B4C defaults to a 0.9 TD fraction

[keckler]

This isn't really a Problem, but it is unexpected, inconsistent, and potentially troublesome.

The B4C class has a built-in non-unity theoretical density fraction that is applied by default when the material is constructed:

armi/armi/materials/b4c.py

Lines 117 to 145 in 37f5479

	def setDefaultMassFracs(self) -> None:
	r"""B4C mass fractions. Using Natural B4C. 19.9% B-10/ 80.1% B-11
	Boron: 10.811 g/mol
	Carbon: 12.0107 g/mol
	4 moles of boron/1 mole of carbon
	grams of boron-10 = 10.01 g/mol* 4 mol * 0.199 = 7.96796 g
	grams of boron-11 = 11.01 g/mol* 4 mol * 0.801 = 35.27604 g
	grams of carbon= 12.0107 g/mol * 1 mol = 12.0107 g
	total=55.2547 g.
	Mass fractions are computed from this.
	"""
	massEnrich = self.getMassEnrichmentFromNumEnrich(naturalB10NumberFraction=0.199)
	gBoron10, gBoron11, gCarbon = self.setNewMassFracsFromMassEnrich(
	massEnrichment=massEnrich
	)
	self.setMassFrac("B10", gBoron10)
	self.setMassFrac("B11", gBoron11)
	self.setMassFrac("C", gCarbon)
	self.p.refDens = DEFAULT_MASS_DENSITY
	# TD reference : Dunner, Heuvel, "Absorber Materials for control rod systems of fast breeder reactors"
	# Journal of nuclear materials, 124, 185-194, (1984)."
	self.p.theoreticalDensityFrac = (
	DEFAULT_THEORETICAL_DENSITY_FRAC # normally is around 0.88-93.
	)

This non-unity fraction is used whenever the material density is queried:

armi/armi/materials/b4c.py

Lines 160 to 178 in 37f5479

	def density(self, Tk: float = None, Tc: float = None) -> float:
	"""
	Return density that preserves mass when thermally expanded in 2D.
	Notes
	-----
	- applies theoretical density of B4C to parent method
	"""
	return material.Material.density(self, Tk, Tc) * self.p.theoreticalDensityFrac
	def density3(self, Tk: float = None, Tc: float = None) -> float:
	"""
	Return density that preserves mass when thermally expanded in 3D.
	Notes
	-----
	- applies theoretical density of B4C to parent method
	"""
	return material.Material.density3(self, Tk, Tc) * self.p.theoreticalDensityFrac

There is also the ability to specify the theoretical density fraction using the TD_frac material modification, which is a good thing:

armi/armi/materials/b4c.py

Lines 31 to 52 in 37f5479

	def applyInputParams(
	self, B10_wt_frac=None, theoretical_density=None, TD_frac=None, *args, **kwargs
	):
	if B10_wt_frac is not None:
	# we can't just use the generic enrichment adjustment here because the
	# carbon has to change with enrich.
	self.adjustMassEnrichment(B10_wt_frac)
	if theoretical_density is not None:
	runLog.warning(
	"The 'threoretical_density' material modification for B4C will be "
	"deprecated. Update your inputs to use 'TD_frac' instead.",
	single=True,
	)
	if TD_frac is not None:
	runLog.warning(
	"Both 'theoretical_density' and 'TD_frac' are specified "
	f"for {self}. 'TD_frac' will be used."
	)
	else:
	self.updateTD(theoretical_density)
	if TD_frac is not None:
	self.updateTD(TD_frac)

The bad part about that is that all other material classes return a fully-dense material by default. If you want a version of the material with lower density, generally you should use the associated material modification TD_frac. But for B4C, you actually need to use TD_frac to get a full-density material. So this means that the B4C class is inconsistent with our other materials, but it is not documented as such.

I did want to change this as part of #1114 , but since it would be an API-breaking change and it proved a bit more challenging than expected, I decided to just report it here for future work instead. B4C, being as ubiquitous as it is, is used allllllll over our repos, and this could be tricky to implement a change.

[ntouran]

B4C is a ceramic and ceramics are always manufactured at less than their max theoretical density. B4C in particular is typically around 90%. This is a lot different from typical metals and coolants and stuff. I still feel that it's reasonable to use a typical B4C TD as the default. I think the big concern otherwise is that if you default to a physically impossible full TD B4C beginners may think their rod worth is a lot higher than it actually is.

I do agree that it should be more clearly documented.

[john-science]

Based on Nick's comment above, I would prefer to leave B4C as-is, and add a comment/docstring explaining the matter.

@keckler Thoughts?

[keckler]

Totally fine with the solution being better documentation. Let's go with that.

I will point out, however, that we don't do the same for other ceramics which suffer from the same physical reality that they are difficult to produce at 100% TD (UO2, ThO2, MOX). So I still view this as inconsistent and somewhat confusing.

I'll go put a comment in the docstring!