx; òy) arechosen so that two adjoining areas exhibit a non-uniform fertilizer distribution - the specific profit ofthe plot A - defined by (Eq.l ) - can be expressed as the mean of the individuai specific profits of eachelementary area:

(4)[ECU/ha]

265

In (Eq.4), the distribution of the fertilizer rate arount its mean value D has no effect on the

terrn (D. cu), although - being Y(D) *'Y(D) - it does influence the terrn (Y(D). Pu).

If f(Dj) is the statistical distribution of the fertilizer rates in the individuaI elementary areas,(Eq.4) becomes:

"cl

Ps(D)=Y(D)·pu -D.cu =Pu ·L)(DJY(Dj}-D.cui= l

[ECU/ha] (5)

where: Ilcl is the number of c1asses into which the fertilizer rate range is subdivided.Equation 5 expresses the concept, observed also in practice, that the specific crop profit Ps is

influenced by the yield and, therefore, by the uniforrnity with which the operator distributes thechosen fertilizer rate.

It is therefore necessary to analyse the behaviour of f(Dj) in relation to the causes of errorpreviously identified.

Irregular Fertilizer Spreader PatternThe application of a specific fertilizer with a specific spreader has an associated "unitary

distribution pattem".The shape and uniforrnity of this pattem depend on the adjustments (ground height, speed of

rotation, fin length and angle, etc.) made by the operator on the spinners in accordance with theManufacturer's recommendations.

The pattem is deterrnined experimentally by means of standardised tests (ASAE S341.2; ISO5690/1) conducted at various testing centers (CEMAGREF, NLH, Statens Maskin-Provningar etc. ).

The overall uniforrnity ofthe lateral distribution - obtained from the partial overlap of severaladjoining passes ("multiple profile") - is expressed by a Coefficient ofVariation, given by:

(JqCV =~·100

p q (6)

where: (Jq (kg/min) is the standard deviation ofthe fertilizer flow rate ofthe collected samples and q(kg/min) is the mean ofthese values.

The reference values commonly used for CV p (Balsari et al., 1994; Yule et al., 1996) areshown in Table 1.

Table 1 - Lateral distribution Coefficient ofVariation

UNIFORMITY of COEFFICIENT ofLATERAL VARlATION

DISTRlBUTION(%)

Excellent O~CVo~ 5Good 5 <evo s lOFair 10<CVo~ 15

Inadequate 15 <CVn ~ 20. Poor CVo> 20

For the purposes of this study, the unitary distribution pattem - deterrnined by the operator'sadjustments and regulations - is assumed to be constant for ali the passes which make up thecomplete field fertilization operation.

266

lncorrect Forward Travel-line (Trajectory) Along the PlotThere exists a useful distance bu (m) between adjoining passes which overlaps two unitary

distribution patterns in such a way as to minimise the Coefficient ofVariation CVp• This distance b,is determined on the basis ofthe distribution pattern, usually through iterative numerical calculations(Balsari et al., 1994; NLH, 1996).

In this way an optimal travel-line, usually rectilinear, is deterrnined; the spreader must followit in order to obtain ali the operation passes equidistantly spaced by bu.

In practice, however, the operator is unlikely to maintain thé optimal travel-line for the entirefertilization operation; therefore, this study assumes that the actual field position of the fertilizerspreader exhibits a normal distribution around the optimal distance b.,

The probability that the spreader is located, at any given time, at a distance b (m) from theoptimal trajectory is given by:

b2

f(b)= ,I .e-2"~,,27t ·crb

(7)

The travel-line error b has a negative effect on the uniforrnity of distribution, generating anon-optimal multiple distribution profile.

The magnitude of the variance (o ,2) depends on the possibility of identifying and maintainingthe optimal travel-line in the plot (driving ability, field conditions, etc.).

Variations in the Work Speed of the FertiIizer SpreaderMany factors can cause variations in the work speed of a spreader, such as: soil slope and

irregularities, shape and size of the plot, slip of the tractor driving wheels, reversing the direction oftravel and the associated manoeuvres, ability of the operator.

Because the experimental "data available only quantify work speed variations for sirnilaroperations (Grisso et al., 1988), we assume that the speed of the spreader follows a nonnaldistribution:

(8)

The value ofthe variance (cr/) depends, obviously, on the magnitude ofthe causes describedabove under the various operating conditions.

Total Distribution ErrorDuring a fertilizer-spreading operation, the causes of error separately analysed above can

coexist, contributing in varying proportion to the non-uniforrn distribution ofthe fertilizer.The actual distribution over the entire plot is in fact determined by summing up the

longitudinal variability effects (y axis, travel direction) - caused by variations in the work speed ofthe machine - and the lateral variability effects (axis x), caused by irregularities in the spreaderpattem and its incorrect overlap due to travel-line deviations.

The plot A can be subdivided into A, elementary areas of width Sx. (m) and length tJ.y (m).Inside each A, = tJ.X"tJ.y,values for both the distribution pattem and the work speed are assumed to beuniforrn.

In practice, the transverse dimension /1x coincides with the resolution used to subdivide theunitary distribution pattem (éx = 0.5 - 1.0 m), whereas the dimension along the travel direction tJ.y,which takes into account the speeds variations, falls in the range tJ.y= 1.0 - 2.0 m.

In the case of a rectangular field, the total number of elementary areas nAwill be:

267

n =(bLA) .(bLU)A t.x t.y (9)

where bLA'(m) and bLU(m) are, respectively, the total plot width and length.By sìmulating the movementof a fertilizer spreader on the field, it is possible to evaluate the -

fertilizer rate Dk (kg/ha) distributed on each elernentary area Ak.To this end, along a single pass, for each forward step t.y ofthe unitary distribution pattem:

L the forward speed varies, using random values calculated by (Eq.8).2. the spreader pattem laterally slùfts from the optimal travel-line, using random values calculated by

(Eq.7).

The procedure is repeated for each pass of the fertilizer spreader, until the whole plot widthbLA(m) has been covered.

DETERMINING THE FUNCTION f(Di)

The fertilizer rates Dk (kg/ha) in each ofthe elementary areas Ak (ha) are given by:

"d

Lmk,j l"dje l "qk,j

Dk =---=-'L..--t.x. t.y t.x je l V k,j

[kg/ha] (lO)

where: na is the number of times (usually 1 or 2) that the unitary spreader pattem "passes over" thatAk; vkj (km/h), mkj (kg) ànd qkj (kg/min) are, respectively, the different work speeds, the rnasses andthe fertilizer flow rates in the Ild distributions which pass over elementary area Ak.

Because the travel-line errors of the fertilizer spreader cause a lateral shift in the spreaderpattern, it is necessary to redefine - for each trajectory variation .: the fertilizer rates distributed on theunderlying elementary areas. For this purpose, by plotting the unitary pattem as a broken-linediagram and 'averaging the values of the broken-line portion that fall upon ~, it is possible todetermine the flow rate qkj (Fig. 3).

o ..,I Al I A

2I A3 I A. I

b

o .1\7Al I A

2I A3 I A. I

Fig. 3 - Effect of unitary profile slùft (b) on the amount of fertilizer distributed on the variouselementary areas AI, .... , ~, ....

The used algorithm can be summed up in the following steps:

268

(bLA)l. for each pass from l to h:- ;

(bLU)2. for eachforward step from l to 60y ;

3. calculate v (km/h) work speed (Eq.9);4. ca1culate b (m) deviationfrom optimal travel-line (Eq.8);5. translate the unit pattern into x = (pass . b, + b); .6. ca1culate qk (kg/min) fertilizer flow rate for each elementary area Ak which falls

under the unit spreader pattern;7. ca1culate Dk (kg/ha)ftrtilizer rate for each Ak ofpoint 6) (Eq.lO);8. add Dk to the fertilizer rates previously distributed on Ak;

9. next forward step;lO. nextpass.

This algorithm, implemented using a common programming language, permits the simulationofvarious operating conditions. These are defined through parameters related to the spreader (unitaryspreader pattern, work speed, nominal fertilizer rate) and to work conditions (speed variations givenby O"v;travel-line deviations given by O"b).

The output of the program is the frequency of the fertilizer rate distributed on the nAelementary areas into which the plot was subdivided. This defines the function f(Di), necessary forca1culating the profit in (Eq.5).

APPLICATION OF THE CALCULATION MODEL

CentrifugaI Fertilizer Spreader Equipped with an Electronic Controller for DistributionProportional to Work Speed (DPAE)

Adopting a DPAE controller on a centrifugal spreader should eliminate non-uniformity in thefertilizer distribution, because the variations in the work speed of the machine are balanced by aproportional variation in the flow rate of fertilizer to the spinners.

This is equivalent to consider the effect of speed variations on distribution equals to zero, i.e.to set o"v= O in (Eq.9).

However, experirnental tests conducted on a fertilizer spreader equipped with a DPAE (Fialaet al., 1997) have shown that the DPAE does not totally eliminate the fertilizer rate errors due tovariations in the work speed. The data coIlected during these tests indicate a Coefficient ofVariationset in the range 2.5 % s CVv s 6.0%.

For the application of th~ simulation model it is possible to assume, for this type ofmachine, an average work speed v = lO km/h and a CVv = 4.0% (c, = 0.4 km/h).

For the other causes of error, two different operating conditions are simulated (Table 2):

• Case A: excellent IateraI distribution quality (CVp = 5.0%); excellent spreader traveI-IineO"b

(m = 0.5 m; CVo. = - = 2.1 %);bu

• Case B: poor Iateral distribution quality (CVp = 17.0%); large deviations ofthe spreader fromthe optimal travel-line (m = 1.5 m; CVo' = 6.3%).

A pIot of width bu = 250 m and length bLU = 250 m, subdivided into llA = 62500elementary areas (each measuring 60X = 0.5 m by 60y = 2 m), fertilized with nitrogen at thenominaI application rate D, = 220 kg/ha by a centrifugal spreader with a working width of b, =24 m is considered,

269,.

Table 2 - Application ofthe model to a spreader equipped with a DPAE in two different operatingconditions: Coefficients of Variation

OPERATING CVv CVp cv,CONDITIONS

Optimal (Case A) 4.0% (o- = 0.4 Ian/h) 5.0% 2.1% (CJb = 0.5 m)Poor (Case B) 4.0% (o- = 0.4 Ian/h) 17.0% 6.3% (m = 1.5 m)

The results of the simulations are reported in Figures 4 and 5; they show how uniformityof distribution is affected by incorrect adjustment of the spreader spinners and by deviations fromthe optimal travel-line. The broader the curve obtained for f(Di), the stronger the negativecorrelation.

~ 10.0r~~ 8.0",>-o

lc:Q)

s.o !:::>CTQ)~ 4.0Q)oc:~ 2.0:::>ooO 0.0

o 20 .60 80 100 120 140Applied rate (% of Dt)

Fig. 4 - Applied rate occurrence frequency: Case A

40

~

10.0 j8.0 +o

I>- 1oc: 6.0Q):::>CT

~ 4.0Q)oc: 2.0~:::>oo 0.0O

o 20 60 80 100 120 1~Applied rate (% of Dt)

Fig. 5 - AppIied rate occurrence frequency: Case B

40

.cv» = 4.0%

cv p= 5.0%

cv ot = 2.1%

160 180 200

evv = 4.0%

ev p = 17.0%

:CVol = 6.3%

160 180 200

The Coefficient of Variation of the applied rate can provide an index of the totaI workquaIity of a centrifugai spreader:

aDCVT= ~ ·100D

(11)

Moreover, the f(Di) obtained from the simuIation permits to calcuIate by (Eq.5) the specificbenefit obtainabIe from a particuIar crop.

270

As an example, a field cultivated with com, whose yield response to nitrogen fertilizationY(D) (t/ha) lies midway between (see Fig. 1):

Yl(D) = 4.42 + 3.32·1O·2·D- 8.84·1O-s·D2+ 6.60 ·1O-8·D3

Y2(D) = 3.94 + 5.56·1O-2·D-3.91·1O-s·D2- 2.22 ·1O-7·D3

is considered. ,-Assuming a com sale price pu = 105 ECU/t (excluding tax), a specific cost of nitrogen c,

= 450 ECU/t (excluding tax) and using into (Eq.5) the f(Di) resulting from the Cases A and Bsimulations, the specific benefits detailed in Table 3 were obtained.

Table 3 - Application of the model to a spreader equipped with a DPAE in two different operatingconditions: technical and economie results

OPERATING CVT ECONOMIC YIELD RESPONSE FUNCTION ECONOMICCONDITIONS PARAMETERS BENEFIT

(%) (ECU/t) (t/ha) (ECU/ha)Optimal (Case A) 9.6 p, = 105 Yl(O) = 4.42 + 3.32·1Q"2·0 - 760+1160

+ 8.84.10.5.02 +6.60.10.8.03

Poor (Case B) 21.6 Cu = 450 Y2(D) = 3.94 + 5.56·1O·2·D- 740+1120+ 3.9HOS·02 - 2.22·1Q"7·D3

Conventional Centrifugai Fertilizer SpreaderFor spreaders which are not equipped with a speed controller, it is necessary to evaluate the

magnitude of CVv.In the lack of experimental data, an average work speed of v = lO km/h and an operating

situation for which the work speed falls in the range 7 ~ v ~ 13 km/h in 95% of the cases (CVv =15.0%; crv= 1.5 km/h), were simulated.

In order to compare the results, for both other error causes (Table 4) and operatingvariables, the same hypotheses of Cases A and B were assumed.

Table 4 - Application of the model to a conventional spreader in two different operatingconditions: Coefficients of Variation

OPERATING CVv, CVp cv,CONDITIONS

Optimal (Case C) 15.0% (c- = 1.5 /cm/h) 5.0% 2.1% (CJb = 0.5 m)Poor (Case D) 15.0% (c- = 1.5 km/h) 17.0% 6.3% (CJb = 1.5 m)

The results obtained for Cases C and D are shown in Figures 6 and 7.

271

10.0

~>. 8.0oc::Q):::l 6.0CT

~s 4.0c::~:::l 2.0ooO

0.0O

. , .

v=

ev p= 5.0%

cv oì = 2.1%

10.0

~ 8.0o

>.oc:: 6.0Q):::lCT

~ 4.0

sc:: 2.0~:::loo· 0.0O

O

20 60 80 100 120 140Applied rate (% of Dt)

Fig. 6 - Applied rate occurrence frequency: Case C

40 160 180 200

evv = 15.0%

ev p= 17.0%

.Cv ol = 6.3%

60 80 100 120 140Applied rate (% of Dt)

Fig. 7 - Applied rate occurrence frequency: Case D

20 40 160 180 200

Considering, for the Cases C and D, the same crop and the same economie parameters, thespecific benefits obtained are shown in Table 5.

Table 5 - Application of the model to a conventionaI spreader in two different operatingconditions: technical and economie results

OPERATING CVT ECONOMIC YIELD RESPONSE FUNCTION ECONOMICCONDITIONS PARAMETERS BENEFIT

(%) (ECU/t) (kg/ha) (ECU/ha)Optirnal (Case C) 18.1 pu = 105 Y,(D) = 4.42 + 3.32·1o-2·D + 745.,.1120

- 8.84·1O·s·D2 +6.60 ·10"·0'Poor (Case D) 28.9 Cu = 450 Y2(D) = 3.94 + 5.56·1O·2·D + 725.,.1070

- 3.91·10'·D2 - 2.22·1o-7·D'

CONCLUSIONS

A simulation model was developed to determine the non-uniformity of distribution of acentrifugaI fertilizer spreader operating on a generic plot. The non-uniformity was expressed as anapplied rate occurrence frequency function f(Di) and as a Total Coefficient ofVariation (CVT) whichprovides an total index ofwork quality.

272

The model takes into consideration errors in applied fertilizer rate caused by: a) irregularitiesin the distribution pattem; b) incorrect forward travel-line along the plot; c) variations in the workspeed.

The algorithm was used to simulate the behaviour of a centrifugai spreader equipped witbDPAE controller under two different operating situations (Case A: optimal conditions; Case B: poorconditions), obtaining CVTs of9.6% and 21.6%, respectively.

Comparison of these indices stresses out the high incidence on work quality of botb thedistribution pattem (spinners regulation) and the spreader travel-line (passes overlapping) and,consequently, the importance to put under constant control tbese work parameters.

For each simulated case and on tbe basis of two different yield response functions tonitrogen fertilization, the function f(Di) has been used to calculate tbe economie benefit related to acom cultivation. Under tbe simulation assumptions, tbe Case B-conditions led, in comparison witbthe Case A-conditions, to a reduction of the economic;,benefitfrom 20 to 40 ECU/ha.

It is important to stress that, unknowing the funetion f(Di), that means to neglect the non-uniforrnity of fertilizer distribution, the economie benefit of the crop would be wrongly calculated onthe basis oftbe average fertilizer rate (according to Eq.I).

The model was also applied to a eonventional fertilizer spreader, assuming eonditions (CaseC and Case D) comparable to tbose used for simulating tbe DPAE spreader. The correspondingCVTs, 18.1% (Case C) and 28.9% (Case D), were significantly worse than the previous ones.

From the economie standpoint, comparing tbe results witb tbose of the DPAE-equippedspreader, a further reduetion (15+50 ECU/ha) in the erop benefit was aehieved.

The study highlights tbat the work quality of centrifugai spreaders can be substantiallyimproved by adopting systems andlor devices capable of, on the one hand, controlling variations inwork speed (DPA, DPAE) and, on the otber hand, identifying the optimal travel-line witb greaterprecision (row-tracing, drive control).

Finally, it should be emphasised that the experimental determination of the actual work speedduring field operations would perrnit an even more rigorous calculation of the economie benefits,under various working conditions, connected witb adoption of DPAE systems on fertilizer spreaders.

REFERENCES

Balsari P., Airoldi G., (1994). Spandiconeime (Fertilizer spreaders). Edagricole, pp. 82Eogland R.A., (1986). Reducing the nitrogen input on arable farms. Journal of AgriculturalEconomics, 37(1): 13-24Fiala M., Oberti R., (1997). Prove di DPAE per il controllo della distribuzione su unospandiconcime ad azione centrifuga. (Experimental tests of a DPAE control system installed on acentrifugai fertilizer spreader). Proceedings 6th National A.I.I.A. Congress, September 10-12,Ancona, Italy, VoI. 3, 557-565Grisso R.D., Hewett E.J., Dickey E.C., Schoieder R.D., Nelson E.W., (1988). Calibrationaccuracy of pesticide application equipment. Applied Engineering in Agriculture, 4(4):310-315Kling A., (1986). Possibilities and limits of nitrogen fertilization. Economie aspects demonstratedfor different crops. Nitrogen, 14, FrankfurtJMainMazzetto F., Vaccaroni M., Lazzari M., (1997). Agricoltura di precisione: realtà eprospettive.(Precision farming: reality and prospects). Proceeding 6th National A.I.I.A. Congress,Ancona, Italy, September 10-12,VoI. 3, 271-280Morrison J., Jackson M.V., Sparrow P.E., (1980). The response ofperennial ryegrass to fertilizernitrogen in relation to climate and soil. Grassland Research lnstitute, Hurley, Techhical report n.27NLH-Iostitutt for Tekniske Fag, (1996). Melding Prove (Nordsten Logic NN), N.3-1996, NorwaySequi P. (ed.), (1997). Progetto finalizzato Produzione Agricola Nella Difesa dell'Ambiente(PANDA).(project "Agricultural productivity protecting the environment"). Agricoltura e Ricerca,167,5-54; 168,5-100

273

Shiles R., (1989). Nitrogen for winter barley. Agtec., 1988/89, Hydro Fertilizers Ltd., Ipswich,EnglandStafford J. (ed.), (1997). Precision Agriculture '97. Proceedings i" European Conference, VoI. I:Spatial Variability in Soil and Crop, VoI. II: Technology, 7-10 September 1997, Warwick UniversityConference Centre, UK, pp. 991Yule I., Crooks E., 1996. Precision farming: the price of imperfection. Landwards, Spring, 5-9

274