A PRIMER ON THE USE OF INFLUENCE COEFFICIENTS IN BUILDING SIMULATION

摘要

In order to create a simulation model of arnbuilding, it is usually necessary to make arnnumber of assumptions and/or approximationsrnabout the building being simulated. Manyrnphysical quantities cannot be known preciselyrnwhen the building is being simulated. Forrnexample, the real amount of infiltration isrnalmost never known because it is impossible tornpredict accurately and difficult to measure.rnOther examples include quantity and type ofrninternal mass, thermophysical properties ofrnbuilding materials, ground temperatures, andrnequipment efficiencies.rnIn addition, occupancy often has a significant,rnbut unpredictable influence on the buildingrnenergy consumption. Such behavior asrnopening doors, leaving windows open,rnchanging the thermostat settings, leavingrnshades open or closed and generating heatrndue to use of appliances all have an impact onrnbuilding energy consumption.rnIn order to gage the accuracy of a simulation,rnit is necessary to estimate the relevantrnsignificance of the assumptions made. Byrndetermining which assumptions havernsignificant impact on the building energyrnconsumption, it is possible to determine whererneffort should be made to refine the simulation.rnIn addition, potential errors can be estimated.rnOne method of determining the significance ofrnthe assumptions made when simulating arnbuilding involves the use of influencerncoefficients. An influence coefficient is thernpartial derivative of a simulation result withrnrespect to a parameter. An example would bernthe partial derivative of total building energyrnconsumption with respect to the effective solarrntransmissivity of a window shade. Thernsimulation result could be any result of interestrnto the user, such as the total energyrnconsumption, heating or cooling loads, annualrnenergy costs, etc. The parameter could be anyrnassumption that affects the simulation result.rnThis paper describes the use of influencerncoefficients to estimate the significance ofrnassumptions made in the building simulationrnprocess. A method for calculating andrnnondimensionalizing influence coefficients isrnpresented. Examples are taken from a studyrncomparing building energy performance ofrnmanufactured family housing units tornconventionally -built family housing units at FortrnIrwin, CA. The building simulation tool is thernBuilding Loads Analysis and SystemrnThermodynamics (BLAST) program.