Abstract: The determination of liquid saturated vapor pressure is a key component of undergraduate physical chemistry laboratory instruction, facilitating students’ systematic understanding of material properties and phase-transition behavior. However, the conventional dynamic method suffers from cumbersome operation, substantial experimental error, limited openness, and safety concerns. The proposed digital laboratory scheme integrates a proportional-integral-derivative (PID)-based automated control system with a web-based virtual laboratory platform and applies machine learning for in-depth analysis of experimental data, thereby providing a safe, efficient, and precise instructional environment. Empirical results show that the mean relative measurement error is reduced from the traditional 8.7% to 0.5%, the instructional cycle is shortened from more than one week to within a week, the temperature measurement error is controlled within ±0.5 °C, and the pressure error within ±0.2 kPa. This approach optimizes conventional teaching practices and effectively remedies deficiencies in current experimental instruction.

Key words: Saturated vapor pressure, Digitalization, Automatic control, Web laboratory, Machine learning