IESL Part I common mistakes
IESL Part I · Understand the traps and how to avoid them.
In the Irrigation and Drainage questions, candidates compute crop water requirement as ETc = Kc × ETo and stop there, ignoring effective rainfall (Pe) and irrigation efficiency, so the net and gross irrigation depths are wrong and the duty in L/s/ha is understated.
Most textbooks present ETc as the headline formula, and candidates memorise that single line without drilling the full water-balance chain: NIR = ETc − Pe, GIR = NIR/eff, then duty = depth/period converted to L/s/ha. Under time pressure they default to the first formula that comes to mind and never check whether rainfall or efficiency belongs in the answer.
What to do instead
Build a fixed four-step template and rehearse it on past-style numbers: (1) ETo from pan or Penman data, (2) ETc = Kc × ETo, (3) NIR = ETc − Pe with Pe taken as 80% of dependable rainfall, (4) GIR = NIR/efficiency, then convert to duty. Always write the units at each step so a missing efficiency term becomes visible.
In the Building Services questions, candidates select a pump, fan, or chiller duty from the calculated peak load but then size the associated pipe or duct and the pump head using the same peak flow without applying the diversity or load-factor allowance, or they mismatch the units between SI and imperial schedules in the same calculation.
Load calculations and hydraulic calculations are usually taught as separate topics, so candidates treat the peak load figure as the final design number. Unit mismatches arise because Sri Lankan practice mixes metric drawings with imperial pipe sizes and pump curves, and the candidate converts mid-calculation rather than up front.
What to do instead
Adopt a two-column working sheet: design load (with diversity factor stated explicitly) and hydraulic duty (flow, head, velocity). Convert all inputs to SI at the top of the page before any arithmetic, and state the assumed diversity factor next to the flow figure so the marker can see the reasoning.
In Structural and Geotechnical questions, candidates apply a partial safety factor combination to the wrong side of the limit-state check-for example factoring down the resisting moment in a retaining wall or using service loads in an ultimate limit state-and produce a factor-of-safety answer that is numerically plausible but conceptually wrong.
Candidates memorise factor values (1.4, 1.6, 1.2) without anchoring them to the load case they belong to. Because the arithmetic still produces a believable number, the error is not caught by a sanity check, and exam pressure rewards finishing the calculation over checking which limit state was asked for.
What to do instead
Before calculating, write one line at the top: 'ULS - factor actions, not resistances' (or SLS as asked). Keep a personal one-page factor table with the load combination named in the left column. After the answer, verify that the utilisation ratio is on the correct side of 1.0 and that the resisting term was not factored.
In the Chemical and Process Engineering questions, candidates balance the overall mass balance but ignore the recycle or purge stream, so the reactor conversion, fresh feed requirement, and purge composition are all wrong even though the overall balance closes.
Textbook problems often present a single-pass reactor, and candidates transfer that mental model to process flowsheets with recycle. They also skip drawing the system boundary because drawing takes time, so they cannot see that the recycle loop changes the fresh feed and the inert build-up.
What to do instead
Always draw the flowsheet and mark the system boundary before writing any balance. Do one balance around the reactor and a separate balance around the overall process including the purge, then check that inert concentration in the recycle reaches steady state. If a purge is present, calculate the purge fraction explicitly rather than assuming zero.
In the Electrical Engineering questions, candidates compute the magnitude of current or voltage in an AC circuit and then report it as a phasor answer, dropping the phase angle and the power factor, so power, torque, and regulation answers derived from it are wrong.
DC-circuit habits carry over: candidates are comfortable with scalar Ohm's law and treat impedance as a resistance. Complex arithmetic with j-operators feels slow under exam conditions, so the imaginary part is quietly discarded, and the resulting real number looks acceptable.
What to do instead
Work in rectangular form for addition and polar form for multiplication and division, and write the phasor with its angle every time. As a check, compute power factor = cos(angle) and confirm it lies between 0 and 1 and matches the expected leading or lagging nature of the load.
In the Electronics and Telecommunication questions, candidates substitute the mid-band gain into a bandwidth or noise calculation without accounting for the amplifier's finite bandwidth, or they compute the noise figure from a single stage and ignore the Friis cascade contribution of later stages.
Op-amp and transistor problems are drilled at mid-band, where gain is constant, so the frequency dependence is treated as a separate topic. Candidates also under-weight the Friis formula because the first-stage noise dominates, and they forget that a lossy first stage (cable, filter) raises the system noise figure.
What to do instead
State the operating frequency relative to the corner or cut-off frequency before choosing a gain formula. For cascaded systems, use F_total = F1 + (F2−1)/G1 + ... and always include the feeder loss as an attenuator with its own noise figure equal to its loss. Sanity-check that the system noise figure is never lower than the first-stage noise figure.
In the IT and Communication Engineering questions, candidates calculate a subnet mask or host range from the classful default rather than the given prefix length, or they forget to subtract the network and broadcast addresses when counting usable hosts.
Classful addressing is the first thing taught and the last thing revised, so under time pressure the candidate sees a 192.168.x.x address and assumes /24. The usable-host subtraction is a small step that is easy to skip when the arithmetic is already done.
What to do instead
Write the prefix length in binary at the top of every subnetting answer and derive the mask from it, never from the address class. Use the fixed checklist: block size = 2^(32−prefix), usable hosts = 2^host bits − 2, and verify that the broadcast address is the next network minus one.
In the Mechanical Engineering questions, candidates compute the thermal or fluid duty without checking the regime-applying the Dittus-Boelter correlation outside its Reynolds and Prandtl range, or using a constant specific heat across a large temperature span-and present a result with no validity statement.
Correlations are memorised as formulas rather than as conditional tools, and the candidate does not habitually compute Reynolds number first. Because the final number has plausible magnitude, nothing signals that the correlation was inapplicable.
What to do instead
Start every convection or flow problem by computing Reynolds number and classifying the regime. Quote the correlation with its stated validity range in the answer, and if the range is exceeded, say so and use an alternative or note the limitation. For large temperature spans, evaluate properties at the mean bulk temperature and state that choice.
Across all domains, candidates spend disproportionate time on one long calculation and leave the short structured-answer questions unattempted, so marks are lost on easy items while the hard item is still incomplete.
The paper rewards depth on some questions, and candidates who have revised one topic heavily try to recover confidence by finishing that question at any cost. There is no rehearsed time budget, so the first sign of trouble appears only when the invigilator calls time.
What to do instead
Before the exam, set a per-question time allocation from the mark weighting and rehearse it in timed mock papers. In the exam, write the allocation on the answer booklet cover, and when a question overruns, leave the calculation at a clean intermediate point with the method stated, move on, and return only if time remains.
Across all domains, candidates present a numerical answer with no units, no stated assumptions, and no engineering conclusion, so method marks are lost even when the arithmetic is correct.
Candidates treat the answer as the number and assume the marker will infer the method and assumptions. In practice the marking scheme awards method and assumption marks explicitly, and a bare number cannot earn them.
What to do instead
Use a fixed answer format: given data, assumptions listed, equations with symbols defined, substitution with units, result with units and significant figures, then one sentence interpreting the result against the design requirement. Rehearse this format until it fits within the time budget.