Three questions for one deck
A screen calculation can ask what fraction of a particular size will pass, what area is required for the feed load, or how width, length and feed distribution interact to determine recovery. Capacity estimates commonly use specific throughput in t/(h·m²) modified by factors for feed sizing, deck arrangement, density, inclination and media. A partition model converts the resulting operating point into two product size distributions. Equipment of equal nominal area can have different passage kinetics because pool depth, trajectory, aperture geometry and active open area vary along a deck.
Percentage undersize efficiency should have an explicit denominator. If the nominal aperture is h, define S₍ₓ<ₕ₎ as the feed solids below h and U₍ₓ<ₕ₎ as the component of undersize solids below h. Then
Overall mass reporting to undersize is U/S. The two agree only in special cases. The purity of the undersize is U₍ₓ<ₕ₎/U. If a flexible aperture or elongated particle permits oversize contamination, both ηᵤ and purity matter. A high purity can accompany a low recovery if little material passes. In the geometric models below, aperture width and particle geometry determine where such passage is feasible.
Efficiency as a given condition
An efficiency-driven partition is useful when there is existing deck performance evidence and the main objective is a mass balance. Given aperture h, a fines bypass floor b and a lower transition size L, one useful oversize curve is b below L, one above h and log-interpolated between.
Solve for L to match measured or specified ηᵤ on the feed size grid. Increasing L lowers pₒ for intermediate sizes and increases recovery to undersize. Target ηᵤ cannot exceed 1−b when all below-aperture material is subject to that floor. With discrete representatives or a large mass very near h, the effective feasible range can be narrower. A load-dependent efficiency map can describe one physical deck across measured flows. When that map is derived from noisy sampling, dependent model parameters are better treated as a coherent curve than as freely adjustable values.
For x = 2 mm, L = 1 mm, h = 4 mm and b = 0.10, the log ratio is 0.50 and pₒ = 0.55. Finer material with x = 0.5 mm has pₒ = 0.10. If their feed masses are 20 and 40 t/h, respectively, and 40 t/h of material above h is retained, passage is 20 × 0.45 + 40 × 0.90 = 45 t/h. ηᵤ is 45/60 = 75%, whereas total solids recovery to undersize is 45%.
Capacity-driven screens
Karra's near-size correction
The projected aperture is a function of nominal aperture h, wire diameter w and deck inclination θ.
For nominal opening 12 mm, wire diameter 3 mm and angle 20°, the projected opening is 15 cos 20° − 3 = 11.095 mm. Industrial screening response is sensitive to the near-size fraction. In the Karra form one factor is G = 0.844(1−Xₙ)³·⁴⁵³, with Xₙ in fraction units. The feed undersize loading Qᵤ/A relative to the product of capacity factors sets the corrected aperture scale, then a Rosin–Rammler-like partition acts on each class. In the implemented correlation, for A the actual deck area and K the combined capacity term,
Here hₑ is in mm, d₅₀ and x are in µm and I is a fit multiplier on the partition exponent. Both Qᵤ and area must be on a consistent one-deck or total fleet basis. If specific load equals K, d₅₀ ≈ 11,095 µm and a class at that size has pₒ ≈ 0.50. Doubling relative load in this particular correlation multiplies d₅₀ by 2⁻⁰·¹⁴⁸ ≈ 0.9025. Interpret this result through the defined modeled partition and factor package. A different convention for the reference destination or relative load changes the interpretation of that exponent.
When Xₙ rises from 0.20 to 0.40, G falls from about 0.391 to 0.145. Near-size burden can consequently dominate the result even on a deck with nominally ample area. Record the reference aperture and the interval used to determine Xₙ. A sieve analysis with no boundaries near the aperture undermines a calibration of this factor.
King's relative feed rate
One King representation computes a basic areal capacity Iᵤ and correction multiplier Kₜ, then sets relative feed R = Q/(Iᵤ Kₜ A) per deck. The efficiency relation in fractional units is
Bound η to the physical interval before applying it. This fit peaks at R = 0.8. An underloaded deck can classify less effectively in the fit, as well as an overloaded one. Above-aperture representative sizes go to oversize. Below-aperture sizes have oversize partition 1−η, a step approximation that cannot match a detailed measured partition that changes continuously with size.
Consider 240 t/h per deck and corrected capacity 200 t/h per deck. R = 1.20, so η = 0.95−0.25(0.4)−0.05(0.4)² = 0.842. If there are 150 t/h below the aperture, undersize solids are 126.3 t/h. Oversize solids are 90 + 23.7 = 113.7 t/h. At the same feed, R = 0.8 gives 142.5 t/h undersize. The oversize change is exactly the undersize gain.
Sequential panel rows
The Whiten–White form separates the probability of passing in one presentation from the number of presentation opportunities. For representative particle size x, rectangular aperture width w, length l and open fraction φ, the approximate probability kernel is
All lengths must have the same units. For N effective trials, the exponential is a survival relation. Each row receives the previous row's retained material. With row survival pⱼ(x), the deck survival is the product of row survivals. Row j passes feed fraction (1−pⱼ)Π𝓀₍𝓀<ⱼ₎pₖ for each class.
For a 2 mm class on a 4 × 8 mm opening with 40% open area, a = 0.40 × 0.50 × 0.75 = 0.15. With eight trials in row 1, p₁ = exp(−1.2) = 0.30119. Row 2 uses 3 × 6 mm openings, 35% open area and six trials. Then a = 0.35 × (1−2/3) × (1−2/6) = 0.07778, p₂ = exp(−0.46667) ≈ 0.62709. Starting with 20 t/h of the class, the first row receives all 20 and passes about 13.976, the second receives about 6.024 and passes about 2.246, and about 3.778 t/h leaves as oversize. Adding separately calculated row recoveries based on full feed would overstate passage.
The area-and-passage vibrating-screen model couples handbook-based specific capacity to this panel concept. It adjusts specific capacity using nominal separation, oversize and half-size feed proportions, moisture, deck arrangement, open area and other factors. It computes required area from feed undersize and corrected specific capacity, forms a percent-load estimate relative to available deck area, and uses that load or a measured efficiency to infer an effective number of trials. It then evaluates the passage law by row. This allows both a capacity check and a product-size estimate. Agreement with a handbook factor table is distinct from validation against the installed media and operating conditions.
Exercise set
1. Identifiability of trials
A deck has one aperture class and constant open area φ across the active length. From partition observations alone, can the product Nφ be distinguished into a unique pair N, φ?
Solution. The exponent contains Nφ as a product. Known measurements of φ, geometry or independent operating data are required to distinguish the terms. A larger N and smaller φ can give the same prediction.
2. Add a panel
If the class in the two-row example passes through a third row with p₃ = 0.80, calculate added undersize and final oversize.
Solution. About 3.778 t/h enters the third row. It passes 0.20 × 3.778 ≈ 0.756 t/h, leaving about 3.022 t/h oversize. Recovery in row 3 is 3.78% of deck feed for that class.
3. Area versus cut adjustment
A deck is feeding 150 t/h of below-aperture material. The corrected specific capacity is 18 t/(h·m²) and an area safety factor is 1.20. Estimate area and load for 9 m² installed.
Solution. Required area is 1.20 × 150/18 = 10.0 m², giving 111.1% load on this basis. Adjusting a curve midpoint to match the present product does not by itself remove the capacity shortfall.
