PSI - Issue 77

Alexander Backa et al. / Procedia Structural Integrity 77 (2026) 143–151 A. Backa et al. / Structural Integrity Procedia 00 (2026) 000 – 000

148

6

Table 2. Measured average values of temperatures, PM particles and ratios.

Plane

PM [mg/m 3 ] 21.33±9.53 17.60±12.56 19.05±11.86 23.45±9.14 19.63±8.01 20.78±25.69 16.96±10.58 16.41±7.41 26.09±43.22 17.35±12.22 17.85±10.40 16.32±8.89 16.49±8.11 19.77±14.82 15.77±12.63

Temperature [°C]

Temperature-to-PM ratio

PM-to-Temperature ratio

XY Plane Average

472±148 563±191 559±189 429±100 463±120 599±166 609±142 503±109 785±88 585±173 539±151 498±107 519±105 577±210 643±147

37.00 64.88 55.11 31.43 37.73 87.96 60.70 46.81

0.0294 0.0291 0.0312 0.0285 0.0286 0.0322 0.0161 0.0196 0.0510 0.0233 0.0248 0.0238 0.0206 0.0346 0.0205

XY1 Average XY2 Average XY3 Average XY4 Average YZ1 Average YZ2 Average YZ3 Average YZ4 Average XZ1 Average XZ2 Average XZ3 Average XZ4 Average Average of all quarters

YZ Plane Average

183.80 103.88 121.40

XZ Plane Average

47.48 48.03

258.93 121.03

19.99±16.76

537±161

82.12

0.0295

To derive a predictive model for PM concentration based on temperature, a linear regression analysis was performed using data from Table 2. Prior to regression, outliers were identified and removed to ensure the robustness of the model. The resulting linear regression equation (Equation 3), with determined regression coefficients, effectively captures 96% of the measured data, demonstrating a strong correlation between temperature ( t , °C) and PM concentration ( PM , mg/m 3 ). = −0.0559 ∙ + 45.947 (3) While the initial regression, including all data points, yielded an R-squared of 0.003, the outlier-corrected model significantly improved the R-squared value to 0.96 with a standard error of 4.85 (Fig. 4).

100 120 140

0 20 40 60 80 PM [mg/m 3 ]

300

400

500

600

700

800

900

Temperature [°C]

Outliers PM [mg/m³]

PM [mg/m³]

Linear Trend Line

Fig. 4. Linear correlation plot of particulate matter vs. temperature.