As the core means of transformer insulation status monitoring, oil dissolved gas chromatography analysis technology can achieve early warning of latent faults inside the equipment by accurately detecting the concentration changes of dissolved gas components in oil samples. This technology is based on the principle of gas chromatography separation. The implementation process includes: after extracting oil samples from the transformer body, degassing of dissolved gases is completed in a high vacuum environment, and quantitative analysis of each component is performed by the detector after separation by chromatographic column. According to GB/T 7252-2001 "Guidelines for Analysis and Judgment of Dissolved Gases in Transformer Oil", transformers with voltage levels of 220kV and below need to implement a periodic monitoring system, among which the key characteristic gas judgment thresholds are: hydrogen (H₂) reaches 150μL/L (may indicate low-energy discharge or electrolysis of water in oil), acetylene (C₂H₂) exceeds 5μL/L (a trace amount can represent high-temperature arc discharge of >800℃), and total hydrocarbons (CH₄+C₂H₆+C₂H₄+C₂H₂) exceed 150μL/L (indicating medium- and high-temperature overheating faults).
The three-ratio method (Rogers coding method) is the core diagnostic tool, which realizes fault type judgment by calculating the three groups of ratio coding combinations of CH₄/H₂, C₂H₂/C₂H₄, and C₂H₄/C₂H₆. Typical coding correspondences include: "0-3-2" code points to severe overheating faults > 700℃, and "1-0-2" code clearly indicates high-energy discharge faults. In the case of excessive gas concentration, a step-by-step treatment plan is required: use vacuum oil filtration equipment with a processing capacity of ≥5T/h, strictly control the oil temperature at 60±5℃, the ultimate vacuum degree <133Pa, and continue purification for ≥48 hours. Practice shows that after 48 hours of vacuum degassing treatment, the acetylene content of a 110kV transformer dropped from the excessive value to the qualified range. The supporting unidirectional circulation process (transformer → oil storage tank → oil filter → transformer) can construct a closed-loop purification path to achieve deep removal of dissolved gases. After the treatment is completed, it needs to be left to stand for 24 hours for retesting, and partial discharge test and insulation resistance test should be carried out simultaneously for double verification.
Equipment returned to the factory for overhaul must meet any of the following criteria: acetylene concentration > 10μL/L or total hydrocarbon > 300μL/L and the three-ratio method diagnoses high-energy discharge (such as "1-0-2" code); internal discharge traces found during disassembly inspection (typical case: a 220kV transformer was returned to the factory for overhaul due to hydrogen concentration reaching 800μL/L, and the bushing end screen discharge fault was finally confirmed). The preventive maintenance system includes: the first chromatographic analysis of newly commissioned equipment must be carried out after 48 hours of quiescence; 220kV transformers implement a half-year inspection cycle, and 110kV equipment implements an annual inspection cycle; focus on strengthening the maintenance of the oil pillow sealing system (the case of Liyuan Hydropower Station confirmed that sealing failure was the main cause of gas exceeding the standard).
The partial discharge detection technology system covers two matrices: electrical measurement method and non-electrical measurement method: electrical measurement method includes pulse current method (IEC 60270 standard, 10pC-level sensitivity detection is achieved through Rogowski coil), ultra-high frequency detection method (UHF, deploying sensors in the flange gap of the box to capture 300MHz-3GHz electromagnetic waves), transient earth voltage method (TEV, detecting nanosecond voltage pulses on the outer wall of the box); non-electrical measurement method includes ultrasonic positioning method (40-200kHz sensor array to achieve ±10cm-level time difference positioning), optical measurement method (built-in optical fiber sensor to capture discharge light radiation), chemical detection method (linked with oil chromatography, when H₂＞150μL/L and trace C₂H₂ is detected, it indicates discharge risk).
Intelligent diagnosis technology includes multi-source information fusion, deep learning application and digital twin system: a 750kV substation successfully identified 0.5mm-level micro-gap discharge inside the bushing through UHF+ultrasound+oil chromatography joint monitoring; the deep learning model based on the ResNet network achieved a discharge pattern recognition accuracy of 96%; the digital twin system constructed a three-dimensional electromagnetic-thermal-mechanical coupling model to achieve dynamic prediction of the discharge development process. In an application example of a converter station in the Southern Power Grid in 2023, oil chromatography monitoring found that the C₂H₂ concentration reached 8.2μL/L, and synchronous UHF detection captured typical discharge signals. Ultrasonic positioning accurately locked the fault point of the B-phase bushing base. Disassembly inspection confirmed that the loosening of the grading ring bolt caused suspended discharge. After repair, the discharge of the equipment was significantly reduced from 3500pC to 15pC.
The research conclusion pointed out that modern transformer fault diagnosis should build a three-dimensional monitoring system of "oil chromatography initial screening-multi-technology positioning-intelligent evaluation". Based on oil chromatography analysis as a basic monitoring method, combined with UHF/ultrasonic positioning technologies, accurate diagnosis of discharge and overheating faults can be achieved. Future development should focus on the construction of digital twin systems coupled with multiple physical fields, and comprehensively improve the intelligent level of power grid equipment operation and maintenance through dynamic insulation status assessment and life prediction.