Examensarbete komfortlab : Undersökning av värmelaster med två-nivåers faktorförsök

Sammanfattning: The goal of this study was to investigate future controls of heating loads in a comfort laboratory. The laboratory room is used to simulate how the room comfort will change by using one or more cooling beams. External walls can be used to test different room sizes, and heat effects can be provided from the floor and one wall. The study was conducted according to a two-level factor design and the five factors used were floor heating, north wall, west wall, fan and facade wall. In a full two-level factor test plan, there are 32 tests for five factors that needed to be done. Instead of making 32 tests, the study continued with a fractional two-level factor design, which meant that only 16 tests were performed. The response variable for the tests was the so called "losses" which may be positive or negative and is a percentage of the floor effect. Via "losses" the average temperature of the measuring room could be 24.5 ˚C, which was the temperature chosen for this survey. Before the results for each test were taken, requirements for "steady state condition" were met. The requirements are taken from an ISO standard that the company uses, and it provides reliability of the results. The analyzes when having "losses" as a response variable also pointed out the floor as the most influential factor. However, there were different results regarding the scheme on the other factors. It is due to different coding of the factors, and "losses" was used in the units % or watts. Some disturbances were detected in the results. One is that "losses" has a connection to the floor heating, which may be the reason it is most significant. The two levels of the floor heating did not match because "losses" changed the levels. For the next two-level factor design experiment, it is recommended that the effect of the floor is constant because of its connection to "losses". The response variable can therefore be "losses". The constant on the floor heating can be changed when the factor plan is repeated several times to see how "losses" varies with different floor effects.