This study examines protein purification modes at various facilities and evaluates the odor of light sunflower protein concentrates obtained with different purification schemes using an array of gas-based chemical piezoelectric sensors. Nine protein concentrate samples with varying pH, salt composition, added reagent type, and the use of filter or sorption materials were analyzed. Volatile compounds emitted from the protein samples were recorded using sensors coated with chromatographic phases and nanostructured sorbents. During the study, several sensors in the array, the method for feeding volatile compounds into the detection cell, and the sorption time of the volatile protein fraction were replaced. Based on the sensor output data, it was established that the minimum content of volatile compounds detected by the sensors is typical for samples with a pH of 5.6-5.8 and treatment with 5% sodium or calcium chloride. These samples are closest to the accepted standard. Samples treated with alkali exhibited distinct odor profiles and increased overall sensor signals, while protein samples filtered with a Taiko Classic 1G filter exhibited a rapid decrease in volatile compound content when exposed to air. The most effective system for sunflower protein concentrate purification, based on sensory analysis, is treatment with 5% calcium chloride followed by filtration through a Benteco F 620 filter. Therefore, the applied sensor-based protein purification quality assessment system will enable the integration of purification decision-making into digital platforms.