Agricultural Engineering · Post-Harvest, Food & Process Engineering · also Agricultural Process Engineering, Crop Processing
Moisture, drying, parboiling, milling, cleaning, size reduction and conveying of grains, pulses and oilseeds.
In the exams: GATE AG Section 6 (~14 marks a year 2019–26; psychrometry in all 8 papers, rice milling in 6 of 8); ICAR AIEEA-PG Unit VI; ICAR AICE subject 41; NCEES PE processing; PRC ABE bioprocess; ASRB NET ASPE
Post-harvest engineering of cereals, pulses and oilseeds is the engineering of everything that happens to a grain between the combine or threshing floor and the bag of rice, flour, dhal or oil that leaves a mill. It covers how much water the grain holds and how to remove it safely, how air carries heat and moisture, how dryers are designed and run, how paddy is parboiled and milled, how wheat becomes flour, how pulses are dehulled and split, how oil is pressed or extracted with a solvent, how grain is cleaned and graded, how it is ground, how it is conveyed, and how seed is conditioned so that it still germinates.
The losses are large and mostly avoidable. Grain harvested wet moulds within days; grain dried too hot cracks in the mill; a badly set husker turns whole kernels into brokens; an expeller that leaves oil in the cake gives away the most valuable product. Each is a question of heat and mass transfer, mechanics and material balance, and each can be calculated.
Examiners test the same core everywhere. GATE Agricultural Engineering asks psychrometry every year and rice milling in most years, with moisture conversion, drying time, separation effectiveness and size-reduction laws close behind; ICAR AIEEA-PG, AICE-JRF/SRF and ASRB NET cover the same ground at increasing depth. The NCEES PE Agricultural and Biological exam puts drying, psychrometrics and bulk solids in Processing Systems, and the Philippine ABE licensure examination covers drying and milling under bioprocess engineering. Crops, dryers, grading standards and definitions such as head rice differ by country, and the manual says so where it matters. Read the chapters in order: every later calculation depends on moisture, sorption and psychrometry.
Water governs a grain's sale weight, storage life, hardness in the mill and susceptibility to moulds, so moisture must be stated unambiguously, and that means stating the basis. Moisture content on a wet basis (wb) is the mass of water divided by the total mass of the moist sample: M_wb = W_w / (W_w + W_d), where W_w is the mass of water and W_d the mass of dry matter. It is used in trade, grading and on moisture meters. Moisture content on a dry basis (db) is the mass of water divided by the mass of dry matter: M_db = W_w / W_d. It is used in drying theory and sorption models because dry matter does not change during drying; it can exceed 100 per cent. The two are related by M_db = M_wb / (1 − M_wb) and M_wb = M_db / (1 + M_db), with both expressed as decimals. A grain at 20 per cent wb therefore holds 0.20/0.80 = 0.25, or 25 per cent db. The commonest examination slip is to subtract wet-basis percentages directly, as if drying from 20 to 14 per cent wb removed 6 per cent of the original mass. It does not, because the total mass (the denominator) shrinks as water leaves. The reliable way to calculate water removed or added is to hold the dry matter constant. If a lot of initial mass W_i is dried from initial wet-basis moisture M_i to final M_f, the dry matter W_i(1 − M_i) equals W_f(1 − M_f), so the final mass is W_f = W_i(1 − M_i)/(1 − M_f) and the water removed is W_i − W_f = W_i(M_i − M_f)/(1 − M_f). For 1,000 kg dried from 20 to 14 per cent wb, the water removed is 1000 × 0.06/0.86 = 69.8 kg, not 60 kg. The same balance run in reverse gives the water to add when conditioning wheat before milling or soaking paddy for parboiling. The loss of mass on drying is called shrink; trading rules in some countries add a handling allowance on top of the pure water shrink, so a commercial shrink factor is not a physical constant.
Direct methods measure the water removed: oven drying (weighing before and after heating at a set temperature and time), distillation (water boiled off in an immiscible oil and collected in a graduated trap), and Karl Fischer titration, which reacts specifically with water. Oven temperature, time and whether the grain is ground are fixed by grain in documents such as ASABE S352 and ISO and national standards; results are comparable only between the same procedures. Indirect methods are faster and are what the trade uses. Electrical resistance meters measure the conductance of a compressed sample; capacitance (dielectric) meters measure the dielectric constant, which rises strongly with water because water has a high permittivity; near-infrared (NIR) instruments measure absorption at water bands. All indirect meters need calibration against a reference oven method for each grain, and they need temperature correction. Resistance meters read the surface more than the core, so grain just out of a hot dryer or just rewetted by rain reads wrongly until moisture has equalised inside the kernel.
Grain is hygroscopic: it exchanges vapour with air until the vapour pressure of its water equals that in the air. The moisture at which exchange stops, for given air temperature and relative humidity, is the equilibrium moisture content (EMC); the humidity of that air is the equilibrium relative humidity (ERH), which as a fraction equals the grain's water activity (a_w). EMC sets the lower limit of drying and predicts whether stored grain will gain or lose water. Moulds respond to a_w, so safe storage moisture differs by crop, climate and storage period; oilseeds are safe at much lower moisture than cereals because oil holds no water. EMC is measured by the static method, in which samples sealed over saturated salt solutions (each holding a fixed humidity) are weighed until constant, which can take weeks, or by the faster dynamic method with moving conditioned air. Isotherms at two or more temperatures give the isosteric heat of sorption through the Clausius–Clapeyron relation (slope of ln a_w against 1/T at constant moisture is −q_st/R). Removing water from dry grain needs more energy per kilogram than evaporating free water, increasingly so at low moisture, so overdrying wastes energy as well as saleable mass.
A sorption isotherm is the curve of EMC against water activity at constant temperature. For cereals, pulses and most foods it is sigmoid, type II in the Brunauer classification. Three regions are usually described. At low water activity water is held as a tightly bound monolayer on polar sites; its enthalpy of sorption is highest, and it is unavailable for reactions. In the middle region further layers form (multilayer sorption) and the curve is nearly straight. At high water activity water condenses in capillaries and pores, the curve rises steeply, and this loosely held water supports microbial growth. Two effects follow that every examiner expects: at constant relative humidity the EMC falls as temperature rises, so warm air dries better and cooling stored grain raises its EMC; and the desorption isotherm (drying) lies above the adsorption isotherm (wetting) at the same water activity. That gap is hysteresis, and it means a grain dried to a given moisture and then exposed to the same air will not regain quite the same moisture.
No single equation fits every grain over the whole range. The early Henderson equation is 1 − RH = exp(−c T M^n), with T absolute temperature and M dry-basis moisture in per cent. ASABE standard D245 adopted modified forms in which temperature enters as (T + C) in degrees Celsius: the modified Henderson, 1 − RH = exp[−K(T + C)M^N], and the modified Chung–Pfost, RH = exp[−A/(T + C) exp(−BM)], together with the modified Halsey and modified Oswin equations. Constants are fitted per grain and valid only over the range fitted. The BET (Brunauer–Emmett–Teller) equation is theoretical and gives the monolayer moisture, but fits only the low-activity region, roughly below a_w 0.45. The GAB (Guggenheim–Anderson–de Boer) equation, M = M_m C K a_w/[(1 − K a_w)(1 − K a_w + C K a_w)], extends BET across most of the range with one extra constant and is widely used for foods. The two-parameter Oswin model is M = A[a_w/(1 − a_w)]^B.
Drying air is a mixture of dry air and water vapour, both close to ideal gases, and by Dalton's law the total pressure P is the sum of their partial pressures p_a and p_v. All psychrometric properties are expressed per kilogram of dry air, because the dry-air mass through a dryer is constant while its water content changes. Humidity ratio (absolute humidity in older texts) W is the mass of water vapour per kilogram of dry air: W = 0.622 p_v/(P − p_v), where 0.622 is the ratio of molar masses of water and dry air (18.015/28.965). Relative humidity φ is the ratio of the actual vapour partial pressure to the saturation vapour pressure at the same temperature, φ = p_v/p_vs. Degree of saturation μ is W/W_s, the humidity ratio divided by that of saturated air at the same temperature; it is close to, but not equal to, φ. Dew-point temperature is the temperature at which air becomes saturated when cooled at constant humidity ratio, so p_v equals the saturation pressure at the dew point.
The wet-bulb temperature is read by a thermometer with a wet wick in moving air, cooled until convective heat gain balances latent heat loss. For air–water mixtures it almost equals the adiabatic saturation temperature, so constant wet-bulb lines are nearly constant-enthalpy lines. Specific enthalpy per kilogram of dry air, with 0 °C as datum, is h = 1.006 t + W(2501 + 1.86 t) kJ/kg dry air, where t is in degrees Celsius, 1.006 kJ/kg K is the specific heat of dry air, 2501 kJ/kg is the latent heat of water at 0 °C and 1.86 kJ/kg K is the specific heat of water vapour. The humid heat, c_s = 1.006 + 1.86 W, is the heat needed to warm one kilogram of dry air plus its vapour by one kelvin. Humid (specific) volume is the volume of the mixture per kilogram of dry air: v = R_a T/(P − p_v) = 0.287 T(1 + 1.608 W)/P, with T in kelvin and P in kilopascals, giving m³/kg dry air. Fans move volume, but mass balances need mass of dry air, so a fan's volumetric flow is divided by the humid volume at the fan's location.
A psychrometric chart plots humidity ratio against dry-bulb temperature at one total pressure, with curves of constant relative humidity and lines of constant wet-bulb temperature, enthalpy and humid volume; any two independent properties fix the state. Charts for 101.325 kPa must be corrected, or equations used, at altitude; ASABE D271 gives the equations. Five processes cover most drying work. Sensible heating moves horizontally to the right: W and dew point stay fixed, temperature rises, relative humidity falls, and the heat required is m_a(h_2 − h_1). Sensible cooling moves left until the dew point is reached, after which water condenses. Adiabatic drying (evaporative cooling) moves up and left along a line of constant wet-bulb temperature: air gives up sensible heat to evaporate water, its temperature falls and its humidity ratio rises, and the water picked up per kilogram of dry air is W_2 − W_1. Heating followed by drying is the normal dryer cycle. Humidification by steam injection moves nearly vertically.
Many dryers save fuel by mixing part of the exhaust with fresh air. When two streams mix adiabatically, mass balances on dry air and on water, and an energy balance, give the mixed state: m_1 + m_2 = m_3, m_1W_1 + m_2W_2 = m_3W_3, and m_1h_1 + m_2h_2 = m_3h_3, where m is the mass flow of dry air. The mixed state lies on the straight line joining the two states on the chart, divided in inverse proportion to the dry-air masses. The weighting must be by mass of dry air, not by volume: using volumetric flow rates directly is a standard error, small when the two streams are at similar temperatures and larger when they differ. A weighted average of temperature is only approximate; enthalpy is exact. If the straight line crosses the saturation curve, the mixture is supersaturated and fog forms, which can wet grain near an inlet.
Moisture (dry basis) against time at constant air conditions is the drying curve; its slope against moisture is the drying-rate curve. For a very wet material three stages appear. In a short initial adjustment period the surface warms or cools to its steady temperature. In the constant-rate period the surface stays covered by a film of free water, evaporation proceeds as from an open water surface, and the rate is set entirely by external conditions: air temperature, humidity, velocity and the exposed area. At the critical moisture content the internal supply of water can no longer keep the surface wet, and the falling-rate period begins, in which the rate is controlled by movement of moisture inside the solid. Some texts split the falling-rate period into a first stage (dry patches spread) and a second (the evaporation plane recedes inside). Drying stops when the material reaches the equilibrium moisture content for the air.
During the constant-rate period the wetted surface sits at the wet-bulb temperature of the air, and heat transfer and mass transfer are coupled: heat arriving by convection is used entirely to evaporate water. The rate per unit area is R_c = h(T_a − T_wb)/λ = k_y(W_s − W_a), where h is the convective heat-transfer coefficient, λ the latent heat at the wet-bulb temperature, k_y the mass-transfer coefficient on a humidity basis, W_s the saturation humidity ratio at the surface and W_a that of the air. The time to dry from initial moisture X_0 to the critical moisture X_c (both dry basis) is t_c = W_s(X_0 − X_c)/(A R_c), where W_s is here the mass of dry solid and A the drying area. Hotter, drier or faster air shortens it. It matters for high-moisture products such as soaked parboiled paddy. In the simplest treatment the rate falls linearly with free moisture, from R_c at X_c to zero at the equilibrium moisture X_e. Integration then gives the falling-rate time t_f = [W_s(X_c − X_e)/(A R_c)] ln[(X_c − X_e)/(X − X_e)], where X is the final moisture. Total drying time is t_c + t_f. The logarithm explains a practical fact: removing the last increments of moisture takes disproportionately long, so drying to below the needed target is expensive. Mechanistically, water moves inside the kernel by liquid diffusion, capillary flow and vapour diffusion, and in the falling-rate period these are usually lumped into an effective diffusivity. Cereal grains harvested at normal moisture are usually below their critical moisture content, so they dry almost entirely in the falling-rate period, which is why grain drying is modelled by diffusion and thin-layer equations rather than by the constant-rate equation. Because the falling-rate period is internally controlled, very hot air mostly overheats the kernel surface. The moisture gradient between the dry outside and wet core sets up differential shrinkage and tensile stresses that produce stress cracks (fissures) in rice and maize; fissured rice kernels break in the mill, lowering head rice yield. Excessive temperature also denatures proteins, damaging wheat gluten and seed germination, and can cause case hardening in some materials. Two practical remedies follow. First, the allowable air or grain temperature is set by end use: seed is dried coolest, milling grain next, and feed grain can tolerate more. Second, drying is often split into passes separated by tempering: the grain is held without airflow so that moisture migrates from core to surface and the gradient relaxes; the next pass then dries faster and with less cracking. Tempering is standard in paddy drying.
A thin layer is a bed of grain thin enough that the temperature and humidity of the air are practically the same everywhere in it, so every kernel sees the same drying conditions. Thin-layer equations are the building blocks of deep-bed simulations, which treat a deep bed as a stack of thin layers, each receiving the air leaving the one below. Results are expressed as the dimensionless moisture ratio MR = (M − M_e)/(M_0 − M_e), where M is the moisture at time t, M_0 the initial and M_e the equilibrium moisture, all on dry basis. MR is 1 at the start and approaches 0. When air humidity fluctuates or M_e is uncertain, some researchers simplify to M/M_0, which is only acceptable when M_e is small compared with M. The Lewis (also called Newton or exponential) model treats drying like Newton's law of cooling: dM/dt = −k(M − M_e), which integrates to MR = exp(−kt). The drying constant k depends on air temperature and, less strongly, on humidity and velocity, usually through an Arrhenius form k = k_0 exp(−E_a/RT). The Page model, MR = exp(−k t^n), adds an empirical exponent n that bends the curve and usually fits grain data better. The Henderson and Pabis model, MR = a exp(−kt), is the first term of the diffusion series, the coefficient a absorbing the shape factor. Two-term, logarithmic and Midilli models add parameters. The half-response time, the time for MR to fall to 0.5, is ln 2/k for the Lewis model and is the time unit used in Hukill's deep-bed analysis. Models are compared by the coefficient of determination R², the root mean square error and the reduced chi-square; a higher R² alone is not proof of a better model when the number of parameters differs, because more parameters always fit better.
The theoretical basis is Fick's second law applied inside the kernel with an effective moisture diffusivity D_eff that lumps all internal transport mechanisms. For a sphere of radius r whose surface reaches equilibrium instantly, with uniform initial moisture and constant D, the solution is MR = (6/π²) Σ (1/n²) exp(−n²π² D t/r²), summed over n = 1, 2, 3 and so on. For an infinite slab of half-thickness L the equivalent is MR = (8/π²) Σ [1/(2n + 1)²] exp[−(2n + 1)²π² D t/(4L²)] for n = 0, 1, 2 and so on. For long drying times only the first term matters, so a plot of ln MR against time becomes a straight line, and D_eff is obtained from its slope: for a sphere the slope is −π²D/r². Effective diffusivities for grains are of the order of 10⁻¹¹ to 10⁻¹⁰ m²/s, and they rise with temperature following an Arrhenius relation whose activation energy is found from the slope of ln D against 1/T. Real kernels are not spheres, shrink and are layered, so the fitted diffusivity is effective and depends on the geometry assumed.
In a deep bed, air entering at the bottom picks up moisture and cools as it rises, so the grain nearest the inlet dries first and a drying zone (front) moves up through the bed. Below it grain nears equilibrium with the inlet air; above it grain is still wet, and near-saturated air can even deposit water on cool upper layers. When drying stops, the bed contains a moisture gradient: overdried at the bottom and underdried at the top. Hukill's analysis expressed this with depth units (the depth of grain that the air can dry to a given extent) and time units (half-response times), giving MR = 2^D/(2^D + 2^Y − 1), where D is depth in depth units and Y time in time units. The Thompson model assumes equilibrium between air and grain in each thin layer and does energy and moisture balances layer by layer; the Michigan State University (Bakker-Arkema) model writes coupled partial differential equations for heat and mass transfer without assuming equilibrium.
Grain dryers are classified by operation (batch or continuous), by the relative directions of grain and air (cross-flow, concurrent-flow, counter-flow or mixed-flow), and by energy source (fossil fuel, biomass such as rice husk, electricity, or solar). Sun drying remains the commonest smallholder method in much of Asia and Africa; it uses no fuel but depends on weather, allows contamination and causes fissuring. Mechanical dryers are judged on capacity, uniformity, grain quality and energy use, measured by specific heat consumption (energy per kilogram of water evaporated, which cannot fall below the latent heat of water at about 2.3–2.5 MJ/kg and in practice is well above it) and thermal efficiency (heat used to evaporate water divided by heat supplied). Indian post-harvest texts also define the heat utilisation factor, HUF = (t_1 − t_2)/(t_1 − t_0), and the coefficient of performance, COP = (t_2 − t_0)/(t_1 − t_0), where t_0 is ambient, t_1 the hot-air and t_2 the exhaust temperature; the two sum to one.
In a bin dryer grain rests on a perforated floor and air from a fan, heated or not, is blown up through it. In full-bin, near-ambient drying the bin is filled at once and air slightly above ambient is blown for days or weeks; airflow per tonne must finish the top layer before it spoils. Layer drying fills the bin in stages, each layer being partly dried before the next is added. Batch-in-bin drying uses higher temperature and shallower depth, then cools and moves the grain. Stirring devices mix the bed to reduce the moisture gradient. Bin design centres on airflow: the static pressure a fan must overcome rises steeply with airflow per unit area and with depth. ASABE D272 gives this resistance for many grains using the Hukill–Ives form ΔP/L = aQ²/ln(1 + bQ), where Q is the superficial air velocity; fines and trash in the grain raise resistance and create uneven airflow, which is one reason to clean before drying. Solar dryers use sunlight directly or through a collector. In a direct (cabinet) dryer the grain sits under a transparent cover and absorbs radiation itself. In an indirect dryer a separate collector heats the air, which then passes through the product in a shaded chamber, protecting colour and nutrients sensitive to light. Mixed-mode dryers combine both. Air may move by natural convection, using the buoyancy of warm air, often with a chimney, or by forced convection using a fan, which may be powered by a photovoltaic panel. They cost no fuel and protect grain from rain and dust, but depend on sunshine, have low throughput per unit area and control temperature poorly; thermal storage or biomass backup extends operation.
The LSU dryer, named after Louisiana State University where it was developed, is a continuous-flow, mixed-flow dryer widely used for paddy in South and South-East Asia. Grain flows down by gravity through a tall rectangular column fitted with rows of inverted-V channels (ducts). Alternate rows are connected to the hot-air plenum and to the exhaust, so air enters from one row of channels, passes through the grain and leaves through the next. Kernels meet air from varying directions and pass through hotter and cooler zones, giving more uniform drying than a cross-flow column. A fluted or rotary discharge roll at the bottom sets the grain flow rate and therefore the residence time. Grain is often recirculated with tempering between passes. Cross-flow columns are simpler but overdry grain nearest the air inlet. A recirculating batch dryer holds one batch that is circulated continuously by a bucket elevator from the bottom of a drying section to the top of a holding (tempering) section. In each cycle the grain spends a short time in the drying zone and a longer time tempering, which suits paddy because it limits fissuring; the operator stops the dryer when a meter shows the target moisture. In a fluidised-bed dryer the air velocity through a shallow bed exceeds the minimum fluidisation velocity, at which the drag on the particles equals their apparent weight and the bed behaves like a boiling liquid. Heat and mass transfer are very high and the bed is well mixed; the cost is high fan power. For grain it is used for rapid, high-temperature, short-time drying of high-moisture material, for example removing surface water from parboiled paddy in a first stage before tempering and final drying, and for puffing and roasting.
Parboiling is a hydrothermal treatment of paddy (rough rice) before milling. It has three steps: soaking the paddy in water until it absorbs enough moisture, heating it (usually with steam) to gelatinise the starch, and drying it back to a safe moisture for storage and milling. Much of the rice in South Asia and parts of West Africa is parboiled; preference varies by region. Starch in the endosperm absorbs water on soaking, and when heated above its gelatinisation temperature the granules swell and lose their crystalline order. On drying, the gelatinised starch sets into a hard, translucent, compact mass that fills internal voids and cements existing cracks, which is the root of most of parboiling's effects. Parboiling changes the grain in ways every examination asks about. Milling breakage falls and head rice yield rises, because gelatinisation heals fissures and chalky areas and makes the kernel hard. The husk is split and loosened, so husking is easier and needs less energy in a sheller. Water-soluble vitamins, notably thiamine, and minerals move from the bran and germ into the endosperm during soaking and are retained after milling, so parboiled milled rice is more nutritious than raw milled rice of the same degree of milling. Heating inactivates enzymes, including lipase, which slows the development of free fatty acids in the bran and improves its stability as a raw material for rice bran oil; the bran from parboiled rice also usually has a higher oil content. Against these, colour darkens, the firmer texture and odour are not wanted in every market, cooking takes longer, and whitening takes more energy.
In traditional methods paddy is soaked in cold or warm water, often overnight or longer, then steamed or boiled in open vessels and sun dried; single and double boiling variants exist. Long cold soaking allows microbial fermentation, giving a characteristic smell. Hot-soaking methods, developed in India and elsewhere, soak paddy in water held near but below the gelatinisation temperature, commonly in the range of about 60 to 70 °C, which cuts soaking to a few hours and avoids fermentation; steaming then completes gelatinisation. Soaked paddy reaches roughly 30 per cent moisture on a wet basis. Pressure parboiling steams paddy under pressure, shortening the process further but giving a darker product. Sand or dry-heat parboiling heats soaked or conditioned paddy in hot sand or on hot surfaces. The degree of parboiling (how completely the starch is gelatinised) is controlled by soaking moisture, steaming time and temperature, and it determines hardness, colour and cooking time. The drying stage matters as much as the soaking: parboiled paddy leaves the steamer very wet, and drying it too fast, particularly below about 16 to 18 per cent moisture, causes cracking; two-stage drying with tempering is normal. The same principle applies beyond rice. Wheat is conditioned (tempered) before milling by adding water and allowing a rest period, which toughens the bran so that it breaks into large flakes, and mellows the endosperm so it reduces cleanly; warm and hot conditioning speed the process. Bulgur is parboiled, dried, cracked wheat. Pulses are conditioned before dehulling by wetting and drying cycles or by oil-and-water treatments that loosen the husk (see Chapter 9). Maize is steeped before wet milling, and oats and rice bran are steamed or kilned to inactivate lipase. The engineering is the same: a water balance for soaking, a heat balance for steaming, and a drying schedule that protects the kernel.
Paddy consists of the silica-rich husk (lemma and palea) enclosing brown rice, which comprises bran (pericarp, seed coat, nucellus, aleurone), germ and starchy endosperm, eaten as white rice. By mass, husk is commonly about a fifth of paddy, and a modern mill typically recovers milled rice (whole plus broken) of the order of two-thirds to seven-tenths of the paddy mass; the exact figures depend on variety, grain shape, drying history and degree of milling. Husk is burnt as boiler and dryer fuel; bran is the raw material for rice bran oil. Milled rice is split into head rice and brokens. Head rice is usually defined as kernels at least three-quarters of the length of a whole kernel, but national standards differ, so the definition used must be stated. Head rice sells well above brokens, so head rice yield is the main measure of mill and dryer performance.
A modern rice mill follows a standard sequence. Pre-cleaning removes straw, dust, stones and other foreign matter with scalpers, sieves, aspirators and destoners, and magnets remove tramp iron. Husking removes the husk to give brown rice. Husk aspiration separates the light husk from the grain. Paddy separation returns unhusked paddy to the husker. Whitening removes the bran layers to give white rice. Polishing removes the bran dust that clings to the kernel and gives a glossy surface, often with a fine water mist. Grading separates head rice from brokens using sieves and indented cylinders (length graders). Optical colour sorting removes discoloured, chalky or foreign kernels. The traditional alternative in South Asia and elsewhere is the steel (Engelberg type) huller, a single machine in which a ribbed steel roller rotating in a screen-lined housing removes husk and bran together by friction. It is cheap but breaks more grain, yields less, and gives a husk–bran mixture unfit for oil extraction.
Two husker types dominate. The under-runner disc husker has two abrasive-coated horizontal discs, the upper stationary and the lower rotating; paddy fed at the centre is husked by pressure and shear as it moves outward through an adjustable clearance. It is robust and cheap but breaks more grain. The rubber-roll husker has two rubber-surfaced rolls rotating in opposite directions at different peripheral speeds; the speed difference shears off the husk of paddy gripped in the gap. It gives higher head rice yield and is standard in modern mills, but the rubber wears. Husker performance is described by three coefficients. The coefficient of husking (hulling) is 1 − (mass of paddy in the output/mass of paddy fed). The coefficient of wholeness of kernels is the mass of whole brown rice divided by the total mass of brown rice produced. The husking (hulling) efficiency is the product of the two. Mills usually husk most but not all paddy per pass and recycle the rest, because harder settings raise breakage.
After aspiration, the husker output is a mix of brown rice and unhusked paddy. Paddy separators exploit differences in specific gravity, surface roughness, size and shape: the compartment type uses a tilted, oscillating deck divided by zig-zag walls, and the tray type uses an oscillating inclined tray with dimpled surfaces; brown rice and paddy migrate to opposite sides. Whitening machines remove bran by abrasion or friction. Abrasive whiteners use rotating stones or emery-coated cones (vertical cone and horizontal types) to scratch off bran; friction whiteners rely on grain-to-grain rubbing under pressure in a chamber with a screen. The degree of milling (degree of polish) is commonly expressed as the percentage of the brown rice mass removed as bran, (1 − mass of white rice/mass of brown rice) × 100. More milling lowers yield and nutrients. Milling yield figures are always relative to a stated basis: total milled rice yield and head rice yield are expressed per unit mass of paddy, whereas degree of milling is expressed per unit mass of brown rice. Confusing these bases is the classic numerical error.
A wheat kernel consists of the bran (pericarp, seed coat and the aleurone layer, which millers remove with the bran), the germ, and the starchy endosperm, which makes up roughly four-fifths or more of the kernel. White-flour milling aims to separate as much endosperm as possible, at flour fineness, with as little bran (which darkens flour) and germ (whose oil turns rancid) as possible. The crease that runs along the kernel makes it impossible to peel bran off as rice bran is removed, so wheat is opened and scraped rather than abraded. The flour extraction rate is the mass of flour produced per 100 units of wheat milled. White flour extraction is commonly in the range of about 70 to 78 per cent; wholemeal flour approaches 100 per cent. In South Asia much wheat is ground whole into atta in stone or disc attrition mills (chakkis), which is a size-reduction operation rather than a separation. Cleaning uses separators, aspirators, destoners, indented cylinders, magnets and scourers. Conditioning (tempering) follows: water is added to raise moisture to a target, and the wheat rests in bins for a set lying time. Conditioning toughens the bran so it stays in large flakes rather than shattering into fine specks that contaminate flour, and it mellows the endosperm so that it breaks cleanly and needs less energy. Hard wheats need more water and a longer rest than soft wheats; targets of roughly 15 to 16.5 per cent moisture for hard wheat, slightly lower for soft, and rests from several hours up to about a day are typical, but each mill sets its own schedule by wheat type. The water to add is found from the constant-dry-matter balance of Chapter 1.
The milling proper uses roller mills: pairs of hardened steel rolls rotating toward each other at different speeds. The break system has corrugated (fluted) rolls with a large speed differential, commonly about 2.5 to 1, which open the kernel and scrape endosperm from the bran in several successive breaks. After each passage, plansifters (stacks of oscillating sieves) separate the stock into bran flakes going to the next break, middlings and semolina (coarse endosperm particles), and a little break flour. Purifiers clean the middlings: an oscillating inclined sieve with an upward air current lifts light bran particles and grades the endosperm. The reduction system has smooth rolls with a smaller differential, commonly around 1.25 to 1.5 to 1, which reduce purified middlings to flour gradually. Many flour streams of differing ash are blended into grades. Flour yield and flour quality pull against each other: pushing extraction higher adds bran particles and raises ash. Durum wheat is milled mainly to semolina for pasta, with more purification and less reduction.
Pulse milling converts whole legume seeds into dhal (dal): dehulled, usually split cotyledons. It is difficult because in many pulses gums and mucilage bind the husk to the cotyledons. Pigeon pea, black gram and green gram are known as difficult to dehull; chickpea and peas are easier. Pretreatments loosen the husk before mechanical dehulling. In the wet method grain is soaked and then sun dried, often with repeated wetting and drying. In the dry method grain is pitted or scratched on an abrasive roller to break the husk surface, smeared with a small quantity of edible oil, then wetted and dried in the sun over several days, and the process may be repeated; heat, chemical (for example sodium bicarbonate solution) and enzyme treatments have been developed to speed this. Dehulling uses abrasive (emery) rollers or cones, under-runner discs or attrition machines; splitting often follows by impact. Output is sieved and aspirated, and unhulled grain recycled. Dhal recovery is always well below the theoretical figure set by the cotyledon fraction, because abrasion removes cotyledon along with the husk as powder and because brokens are produced; the commonly quoted example is pigeon pea, whose theoretical dhal yield of roughly 85 per cent contrasts with commercial recoveries often around 70 to 75 per cent.
Oilseeds differ greatly in oil content, which decides the process route. Approximate ranges quoted in standard texts are about 40 to 50 per cent for groundnut kernels and sesame, 35 to 45 per cent for rapeseed-mustard and sunflower (whole seed), 18 to 22 per cent for soybean, and 15 to 20 per cent for cottonseed and rice bran; values vary with variety, growing conditions and whether hull is included. Preparation improves yield in every route. Dehulling or decorticating (groundnut, sunflower, cottonseed) removes hulls that carry little oil but absorb it and add fibre to the meal. Size reduction by cracking and then flaking between smooth rolls ruptures oil cells and creates thin flakes with a short diffusion path. Cooking or conditioning with steam then coagulates protein, lowers oil viscosity, adjusts moisture to the value that gives the best pressing, and in some seeds inactivates enzymes and anti-nutritional compounds; in cottonseed it binds gossypol. The optimum moisture and temperature differ for each seed.
Mechanical expression squeezes oil out under pressure. The traditional ghani (kolhu) is a rotating pestle in a mortar, once bullock-driven, now motorised; it is slow, leaves much oil in the cake, and is valued for traditional-flavour oils such as mustard. The hydraulic press applies pressure to seed held in cloth or cages; it is a batch machine used for high-value oils. The screw press (expeller) is the main continuous machine: a rotating worm shaft conveys seed through a barrel made of spaced bars (a cage), and the volume between successive flights decreases along the shaft, so pressure rises; oil escapes between the bars and the cake leaves through an adjustable annular gap (the choke) at the discharge end. Pressure, and therefore residual oil, is set by the shaft's compression ratio, the choke setting, shaft speed, and the seed's moisture and temperature. Full-press expellers typically leave several per cent oil in the cake, commonly about 4 to 8 per cent, and ghanis more; the figure depends on seed and operation.
Solvent extraction dissolves oil out of prepared flakes or pressed cake with a solvent; commercial hexane, a petroleum fraction boiling at about 65 to 70 °C, is the industrial standard; it brings residual oil in the meal below about 1 per cent. Extractors work by percolation or immersion, with flakes and solvent counter-current so the freshest solvent meets the most exhausted flakes. The oil–solvent solution (miscella) is distilled to recover oil, and the solvent is condensed and reused. The spent flakes pass through a desolventiser-toaster, where steam removes residual solvent; for soybean meal the toasting also destroys trypsin inhibitors, making the meal suitable for animal feed. Low-oil seeds such as soybean are usually extracted directly; high-oil seeds are pre-pressed in an expeller to a cake of intermediate oil content and the cake is then solvent extracted (pre-press solvent extraction). Hexane vapour is explosive, so plants are flame-proof, recover vapour and are regulated; supercritical carbon dioxide serves specialised uses.
Crude oil contains phospholipids (gums), free fatty acids, pigments, waxes and odorous compounds. Refining removes these in steps: degumming with water or acid; neutralisation with alkali, which turns free fatty acids into soapstock (chemical refining), or their removal by steam distillation at high temperature and vacuum (physical refining); bleaching with activated earth or carbon; dewaxing by chilling for oils such as sunflower and rice bran; and deodorisation by steam stripping under vacuum. Rice bran must be stabilised by heat soon after milling, because lipase in the bran rapidly produces free fatty acids that make oil refining costly. Oil recovery is calculated by a mass balance on the oil-free solids, which pass unchanged into the cake. If seed of mass F holds oil fraction x_F and the cake holds oil fraction x_C, the oil-free solids F(1 − x_F) equal C(1 − x_C), so C = F(1 − x_F)/(1 − x_C); oil in cake is C x_C, oil expressed is F x_F − C x_C, and recovery is this divided by F x_F.
Cleaning removes material that is not the crop; grading sorts the crop into classes. Every separator exploits a property in which wanted and unwanted particles differ: width and thickness (sieves with round and slotted holes), length (indented cylinders and discs), shape and ability to roll (spiral separators, inclined belts), terminal velocity and aerodynamic properties (aspirators and air columns), specific gravity at the same size (gravity tables and destoners), surface texture (velvet rolls, magnetic separators using iron powder that sticks to rough seed coats), colour and reflectance (optical sorters), and magnetic susceptibility (magnets for tramp iron). The terminal velocity of a particle is the air velocity at which drag balances weight, v_t = [2mg/(ρ_a A C_d)]^0.5, with A the projected area and C_d the drag coefficient. Aspirators run an air velocity between the terminal velocities of the light impurities and the grain so the impurities lift away. The air-screen cleaner combines aspiration with a scalping screen (grain passes, large trash stays on top) and a grading screen (grain stays on top, sand and small seeds pass).
A screen splits a feed F into an overflow P (oversize) and an underflow U (undersize). If the desired product is the oversize and x_F, x_P and x_U are the mass fractions of oversize material in feed, overflow and underflow, the material balances are F = P + U and F x_F = P x_P + U x_U, which give P/F = (x_F − x_U)/(x_P − x_U). The effectiveness based on the oversize is E_P = P x_P/(F x_F), the fraction of feed oversize recovered in the overflow; the effectiveness based on undersize is E_U = U(1 − x_U)/[F(1 − x_F)], the fraction of undersize that passes. The overall effectiveness is their product, E = E_P E_U. A screen with no apertures recovers all the oversize (E_P = 1) but passes no undersize (E_U = 0); the product penalises both kinds of misplacement. The same form applies to any two-way separator. Effectiveness drops when feed rate is too high or near-size particles blind the apertures.
The indented cylinder (trieur) is a slowly rotating horizontal cylinder whose inside surface has pocket-shaped indents. Short particles such as broken rice kernels or round weed seeds lodge in the indents, are lifted and drop into a collecting trough, while long particles slide out of the indents and leave at the far end; indent size is chosen to fall between the lengths to be separated. Disc separators use indented discs rotating through the grain mass for the same length-based separation. The spiral separator lets seeds slide down an inclined spiral: round seeds roll fast and fly out to an outer flight, while flat or angular seeds stay inside. The gravity separator (specific-gravity table) is an inclined, vibrating, porous deck with air blown up through it; the air stratifies the bed and the vibration conveys the heavy particles uphill while the light ones float and flow downhill, so particles of the same size are separated by density. It removes light, insect-damaged, shrivelled and immature seeds and is central to seed upgrading. Optical sorters reject discoloured kernels with air jets.
Seed must be genetically pure, clean and able to germinate vigorously, so seed processing (conditioning) is designed for purity and viability rather than throughput. Seed is received and sampled, pre-cleaned, dried at lower temperature than grain to protect the embryo, specially prepared where needed (debearding barley, delinting cotton, shelling maize), cleaned on an air-screen cleaner, upgraded on indented cylinders, spirals, gravity tables and colour sorters, treated with protectants at the label rate, then weighed, bagged and labelled. Drops, fast elevators and screw conveyors damage seed, so gentle handling is preferred. Purity, germination, moisture and vigour are tested under the International Seed Testing Association (ISTA) Rules or national rules, such as those under India's Seeds Act of 1966 or the AOSA rules in North America. Certification schemes (national, and the OECD Seed Schemes for international trade) set field and laboratory standards, which differ by country and crop. Harrington's rules of thumb express the effect of storage conditions on seed life: within the usual ranges, each 1 per cent fall in seed moisture or each 5 °C fall in temperature roughly doubles storage life; they are approximations, not design values.
Size reduction breaks particles by compression, impact, shear or attrition; nearly all the energy ends as heat. Three empirical laws relate energy to size. Rittinger's law says energy is proportional to new surface created: E = K_R(1/L_2 − 1/L_1), where L_1 and L_2 are feed and product sizes; it fits fine grinding best. Kick's law says energy is proportional to the size-reduction ratio: E = K_K ln(L_1/L_2), so the same energy is needed to go from 10 mm to 5 mm as from 1 mm to 0.5 mm; it fits coarse crushing best. Bond's law lies between them: E = W_i(10/√L_2 − 10/√L_1), with sizes in micrometres taken as the 80 per cent passing sizes of product and feed, and W_i the work index, the energy (kWh per tonne; check whether short or metric) to reduce the material from very large size to 80 per cent passing 100 µm. All three are special cases of the general differential form dE/dL = −K/Lⁿ with n = 2 (Rittinger), n = 1 (Kick) and n = 1.5 (Bond). To scale a known duty to a new one, take the ratio of the size terms; the constant cancels.
Hammer mills shatter material by impact, with a screen setting fineness; attrition (plate) mills grind by shear between discs, as in the chakki; roller mills crush between rolls, as in flour milling and flaking. For a pair of smooth rolls to draw a particle in, the nip angle must satisfy friction: if the half-angle of nip is b, then cos b = (R + d)/(R + r), where R is the roll radius, d half the roll gap and r the particle radius, and the particle is drawn in only if the coefficient of friction μ ≥ tan b. Ball mills tumble balls in a rotating drum; the critical speed at which balls just stay on the wall is N_c = 42.3/√(D − d) revolutions per minute (D drum diameter and d ball diameter in metres), and mills run below it, commonly around two-thirds to three-quarters of critical. Product fineness for feed is described by the fineness modulus from a standard sieve set; the average particle size is related to it by the Henderson–Perry form D = 0.104 × 2^FM mm (equivalently 0.0041 × 2^FM inch). ASABE S319 uses geometric mean diameter instead.
Belt conveyors are efficient and gentle but limited in incline. Capacity is Q = ρAv (A load cross-section, v belt speed); power covers the empty belt, horizontal friction of the load, and lift, Q g H. The bucket elevator lifts grain vertically in buckets fixed to a belt or chain. Centrifugal-discharge elevators run fast with spaced buckets, so material is thrown out at the head pulley; the speed at which the contents become weightless at the top of the pulley is v = √(gr), i.e. N = 42.3/√D rpm for head pulley diameter D in metres, and design speed is set so discharge clears the next bucket into the spout. Continuous-bucket elevators run slower with closely spaced buckets, the contents sliding over the back of the bucket in front, which is gentler for seed and fragile products; positive-discharge elevators invert buckets over a snubbing pulley. Capacity is (bucket volume × fill fraction × bulk density × belt speed)/bucket spacing. Screw (auger) conveyor capacity is Q = (π/4)(D² − d²) p N ρ φ, with D screw diameter, d shaft diameter, p pitch, N speed, ρ bulk density and φ the loading (fill) fraction, and capacity falls as inclination rises. Screws are compact but need more power per tonne than belts and damage seed.
Pneumatic conveyors carry grain in an air stream through pipes. Suction systems collect from several points; pressure systems deliver to several. Dilute-phase conveying suspends particles at high velocity and low solids loading; dense-phase pushes slugs slowly, more gently. In horizontal dilute-phase flow, particles fall out and form dunes if velocity drops below the saltation velocity; in vertical flow the corresponding limit is the choking velocity, and air velocity must be well above the terminal velocity of the particles. Cyclones separate air from grain and rotary airlocks feed pressurised lines. They are flexible but power-hungry and crack grain at bends. Grain dust suspended in air and ignited by a spark or hot bearing explodes, and the primary blast can raise dust for a larger secondary one. Elevator legs are a known origin, so plants use dust control, explosion venting, bearing and belt-alignment monitors and grounding, under regulations that differ by country.
A student uses this subject in the post-harvest laboratory and in design assignments: determining moisture by the oven method and checking a meter against it, plotting a thin-layer drying curve and fitting Lewis and Page constants, measuring husking efficiency on a laboratory rubber-roll sheller, finding screen effectiveness from sieve analysis, and computing the water to add for wheat conditioning. For examinations, the priority is fluency with the constant-dry-matter balance, psychrometric equations and mixing, husking coefficients, screen effectiveness and the three size-reduction laws, because these carry most of the calculable marks, and the ability to say why parboiling raises head rice yield.
An engineer five years into work at a rice mill, dryer manufacturer or oil plant runs these calculations daily: sizing a dryer batch and its fuel from the water to be removed, setting LSU discharge rate and air temperature to hit a moisture target without fissuring, tracking head rice yield and husker coefficients shift by shift to decide when to adjust or replace rubber rolls, balancing expeller choke settings against residual oil in cake, specifying bucket elevator speed and screw conveyor size for a new line, and diagnosing a drop in dhal recovery. The same person handles safety: dust control, hexane handling, and lock-out on conveyors.
With twenty-five years' experience, the engineer decides at plant and policy level: choosing between parboiled and raw milling lines for a market, selecting dryer type and fuel (husk, biomass, gas, solar hybrid) for a region's harvest season, setting a mill's drying and tempering protocols to protect head rice across varieties, planning capacity and layout of an integrated rice, pulse or oil complex, and specifying equipment against national standards. They train operators, negotiate quality clauses with buyers on moisture and broken percentages, and advise government procurement agencies on drying and storage infrastructure to cut post-harvest losses.
A researcher with this depth works on the open problems: mechanistic models of simultaneous heat, mass and stress development in kernels that predict fissuring; sorption and diffusivity data for underused crops and new varieties; low-energy and renewable drying, including heat pumps and solar–biomass hybrids; parboiling kinetics and their links to starch structure, glycaemic response and nutrient retention; enzyme and other pretreatments that lift dhal recovery; safer alternatives to hexane; and sensor-driven control of dryers and mills. They also design the standard test methods and supervise students who measure the properties that practitioners rely on.
This manual does not yet cover: storage structures and bin loads (Janssen, Airy), which sit in the storage-engineering subject; maize dry and wet milling, oat and barley processing, and millets in any depth; detailed dryer design worked through to fan selection with ASABE D272 constants (the constants are in a paid standard and are not reproduced); crop-specific safe storage moisture tables, oven-method times and temperatures, and conditioning schedules, which vary by standard and are left to the governing document; quantitative rice bran oil and refining process conditions; and country-by-country grading standards for head rice, dhal and flour. Numerical ranges given for yields and oil contents are typical textbook ranges, not design values.
Original practice questions written to the syllabus and pattern of the exams named in each question's examTags. They are not past papers and do not reproduce any board's questions. Every numerical answer was computed by script; constants marked hypothetical are for practice only.
These notes are free to read here, with no account.
What the Laisha app adds:
Free for ten days, then a subscription. We would rather write that here than have you find it later.