[A peak recognition algorithm designed for chromatographic peaks of transformer oil]

Area of Science:

• Analytical Chemistry
• Chromatography
• Spectroscopy

Context:

• Transformer oil analysis is crucial for assessing equipment health.
• Traditional chromatographic peak identification methods have limitations in automation and distortion.
• Dissolved gas analysis (DGA) in transformer oil requires precise peak identification.

Purpose:

• To improve the accuracy and automation of chromatographic peak identification in transformer oil.
• To overcome the limitations of the traditional first-order derivative method, such as low automation and susceptibility to distortion.
• To develop a robust algorithm suitable for online monitoring devices.

Summary:

• An improved method for chromatographic peak identification in transformer oil analysis is presented, combining moving average iteration and normalized analysis techniques.
• This algorithm utilizes multiple iterations of moving average and square wave curves to optimize normalized peak identification parameters, incorporating absolute retention times and peak windows.
• The enhanced method demonstrates accurate peak identification, insensitivity to noise, peak width, and shape variations, making it highly adaptable.

Impact:

• Provides a more automated and reliable method for chromatographic peak identification in transformer oil analysis.
• Enhances the accuracy of dissolved gas analysis in transformer oil, leading to better condition monitoring.
• Offers a robust solution adaptable for on-site, online monitoring devices, improving transformer maintenance and safety.