GATE Production and Industrial Engineering (PI): Metrology, Quality and Reliability (limits and fits, vernier and screw-thread measurement, surface roughness, control charts, acceptance sampling, six sigma, reliability, MTBF and replacement), previous-year questions solved

Every quality and reliability question from the GATE Production and Industrial Engineering papers of 2024, 2025 and 2026, solved: limits and fits, vernier and two-wire screw-thread measurement, flatness and surface roughness, control charts, acceptance sampling and OC curves, DPMO and six sigma, Poisson defect counts, system reliability, hazard rate, availability and MTBF, total productive maintenance and machine replacement. Each answer matches the official key and was recomputed by code with the numbers printed in the paper, by a second method where one exists.

Questions from official GATE PI papers (IISc Bengaluru 2024, IIT Roorkee 2025, IIT Guwahati 2026); solutions by Skill-Lync. · Updated 19 Nov 2026 · 18 questions

GATE 2024 · PI · Q.57NAT · 2 marksMetrology: limits, fits, gauges and measurement

Two-wire method on a single-start metric thread of 1.4 mm pitch: diameter of the best wire in mm?

The pitch of a metric screw thread is calculated from pitch circle diameter measurement through two-wire method. If the thread is single-start with calculated pitch of 1.4 mm then the diameter (in mm) of the best-wire is __________ . (Rounded off to 2 decimal places)

Answer: 0.81 mm Official key: 0.79 to 0.81

  1. In the two-wire method two wires of equal diameter are laid in opposite thread grooves and the distance over them is measured with a micrometer. The effective (pitch) diameter then follows from that reading, the wire diameter and the thread form.
  2. The best wire is the size that touches the flanks exactly on the pitch line, where the width of the thread space equals half the pitch. Then the effective diameter does not depend on small errors in the flank angle. For a thread of included angle this gives .
  3. A metric thread has an included angle of , so . A single-start thread has lead equal to pitch, so .
  4. 0.81 mm. This is inside the key range 0.79 to 0.81.

Check: we modelled the 60 degree groove, put a wire in it touching both flanks, and searched for the wire size at which the width of the groove at the contact points was . The search gave 0.8083 mm, the same as the formula. Trap: using the whole pitch angle or the angle in the cosine gives 1.40 mm.

Where this shows up in industry

Thread gauge rooms measure the pitch diameter of taps, lead screws and critical fasteners with a micrometer over best-size wires from a calibrated set. The measured value is compared with the limits in the thread standard to accept or reject the part.

GATE 2025 · PI · Q.35NAT · 1 markMetrology: limits, fits, gauges and measurement

Maximum possible clearance between a hole 30 +0.04/-0.02 mm and a shaft 30 +0.06/+0.03 mm

The hole and the shaft dimensions (in mm) are given as

Hole dimension

Shaft dimension

The maximum possible clearance (in mm) is________. (Rounded off to two decimal places)
Tolerance zones: hole from 29.98 to 30.04 mm and shaft from 30.03 to 30.06 mm, maximum clearance 0.01 mmbasic size 30 mmHoleShaft+0.04-0.02+0.06+0.03max clearance 0.01 mmDeviations from 30 mm (mm)

Answer: 0.01 mm Official key: 0.01 to 0.01

  1. Limits and fits: clearance is hole size minus shaft size. It is largest when the hole is at its largest and the shaft at its smallest: . It is smallest (most negative, meaning interference) when the hole is smallest and the shaft largest: .
  2. Limit sizes. Hole: mm and mm. Shaft: mm and mm (both deviations of the shaft are positive, so the whole shaft zone lies above the basic size).
  3. 0.01 mm.
  4. For the other extreme, mm, an interference of 0.08 mm. Because the fit can give both clearance and interference, it is a transition fit.

Trap: subtracting the wrong pair of limits, for example mm, or taking the shaft's smaller limit as 30.00 because 'the tolerance starts at the basic size'. The shaft's lower deviation is , not zero. We checked this by running through every hole and shaft size on a 0.001 mm grid inside the two zones: the largest hole minus shaft is 0.01 mm.

Where this shows up in industry

Fits on shafts and bores, such as bearing seats and gear hubs, are checked from the deviation tables of ISO 286 before drawings are released. A toolroom uses the maximum clearance to know the loosest assembly it must tolerate, and the minimum clearance to know how much press force an interference case needs.

GATE 2025 · PI · Q.61NAT · 2 marksMetrology: limits, fits, gauges and measurement

Vernier scale reading: zero between 20 and 21 mm, 4th vernier division coincides, 5 VSD = 4 MSD

The zero line of the Vernier scale lies between divisions 20 and 21 of the main scale. The 4th Vernier scale division exactly coincides with a main scale division. The 5 divisions of the Vernier scale are equal to 4 divisions of the main scale. If one main scale division is 1 mm, the measured value (in mm) is __________. (Rounded off to one decimal place)
Vernier scale: zero line at 20.8 mm on the main scale, 4th vernier division in line with the 24 mm mark202122232425mm0123454th vernier division is in line with 24 mmvernier zero line at 20.8 mm

Answer: 20.8 mm Official key: 20.8 to 20.8

  1. Vernier principle: the vernier divisions are slightly shorter than the main scale divisions, so the least count is the smallest length the instrument can resolve. The reading is the main scale value just before the vernier zero, plus (number of the coinciding vernier division) .
  2. mm, so mm and mm.
  3. Main scale reading mm (the zero line is between 20 and 21). Vernier part mm.
  4. Reading 20.8 mm.

Check: the vernier zero sits at 20.8 mm, and the 4th vernier division is mm further on, at 24.0 mm, exactly on a main scale line, as the question says. Trap: adding the coinciding division number as whole main divisions, which gives 24.0 mm, or using the vernier division length 0.8 mm as the least count and getting mm.

Where this shows up in industry

Vernier calipers are the everyday inspection-room tool, and a quality lab states the least count of each one on its calibration record. A least count of 0.02 mm is common on shop calipers, and the coinciding-line reading is how an inspector gets the value to that resolution.

GATE 2026 · PI · Q.17MCQ · 1 markMetrology: limits, fits, gauges and measurement

Match sine bar, profilometer, vernier calliper and ring GO gauge to their most common applications

Which ONE of the following CORRECTLY matches the metrology instruments and their respective most common applications?
Metrology instrumentApplication
PSine bar1Surface roughness of milled planar surface
QProfilometer2Planar surface inclination of a part
RVernier calliper3Checking shaft diameter tolerance without direct measurement
SRing GO gauge4Length of any side of a cuboid
  1. (A)P-2; Q-1; R-4; S-3
  2. (B)P-4; Q-3; R-2; S-1
  3. (C)P-4; Q-1; R-2; S-3
  4. (D)P-2; Q-1; R-3; S-4

Answer: (A) Official key: A

  1. Rule: match each instrument to the quantity it senses. A sine bar sets an angle from a known length and a height of gauge blocks, so it gives the inclination of a surface (P-2). A profilometer drags a stylus over the surface and records its height profile, so it gives surface roughness (Q-1).
  2. A vernier calliper reads a length directly from a scale, so it gives the length of any side of a cuboid (R-4). A ring GO gauge is a limit gauge: the shaft either enters the ring or does not, so the diameter tolerance is judged without measuring the diameter (S-3).
  3. The combination is P-2, Q-1, R-4, S-3, which is option (A).
  4. Elimination: (B) and (C) give the sine bar a length job (P-4) and the vernier an angle job (R-2); (B) also gives the profilometer a gauging job. (D) gives the vernier the gauging job (R-3) and the gauge a length job (S-4).

Trap: a GO gauge does not give a number. It only says the size is not beyond one limit, so S-4 can never be right. We checked each option against what each instrument can do, and only (A) has every pair possible.

Where this shows up in industry

Every machine shop inspection room keeps these four: a sine bar with gauge blocks for angled faces, a stylus roughness tester for finished surfaces, a vernier or digital calliper for quick lengths, and go/no-go ring gauges at the line to accept or reject shafts at once.

GATE 2026 · PI · Q.40MCQ · 2 marksMetrology: limits, fits, gauges and measurement

Flatness and parallelism of a top surface from a figure: which zone, a or b, is which

A student is trying to determine the flatness and parallelism of the top surface of a part whose cross-section, assumed uniform in the Z direction, is shown in the figure.

Which ONE of the following options is CORRECT?

Two views of the same part, a block whose top surface rises slightly from left to right with a small ripple, standing on a datum. Left view: two parallel lines tilted to follow the surface enclose it, with a dashed line between them, and the gap between them is marked a. Right view: two horizontal lines parallel to the datum enclose the same surface, with a dashed line between them, and the gap is marked b. X points right and Y points up.aDatumbDatumXY
  1. (A) is flatness while is parallelism.
  2. (B) is parallelism while is flatness.
  3. (C) is parallelism while is flatness.
  4. (D) is flatness while is parallelism.

Answer: (A) Official key: A

  1. Rule (ISO 1101 form tolerances): flatness is the width of the narrowest pair of parallel planes that contains the whole surface, and those planes may take any orientation. The surface is judged against itself, not against another feature.
  2. Parallelism is an orientation tolerance: the surface must lie between two planes that are parallel to a datum, so the planes cannot tilt. Any slope of the surface relative to the datum counts as error.
  3. In the left view the two lines tilt with the surface, so their gap is the flatness. In the right view the two lines are parallel to the datum line, so their gap is the parallelism.
  4. Because the tilted zone can hug the surface while the datum-bound zone cannot, always . Options (C) and (D) combine them as and , which are neither of the two zones. Option (B) swaps the two. The answer is (A).

Check: for a rippled, rising profile like the one drawn, the tightest zone is 5.88 units wide when tilted by 16.6 degrees, but 28.85 units wide when held parallel to the datum. On 2000 random profiles the free-orientation zone was never wider than the datum-bound zone.

Where this shows up in industry

On a surface-plate or CMM report, flatness of a machined face and its parallelism to the base are two separate callouts, and a milled face that is perfectly flat but slightly tilted passes the first and fails the second.

GATE 2026 · PI · Q.54NAT · 2 marksMetrology: limits, fits, gauges and measurement

Average surface roughness Ra from a 2D stylus profilometer trace with areas above and below the datum

In the measurement of surface roughness using a 2D stylus profilometer, a surface profile measured over a length of 0.8 mm was recorded on a graph paper. During recording, vertical magnification of 10,000 and horizontal magnification of 100 were used. The areas in the as-recorded graph above and below the datum line are as follows:

Above (mm2): 140, 60, 150, 50
Below (mm2): 70, 50, 120, 160

The average surface roughness () of the surface is ______ m (rounded off to two decimal places).

Answer: 1.00 um Official key: 1 to 1

  1. Rule: is the arithmetic mean of the absolute heights from the datum (mean) line over the sampling length: . On a recorded graph, every area counts as positive, whether above or below the datum.
  2. Total area on the graph: above mm; below mm; together mm.
  3. Length on the graph: mm. Mean height on the graph: mm.
  4. Undo the vertical magnification: mm 1.00 m.

Traps: using the net area (above minus below ) gives a zero that is meaningless, since opposite signs must not cancel in . Forgetting the horizontal magnification gives 100 m, which is 100 times too large. We rebuilt an eight-lobe trace with these areas, scaled it back by both magnifications and integrated numerically over 0.8 mm: 1.00 m, and the signed integral is zero.

Where this shows up in industry

Roughness on machined faces such as bores, sealing surfaces and bearing seats is checked with a stylus instrument that reads Ra directly; the arithmetic behind it, absolute areas over the sampling length, is the same as on this old chart-recorder printout.

GATE 2026 · PI · Q.63NAT · 2 marksMetrology: limits, fits, gauges and measurement

Clearance between a hole at maximum material and a shaft at minimum material, 22 mm limits

In a hole-shaft assembly, the following are the dimensions and tolerances (all in mm) provided for the hole and shaft:

Hole:

Shaft:

The magnitude of the clearance when the shaft is at the minimum material condition and hole is at the maximum material condition is ______ mm (to three decimal places).

Tolerance bands drawn on a size axis in micrometres above 22.000 mm. The hole band runs from 10 to 18 and the shaft band from 4 to 75. The hole at maximum material is 10 and the shaft at least material is 4, so the gap between those two sizes is 6 micrometres, which is 0.006 mm020406080size above 22.000 mm (um)HoleShaft18 (hole LMC)10 (hole MMC)75 (shaft MMC)4 (shaft LMC)6 um

Answer: 0.006 mm Official key: 0.006 to 0.006

  1. Limits: hole 22.010 to 22.018 mm; shaft 22.004 to 22.075 mm.
  2. Rule: maximum material condition (MMC) is the size at which a part holds the most material. For a hole that is its smallest size, 22.010 mm. Least (minimum) material condition (LMC) is the size with the least material. For a shaft that is its smallest size, 22.004 mm.
  3. Clearance hole size shaft size 0.006 mm.

Trap: for a shaft, MMC is the largest size and LMC the smallest, the reverse of the way many people picture it. Taking the shaft at 22.075 mm gives mm, an interference of 0.065 mm. Taking the largest clearance instead gives 0.014 mm. We enumerated every hole and shaft size on a 1 micrometre grid: the clearance runs from -0.065 mm (interference) to 0.014 mm, so this is a transition fit, and the asked case is 0.006 mm.

Where this shows up in industry

Fit calculations on drawings use exactly these cases: a toolroom checks the tightest and loosest assembly of a bearing seat and a shaft from the limits, and a fit that can swing between a clearance and an interference is flagged to the assembler as a transition fit.

GATE 2024 · PI · Q.44MSQ · 2 marksStatistical quality control and six sigma

Acceptance sampling: which statements on type I and type II error, sample size and risks are true?

Which of the following statement(s) is/are TRUE for a given acceptance sampling plan?
  1. (A)Type II error decreases with an increase in type I error.
  2. (B)The probability of rejecting a good quality lot is producer’s risk.
  3. (C)Type II error decreases with a decrease in sample size.
  4. (D)The probability of rejecting a good quality lot is consumer’s risk.

Answer: (A) and (B) Official key: A;B

  1. An acceptance sampling plan inspects items from a lot and accepts the lot if at most are defective. Type I error is rejecting a lot of good quality (at the acceptable quality level). It is the producer's risk. Type II error is accepting a lot of bad quality (at the limiting quality). It is the consumer's risk.
  2. (B) is true by definition. (D) swaps the two risks, so it is false.
  3. (A): for a given sample size, the two errors trade off. A stricter plan (smaller ) rejects more good lots, so rises, and also accepts fewer bad lots, so falls. Hence falls when rises: true.
  4. (C): a smaller sample carries less information, so more bad lots slip through. For a fixed , cutting raises , it does not lower it: false.

Check: with exact binomial OC curves for , AQL 2 % and LTPD 10 %, going from up to each step lowered and raised (opposite directions at every step). At the producer's risk was 0.078 and the consumer's risk 0.112, so they are different numbers. With fixed, went from 0.053 at up to 0.930 at as shrank. Trap: mixing up which risk belongs to which party.

Where this shows up in industry

Receiving inspection departments choose a sample size and an acceptance number from tables such as ISO 2859-1 and agree the risk split with suppliers in the quality agreement. A buyer who wants a lower chance of accepting bad lots usually asks for a bigger sample or a lower acceptance number.

GATE 2025 · PI · Q.34NAT · 1 markStatistical quality control and six sigma

Defects per million opportunities of a gear box line: 15 CTQ features, 40000 boxes, 1500 defects

While inspecting final assembly of automobile-gear-boxes, 15 features were considered critical-to-quality (CTQ). During last quarter, 40000 gear boxes were produced among which 1500 defects were found of the CTQ features. The defects per million opportunities (DPMO) is________. (Answer in integer)

Answer: 2500 Official key: 2490 to 2510

  1. Six sigma measure: DPMO counts defects against every chance to go wrong, not only against units. An 'opportunity' is one critical-to-quality (CTQ) feature that could be wrong, so .
  2. Opportunities in the quarter: (600,000) chances for a defect.
  3. 2500.

Trap: leaving out the 15 features and dividing by the 40,000 boxes, which gives (that is defects per million units, DPMU, and is 15 times too large). Defects per unit is 0.0375 here, so per opportunity it is 0.002500, or 2500 in a million.

Where this shows up in industry

Automotive suppliers report DPMO (and the sigma level that follows from it) to customers because it lets a simple gear box and a complex one be compared fairly. The count of opportunities per unit is agreed with the customer in the quality plan, since a larger count lowers the figure without any process improvement.

GATE 2025 · PI · Q.62NAT · 2 marksStatistical quality control and six sigma

Upper control limit of an X-bar chart: mean 10.56 mm, standard deviation 0.05 mm, sample size 5

A broaching machine makes key slots with a mean dimension of 10.56 mm and a standard deviation of 0.05 mm. The upper control limit for mean of sample size 5 calculated using X-bar () chart is __________. (Rounded off to two decimal places)

Answer: 10.63 mm Official key: 10.61 to 10.65

  1. X-bar chart: it plots the mean of each sample. The mean of readings varies less than a single reading, with standard deviation (the standard error). The 3-sigma limits are , which enclose 99.73 percent of the sample means while the process is steady.
  2. Standard error: mm.
  3. , which is 10.63 mm to two decimals. (The lower limit is mm.)

Trap: forgetting to divide by , which gives mm. That is the limit for individual slots, not for sample means, and it sits outside the key range. We simulated 200,000 samples of size 5 from a normal distribution with mean 10.56 mm and standard deviation 0.05 mm: 0.27% of the sample means fell outside the two limits, close to the 0.27 percent expected for 3-sigma limits.

Where this shows up in industry

Control charts on slot widths, bore diameters and similar machined features run on the shop floor with samples of 4 or 5 parts taken at set intervals. The tighter limits on the mean make a small shift in the tool or fixture show up on the chart before many parts go out of tolerance.

GATE 2026 · PI · Q.49MSQ · 2 marksStatistical quality control and six sigma

Ideal operating characteristics (OC) curve: which statements about producer's and consumer's risk are true

Which ONE or MORE among the following options is/are TRUE for the ideal operating characteristics (OC) curve given in the figure?

Note: represents the acceptable quality level (AQL) and represents the lot tolerance percent defective (LTPD).

Ideal operating characteristics (OC) curve. Horizontal axis: proportion of defectives in the lot from 0 to 1, with p0 (the AQL) and p1 (the LTPD) marked. Vertical axis: probability of accepting the lot, with the value 1 marked. The thick curve is at probability 1 from 0 to p0 and at probability 0 from p1 to 1, with a dashed vertical line at p010p0p11Proportion of defectives in the lotProbability of accepting the lotIdeal OC curve
  1. (A)Producer's risk is 100%.
  2. (B)Consumer's risk is 100%.
  3. (C)Producer's risk is zero.
  4. (D)Consumer's risk is zero.

Answer: (C) and (D) Official key: C;D

  1. Definitions. An OC curve gives the probability of accepting a lot against its fraction defective . Producer's risk is the chance that a lot of good quality, at the AQL , is rejected: . Consumer's risk is the chance that a lot of bad quality, at the LTPD , is accepted: .
  2. Read the figure: for every , so . That makes (C) true and (A) false.
  3. Read the figure: for every , so . That makes (D) true and (B) false.
  4. The ideal curve is a perfect screen: every good lot is accepted and every bad lot is rejected, so both risks are zero. The answer is (C) and (D).

Trap: swapping the two risks, or reading the 1 and 0 levels as the risks themselves. A real plan never reaches this: a single sampling plan with and has a producer's risk of 0.264 at and a consumer's risk of 0.083 at . Both risks reach zero only if the lot quality can be known exactly, which is the ideal this curve stands for.

Where this shows up in industry

Incoming-inspection plans for bought-in parts are chosen by reading the OC curve of a candidate sample size and acceptance number and agreeing with the supplier on the two risks, since a curve as steep as the ideal one would need testing every piece.

GATE 2026 · PI · Q.61NAT · 2 marksStatistical quality control and six sigma

Poisson soldering defects with mean 3 per device: probability of at least two defects

The number of soldering defects that occur in a semiconductor device follows a discrete Poisson distribution with a probability mass function .

The average number of soldering defects per semiconductor device is 3.

The probability that a randomly selected semiconductor device will have at least two soldering defects is ______ (rounded off to two decimal places).

Answer: 0.80 Official key: 0.78 to 0.82

  1. Rule: for a Poisson count the mean equals , so . 'at least two' is the complement of 'zero or one': .
  2. .
  3. .
  4. 0.80.

Trap: the boundary. (more than two) and are both different from 'at least two'. A seeded simulation of 400 000 devices gave 0.801 for and a mean of 3.00 defects, in line with the closed form.

Where this shows up in industry

Electronics assembly lines track solder defects per board on a c-chart, which rests on this same Poisson model, and a quick tail probability like this tells a quality engineer how often a board with several defects should be expected before calling the process out of control.

GATE 2024 · PI · Q.26MCQ · 1 markReliability, maintenance and replacement

Which statement about Total Productive Maintenance (TPM) is NOT true?

Which one of the following is not true about Total Productive Maintenance (TPM)?
  1. (A)It allows operators to perform preventive maintenance on the machines.
  2. (B)It allows operators to perform reactive maintenance on the machines.
  3. (C)It is consistent with the Just-in-Time (JIT) system.
  4. (D)It is consistent with the Lean system.

Answer: (B) Official key: B

  1. TPM (developed in Japan from the 1970s, described by Seiichi Nakajima) aims for zero breakdowns, zero defects and zero accidents by involving everyone, from operators to managers, in keeping equipment healthy. Its pillars include autonomous maintenance, planned maintenance, focused improvement, early equipment management and education and training.
  2. In autonomous maintenance operators clean, inspect, lubricate, tighten and make small adjustments. That is preventive care done daily, so (A) is true.
  3. The whole point of TPM is to stop breakdowns happening. Repair after a failure, reactive (breakdown) maintenance, is what it tries to remove, and major repairs stay with trained maintenance staff. So the statement that TPM allows operators to perform reactive maintenance is the one that is not true: (B).
  4. JIT and lean both depend on machines being ready when called on, with no buffer stock to hide a stoppage. TPM supplies that reliability, so (C) and (D) are true.

Check: each statement was compared with the TPM pillars. Only (B) contradicts them. Trap: reading "not true" as "true" and ticking (A).

Where this shows up in industry

Plants that run TPM post daily clean-inspect-lubricate checklists at each machine and track overall equipment effectiveness. Operators tag abnormalities such as leaks or noise so that maintenance can fix them in a planned stop, before they cause a breakdown.

GATE 2024 · PI · Q.60NAT · 2 marksReliability, maintenance and replacement

Machine runs 2400 h with 8 breakdowns: availability gain in % if mean repair time falls from 20 h to 10 h?

A repairable machine operated for 2400 hours in a year and for that year the machine broke down 8 times. The mean time to repair including waiting time is found to be 20 hours for that year.

If the mean time to repair including waiting time could have been reduced to 10 hours for that year, then the improvement in the availability of that machine would be ___________ %. (Rounded off to 2 decimal places)

Answer: 3.23 % Official key: 3.19 to 3.23

  1. For a repairable machine, availability is the fraction of time it is ready to work: , where MTBF is the mean time between failures (mean operating time per breakdown) and MTTR is the mean time to repair.
  2. .
  3. Before: . After: .
  4. The improvement is taken relative to the original availability: 3.23 %, which is inside the key range 3.19 to 3.23.

Check: we built the year as 8 cycles of 300 h up and a repair, giving the same availabilities. The official range matches the relative gain. The plain difference of the two availabilities is 3.02 percentage points, which is outside the key, so that is the trap. Taking the MTBF as (2400 + 8 × 20) / 8 = 320 h would give 3.03 %, also outside.

Where this shows up in industry

Maintenance managers track availability, MTBF and MTTR for each critical machine in a computerised maintenance system. Faster repair, by stocking spares and cutting waiting time for a technician, often raises availability more cheaply than making the machine fail less often.

GATE 2025 · PI · Q.17MCQ · 1 markReliability, maintenance and replacement

Reliability of an aircraft with two engines, each of reliability R, that crashes only if both stop

An aircraft has two engines, each having a reliability R. The aircraft will crash only when both engines stop working. The reliability of the aircraft flying without crash is
  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (C) Official key: C

  1. Parallel (redundant) system: the system works if at least one unit works, so it fails only when all units fail. For independent units, . A series system, where every unit must work, would use instead.
  2. The probability that one engine fails is . Both fail with probability (independent engines).
  3. . Option (C).
  4. Sense check with : the aircraft survives with , better than one engine alone (0.9). Option (A) is the series value , worse than one engine. Option (B) gives and option (D) a negative number, and a probability can be neither.

Trap: choosing because the word 'both' appears in the question. is the probability that both engines work, which is the series case. We enumerated all four up/down states of the two engines and compared each option over : only matches at every .

Where this shows up in industry

Aircraft, hospital power supplies and plant cooling systems carry duplicate units so that one failure does not stop the system. A reliability engineer uses this one-minus-product-of-failures rule to show an auditor how much a second pump or engine raises the chance of getting through the mission.

GATE 2025 · PI · Q.63NAT · 2 marksReliability, maintenance and replacement

Replacement age that minimises the average total cost per year, original machine cost INR 70,000

The table shows the data of running a machine for five years. The original machine cost is Rupees 70,000. In order to minimize the average total cost per year for running the machine, the machine should be replaced after________ years. (Answer in integer)
1st year2nd year3rd year4th year5th year
Resale value (Rupees)4000030000250002200020000
Maintenance cost (Rupees)1910020300235003050040000

Answer: 4 years Official key: 4 to 4

  1. Replacement of a machine whose running cost rises with age and whose resale value falls: if the machine is kept for years, the average cost per year is , where is the purchase price, the resale value at the end of year and the maintenance cost in year . Choose the with the lowest average. Money is not discounted in this question.
  2. Build the table year by year (all amounts in INR):
Year Resale value (INR)Maintenance (INR)Cumulative maintenance (INR)Total cost (INR)Average cost per year (INR)
140,00019,10019,10049,10049,100
230,00020,30039,40079,40039,700
325,00023,50062,900107,90035,967
422,00030,50093,400141,40035,350
520,00040,000133,400183,40036,680
  1. The average falls from 49,100 to 39,700, then 35,967, then 35,350, and rises again to 36,680. The lowest is at 4 years, so replace after 4 years.
  2. Check with the marginal rule: keep the machine another year while that year's cost is below the running average. The year 4 cost is INR, below the 3-year average of 35,967 INR, so keep it through year 4. The year 5 cost is INR, above the 4-year average of 35,350 INR, so replace.

Trap: choosing the year with the lowest maintenance cost (year 1) or the year with the lowest total cost so far, both of which ignore that the purchase price is spread over the years of service. We computed the average for every year and the minimum is at year 4; the marginal rule gave the same year.

Where this shows up in industry

Fleet, press and machine-tool owners use this replacement calculation to decide when a maintenance bill and a falling resale value make a new machine cheaper per year. Plant maintenance departments back it with cost history from each machine's maintenance log, usually with discounting added when the sums are large.

GATE 2026 · PI · Q.41MCQ · 2 marksReliability, maintenance and replacement

Hazard rate r(t) = f(t)/(1 - F(t)) for F(t) = 1 - exp(-a t^b)

The function, , signifies the likelihood that a component has survived up to time and fails in the next instant of time.

where denotes the probability density function for time to failure with corresponding cumulative distribution function given by,

Which ONE of the following represents ?

  1. (A)
  2. (B)
  3. (C)
  4. (D)

Answer: (B) Official key: B

  1. Rule: is the hazard rate, the chance of failing in the next instant given survival so far. It equals the density divided by the reliability: with .
  2. Reliability: .
  3. Density: .
  4. Divide: . The exponentials cancel, so the answer is (B).

Trap: option (A) is the density itself, which forgets to divide by the reliability. Reading the result: for the hazard falls with time (early failures), for it is constant (, the exponential case), and for it rises (wear-out). We checked this numerically by differentiating with finite differences on a grid of , and : only (B) matched.

Where this shows up in industry

Maintenance planners fit this Weibull shape to bearing, pump and cutting-tool failure records: a shape above 1 says parts wear out and a planned replacement pays, while a shape below 1 points to early-life defects that call for better assembly and burn-in instead.

From GATE prep to a quality, inspection or manufacturing career

Talk to a counsellor.

Talk to a career counsellor

Free 20-minute call. We'll map the right course to where you want to work.

By submitting you agree to our Terms and Privacy Policy.

Talk to a career counsellor

Talk to a career counsellor

Free 20-minute call. We'll map the right course to where you want to work.

By submitting you agree to our Terms and Privacy Policy.