LOGO ANALYSIS OF INSULATION OF MATERIAL Wang Deyu, Li Dejun Zhong Haoyuan Xu Shanshan Li Yaqiong, Yan Li 1
Contents Executive Summary 1 Literature Review 2 Data analysis & Conclusion 3Choice of Experimental Design 4 Performing the Experiment 5 Eliminating Noise 6 2
Executive Summary 3
Executive Summary(1/2) Definition The experiment is aimed to compare the performance of different kinds of heat insulation materials under normal conditions. The results of the experiment would be quantified into the details including the texture, thickness, exterior color and ventilation. 4
Executive Summary(2/3) Cause and Effects 5
Executive Summary(3/3) Regression Model 6
Literature Review 7
Literature Review(1/2) Define how various factors would accelerate or decelerate the cooling rate. 《 Fabric Selection for a Liquid Cooling Garment 》 《水压机泵站工作液体降温问题分析》 8
Literature Review(2/2) △ T=f(LP, S, M, C, T, HC, CA) 9 properties of the liquid size of the container Heat conduction property of container physical properties of insulation material contact of the air color of the material thickness of the material
10 Choice of Experimental Design
Preparation & Location Preparation Material Container: Beaker Kerosene thermometer Experiment Location C Builiding, Room 306. The room temperature is 26 centigrade. 11 Flax, Black Flax, White Cotton, Black Cotton, White Beaker,1 50ml
Variables & Blocking 12 Variable Selection Material, Color, Layer, and Ventilation Setting Variables Blocking Two thermometers which have different calibration FactorMaterialColorLayerVentilation +HeavyBlackMultipleYes -LightWhiteSingularNo 4 variables, 2 levels per variable, 2 replications per treatment, 2 blocks, Full factorial
13 Performing Experiment
14 We boil tap water to approximately 100 degrees Celsius, and then quickly pour 200 ml boiling water into the two beakers and read the temperature. We begin reading when we first see the temperature is steady and begins to drop. We count 3 minutes before a second reading. Using the two readings with 3-minute interval, the drop of temperature within the 3 minutes could be calculated.
15 Eliminating Noise
Warm up of the beakers and the thermometers To ensure that the heat won’t lose through other channels. Wrap the cloth tightly to the beaker Use slim clip to ensure that the least width is overlapped. Pad the cup with a paper dish underneath To minimize the heat conducted through the bottom. 16
Data analysis & Conclusion 17
Initial Model (1/2) 18 No.FactorsNo.Factors 1Material9Color* ventilation 2Color10Layer*Ventilation 3Layer11Material*Color*Layer 4Ventilation12Material*Color*Ventila tion 5Material*Color13Material*Layer*Ventil ation 6Material*Layer14Color*Layer*Ventilati on, 7Material*Ventil ation 15Material*Color*Layer* Ventilation 8Color*Layer Use these 15 factors in a GLM and calculate coefficients in Minitab.
Initial Model (2/2) 来源 自由度 Seq SS Adj SS Adj MS FP Material Color Layer Ventilation Material*Color Material*Layer Material*Ventilation Color*Layer Color*Ventilation Layer*Ventilation Material*Color*Layer Material*Color*Ventilation Material*Layer*Ventilation Color*Layer*Ventilation Material*Color*Layer*Ventilation 误差 合计
Modified Model (1/3) Then we delete the following factors and recalculate one by one. Material*Color*Layer*Ventilation Material*Color*Layer Material* Layer*Ventilation Color*Layer*Ventilation Material*Color*Layer Color*Layer Material*Layer 20
Modified Model (2/3) Also we get S = R-Sq = 96.32% R-Sq (调整) = 95.04% 项 系数 系数标准误 T P 常量 Material Light Color White Layer Singular Ventilation No Material*Color Light White Material*Ventilation Light No Color*Ventilation White No Layer*Ventilation Singular No
Modified Model (3/3) 22
Modified Model (3/3) 23
Modified Model (3/3) 24
Modified Model (3/3) 25
拟合因子 : TempDiff 与 Material, Color, Layer, Ventilation TempDiff 的效应和系数的估计(已编码单位) 项 效应 系数 系数标准误 T P 常量 Material Color Layer Ventilation Material*Colo r Material*Ventilation Color*Ventilation Layer*Ventilation S = PRESS = R-Sq = 98.71% R-Sq (预测) = 97.38% R-Sq (调整) = 98.23% 26
拟合因子 : TempDiff 与 Material, Color, Layer, Ventilation 对于 TempDiff 方差分析(已编码单位) 来源 自由度 Seq SS Adj SS Adj MS F P 主效应 2 因子交互作用 残差误差 失拟 纯误差 合计
拟合因子 : TempDiff 与 Material, Color, Layer, Ventilation TempDiff 的系数估计,使用未编码单位的数据 项 系数 常量 Material Color Layer Ventilation Material*Color Material*Ventilation Color*Ventilation Layer*Ventilation
29 拟合因子 : TempDiff 与 Material, Color, Layer, Ventilation
30 拟合因子 : TempDiff 与 Material, Color, Layer, Ventilation
31 拟合因子 : TempDiff 与 Material, Color, Layer, Ventilation
Regression Model (1/4) 32 The residuals fits well in a normal distribution, and the main effects and all the 4 interactions are significant. Thus,
Regression Model (2/4) 33
Regression Model (3/4) 34
Regression Model (4/4) We transform TempDiff into Exponential form, and get the residual plot as below: We see some obvious patterns, we don’t recommend to transform the data in this way. 35
Results explanations No ventilation can remarkably maintain the high level of heat preservation. Materials have main effect of heat preservation as well. Colors of material have main effect of heat preservation as well. Thickness of material has less but also main effect of heat preservation as well. Interaction explanation. 36
Possible causes Ventilation-absence condition has the best ability of maintaining heat may result in that heat is lost mostly from the top of the cup, more than from the wall of cup. White color surprisingly has better ability of maintaining heat. Heavy cloth has better heat maintaining ability, which corresponds to our intuition. However, layers have less effect. 37
Error sources Inequity of preliminary heating results the different original conditions of materials such as cloth and the cups. Two thermometers have different abilities of measuring such as sensitivity to temperature changes and measurement resolution. System errors from two experimenters reading the thermometers such as view angular. Water incrustation or impurities in later treatments because of repetitive uses. Impurities in water may affect the temperature decrease rates. Room temperature may change during the relatively long period time during the experiment process. 38
Q & A 39
LOGO 40