The Elementary Ontology

The ontology can be referred using IRI http://w3id.org/elementary (you can download it from here). The ontology bridges to Brick, TSO, and OntoCape which are required if SPARQL queries refer to the classes and relationship contained in these.

Ontology of HVAC Stereotypes

Component stereotypes for common HVAC components (for e.g., corresponding to brick:HVAC_Equipment) can be found here. They can be referred to by components in a system description by using the predicate elem:hasStereotyp.

Systems for Evaluation (Modeled Using Brick)

System Kind 1: A simple fresh-air handling unit can be found here.

System Kind 2: A recirculated-air handling unit can be found here.

System Kind 3: Heat-wheel Energy Recovery FAHU can be found here.

System Kind 4: Fresh-air AHU With Parallel Fans can be found here.

SPARQL Queries

1. Use Case - 1: Find the connection point of a component stereotype which manipulates a process variable and dependent variable which will be affected:

PREFIX elem: 
SELECT * WHERE {
    ?cs a elem:Stereotype.
    ?cs elem:hasPhysicalMechanism ?m.
    ?m elem:hasDependentVariable ?dv.
    ?m elem:hasManipulatedVariable ?mv.
    ?cs elem:hasConnectionPoint ?cp.
    ?cp elem:manipulates ?mv.
}

2. Use Case - 2: Between two components that are linked using a generic "feeds" relationship, find the process dependency:

The following query matches the connection port and the variable dependency. For SK1 (FAHU), the two downstream dependencies would be damper->filter (air flow), and valve->heating coil (water flow)

PREFIX elem: 
PREFIX hvac: 
PREFIX brick: 
PREFIX ahu: 
SELECT * WHERE {

    ?c1 a elem:Component.
    ?c2 a elem:Component.
    ?c1 elem:feeds ?c2.
    ?c1 elem:hasMechanism ?m1.
    ?c2 elem:hasMechanism ?m2.
    
    ?c1 elem:hasConnectionPoint ?c1p.
    ?c1p a elem:OutletConnectionPoint.
    ?c2 elem:hasConnectionPoint ?c2p.
    ?c2p a elem:InletConnectionPoint.
    ?c1p elem:dealsWithStuff ?stuff.
    ?c2p elem:dealsWithStuff ?stuff.
    ?c1p elem:suitableFor ?streamKind.
    ?c2p elem:suitableFor ?streamKind.
    
    ?c1p elem:isRelatedToVariable ?c1pv.
    ?c1 elem:hasDependentVariable ?c1pv.
    
    ?c2p elem:isRelatedToVariable ?c2pv.
    ?c2 elem:hasIndependentVariable ?c2pv.
    
    ?c1pv elem:dealsWithStuff ?vstuff.
    ?c1pv elem:hasQuantity ?vq.
    ?c2pv elem:dealsWithStuff ?vstuff.
    ?c2pv elem:hasQuantity ?vq.

}

3. Use Case - 3: Find components which influence a variable quantity at a give point in the stream.

First we need to infer interactions between components:

PREFIX elem: 
PREFIX brick: 
PREFIX ahu: 
PREFIX hvac: 

CONSTRUCT
{
    _:con a elem:Connection.
    ?a elem:connectedThrough _:con.
    ?b elem:connectedThrough _:con.
	_:con elem:connectsSystemThrough ?ap.
    _:con elem:connectsSystemThrough ?bp.
    _:con elem:manipulates ?bpv.
    ?a elem:affects ?b.
	construct{?b elem:affects ?b}
	where{?a elem:hasIndependentVariable ?aipv. ?b elem:affectedUpstreamVariable ?buv.}
	?b elem:affects ?a.
}
WHERE {
	?a elem:feeds ?b.
    
	?a elem:hasConnectionPoint ?ap.
    ?ap a elem:OutletConnectionPoint.
    ?ap elem:dealsWithStuff ?stuff.
    ?ap elem:isRelatedToVariable ?apv.
    ?a elem:hasDependentVariable ?apv.
	
    ?apv elem:hasQuantity ?quantity.
    
    ?b elem:hasConnectionPoint ?bp.
    ?bp a elem:InletConnectionPoint.    
    ?bp elem:dealsWithStuff ?stuff.
    ?bp elem:isRelatedToVariable ?bpv.
    ?b elem:hasIndependentVariable ?bpv.
	
	
	BIND(?reveristrue, IF(?apiv == ?buv
	
    ?bpv elem:hasQuantity ?quantity.
    
    OPTIONAL {
        ?skind a elem:StreamKind.
     	?ap elem:intendedFor ?skind.
        ?bp elem:intendedFor ?skind.
    }
}

Based on this, a test point (TP) can be specified a context and then upstream and downstream components can be located transitively:

PREFIX elem: 
PREFIX hvac: 
PREFIX brick: 
PREFIX ahu: 
SELECT ?c ?cup ?cdw WHERE {
	#Where
    BIND(ahu:conditioned_space as ?context).
    #What
    BIND(brick:Air as ?stuff).
    BIND(brick:Temperature as ?quantity)
    
    #Find the first upstream component of the context which affects ?stuff ?quantity
    ?c a elem:Component.
    ?c elem:feeds? ?context.
    ?c elem:hasDependentVariable ?dvar.
    ?c elem:hasIndependentVariable ?ivar.
    ?dvar elem:hasQuantity ?quantity.
    ?dvar elem:dealsWithStuff ?stuff.
    #For all such components find its upstream
    #components that is linked via process variable
    {
    ?cup a elem:Component.
    ?cup elem:feeds? ?c.
    ?cup elem:hasDependentVariable ?ivar.}
    #But it could also be downstream components
    #which has effect upstream (like a fan), so
    UNION
    {
    ?cdw a elem:Component.
    ?c elem:feeds* ?cdw.
    ?cdw elem:affectedUpstreamVariable ?ivar.
    }
} limit 100

How to get things running

The easiest way to try out the sparql queries is to load the datasets in a graph database which supports OWL-RL reasoning (basically: sub-class, sub-property, and property chains). The RAHU, ERAHU, and PFAHU use FAHU as basis because the supply stream is the same in all.