funCyclone: Objective function - Cyclone Simulation: Barth/Muschelknautz

In SPOT: Sequential Parameter Optimization Toolbox

Objective function - Cyclone Simulation: Barth/Muschelknautz

Description

Calculate cyclone collection efficiency. A simple, physics-based optimization problem (potentially bi-objective). See the references [1,2].

Usage

funCyclone(
  x,
  deterministic = c(TRUE, TRUE, TRUE),
  cyclone = list(Da = 1.26, H = 2.5, Dt = 0.42, Ht = 0.65, He = 0.6, Be = 0.2),
  fluid = list(Mu = 1.85e-05, Ve = (50/36)/0.12, lambdag = 1/200, Rhop = 2000, Rhof =
    1.2, Croh = 0.05),
  noiseLevel = list(Vp = 0.1, Rhop = 0.05),
  model = "Barth-Muschelknautz",
  intervals = c(0, 2, 4, 6, 8, 10, 15, 20, 30) * 1e-06,
  delta = c(0, 0.02, 0.03, 0.05, 0.1, 0.3, 0.3, 0.2)
)

Arguments

x	vector of length at least one and up to six, specifying non-default geometrical parameters in [m]: Da, H, Dt, Ht, He, Be
deterministic	binary vector. First element specifies whether volume flow is deterministic or not. Second element specifies whether particle density is deterministic or not. Third element specifies whether particle diameters are deterministic or not. Default: All are deterministic (TRUE).
cyclone	list of a default cyclone's geometrical parameters: fluid$Da, fluid$H, fluid$Dt, fluid$Ht, fluid$He and fluid$Be
fluid	list of default fluid parameters: fluid$Mu, fluid$Vp, fluid$Rhop, fluid$Rhof and fluid$Croh
noiseLevel	list of noise levels for volume flow (noiseLevel$Vp) and particle density (noiseLevel$Rhop), only used if non-deterministic.
model	type of the model (collection efficiency only): either "Barth-Muschelknautz" or "Mothes"
intervals	vector specifying the particle size interval bounds.
delta	vector of densities in each interval (specified by intervals). Should have one element less than the intervals parameter.

Value

returns a function that calculates the fractional efficiency for the specified diameter, see example.

References

[1] Zaefferer, M.; Breiderhoff, B.; Naujoks, B.; Friese, M.; Stork, J.; Fischbach, A.; Flasch, O.; Bartz-Beielstein, T. Tuning Multi-objective Optimization Algorithms for Cyclone Dust Separators Proceedings of the 2014 Conference on Genetic and Evolutionary Computation, ACM, 2014, 1223-1230

[2] Breiderhoff, B.; Bartz-Beielstein, T.; Naujoks, B.; Zaefferer, M.; Fischbach, A.; Flasch, O.; Friese, M.; Mersmann, O.; Stork, J.; Simulation and Optimization of Cyclone Dust Separators Proceedings 23. Workshop Computational Intelligence, 2013, 177-196

Examples

## Call directly
funCyclone(c(1.26,2.5))
## create vectorized target funcion, vectorized, first objective only
## Also: negated, since SPOT always does minimization.
tfunvecF1 <-function(x){-apply(x,1,funCyclone)[1,]}
tfunvecF1(matrix(c(1.26,2.5,1,2),2,2,byrow=TRUE))
## optimize with spot
res <- spot(fun=tfunvecF1,lower=c(1,2),upper=c(2,3),
   control=list(modelControl=list(target="ei"),
   model=buildKriging,optimizer=optimLBFGSB,plots=TRUE)) 
## best found solution ...
res$xbest
## ... and its objective function value
res$ybest

SPOT documentation built on June 26, 2022, 1:06 a.m.