Oil chromatography data together with its variation trend provides the key basis for the evaluation of the transformer health state. Existing studies on Deep Belief Network (DBN), support vector machine (SVM), et al. have reported a few results in the field of oil chromatography data prediction. However, the above methods are proposed for the classification problem, so there is no theoretical basis for applying the above methods to the time-series prediction. The wrong usage limits the accuracy of the predicted results, which is observed as the obvious “time-shift” error in the prediction curve, leading to the predicted state inconsistent with the actual situation. To fill the gap, a Deep Recurrent Belief Network (DRBN) model for transformer state prediction was proposed based on time series theory and oil chromatography data characteristics. In this model, the construction shortage of DBN in time series prediction was analyzed as a prototype pattern to structure the stereoscopic mapping relation. The self-adaptive delay network based on the principle of autocorrelation, together with the corresponding error feedback network, realized the stereoscopic flow of data in multi-dimensional space and thus ensured the model’s valid expression of time domain correlation. In addition, the cross-entropy loss function, based on Kullback-Leibler divergence and Weibull distribution with an obvious characteristic of DGA, was constructed to eliminate the uncertainty of initialization process and also effectively control the direction and step size of the error gradient. Examples in the field were used to verify the method to find that the model proposed in this paper can availably overcome the “time-shift” error, and its prediction accuracy can reach more than 95.16%.

Dissolved gas analysis can provide powerful technical support for assessment and fault diagnosis of power transformers. However, due to inadaptability of model structure, the performance of diagnosis model can hardly be improved with theoretical support. To fill the gap, a novel self-decision fault diagnosis model for power transformer has been proposed, taking into consideration of both the characteristics of faults and adaptability of conventional deep brief network. Specifically, an active-correction unit, as the core of this novel model, was established to detect the real-time state of the diagnosis model and then select the corresponding correction strategy. An error neuron is also placed to serve as the starting point of the state-oriented correction and load the correction signal that is sent from the active-correction unit. Verifications in the field indicates that the fault correlation extracted by the sparse autoencoder can eliminate error in the training process. When the results of IEC three-ratios method is integrated into the input neuron, the accuracy of proposed model can reach more than 92.3%. Based on the proven effectiveness and feasibility of the proposed method in state-oriented correction, this method can provide a reliable support for substation preventive maintenance.

The present research focuses on the transient stability of power systems embedded with a high penetration of converter-interfaced generation (CIG). It is investigated that under large disturbances, power systems with CIGs may exhibit the characteristic of multiple stable equilibrium points (SEPs). This characteristic and its associated stability potential have not been thoroughly studied in conventional power systems, primarily due to the prevalent assumption of a single SEP constrained by the physical limitations of the large-capacity synchronous generator. Since the critical clearing time (CCT) is no longer able to determine system transient stability with multiple SEPs, the absolute instability boundary (AIB) condition is proposed. It is found that if the trajectory surpasses the AIB, the power system will be unable to maintain stability after large disturbances. In what follows, an energy function is introduced to assess transient stability considering multiple SEPs. At the end of the study, the impact of CIG's frequency limiters on transient stability is studied, and a transient stability enhancement method is derived based on the AIB condition and safety constraints. This provides a basis for CIGs with millisecond-level dynamics to respond to power systems with near-second dynamics. Numerical simulations demonstrate the effectiveness of the proposed theoretical results.

As a widely configured controller, the limiter is essential to ensure the security of the voltage source converter (VSC), which affects the stability characteristics of the grid-connected VSC system under severe faults. However, due to limiters' hard algebraic constraints, the existing methods have yet to reveal the stability mechanism related to the multi-limiters thoroughly. In this paper, the stability mechanism of the grid-connected VSC system with multi-limiters is investigated, and an improved Filippov Theorem is proposed to study the dynamic characteristics. Two kinds of points are essential to the stability, namely, the fixed switching point(FSP) and the fixed ending point(FEP). Expressly, transient stability can be guaranteed by the condition that FSP is in the region of attraction(ROA), even if other parts of the trajectory are outside the ROA. Moreover, the existence of FEP contributes to some control loop failure, which does not induce instability but may cause the power into the grid to be inconsistent with the reference value. This paper presents the analytical methods for these crucial points, which avoid time-domain simulation. What follows illustrates how to enhance the stability characteristics by configuring the FSP and FEP. Another advantage of the proposed method is easily extended to the VSC with multi-limiters and different control structures.

Transformer is the key equipment of the power system. The reliability of power system is reduced by transformer faults. The device, which monitors and predicts the operational status of the transformer, has important implication for the safe operation of the power system. Currently, most of the relevant work is based on the dissolved gas analysis (DGA), but the raw data usually has short-period fluctuations, which poses great difficulties for the accuracy of prediction model. Transformer fault is a comprehensive performance of the historical characteristics and abrupt characteristics. Model building of existing methods usually leads to lasting dependence on data quality. In this paper, state prediction of power transformer based on Grey-Lagrange method with weighted coefficient of variation was proposed. Firstly, the characteristics of the faults development are taken into account comprehensively. The long-term and short-term data sets are prepared to build the prediction model, which is conducive to improving the accuracy of the model. Secondly, the coefficient of variation matrices was calculated and used as the input characteristics to train the Grey-Lagrange model. Finally, the models which have been trained were merged by genetic algorithm (GA). In summary, the method, proposed in this paper, can recognize the change in detail and predict the trend of gas content in oil, which can provide assistance for condition maintenance. The results of the prediction model offer the possibility to estimate the states of the equipment, especially for the situation of weak insulation degradation and abrupt faults.

Recent studies indicate that transient instability in power systems with converter interfaced generation (CIG) is primarily related to unstable periodic orbit (UPO) rather than unstable equilibrium point (UEP). This phenomenon is noticeable in weak grids, emphasizing the importance of studying the evolution of the stability boundary's dominant, from UEP to UPO. To address this, the present study introduces a novel transient stability analysis framework for power systems, which relies on the energy-angle complex vector. Both the identification of transition point of the stability boundary and the stability margin assessment are conducted under this framework. It is revealed that variations in parameter conditions, such as an increased proportion of CIG, a decrease in system strength, and inadequate control parameters for CIG, may lead to the dominance of UPO in shaping the stability boundary. This may lead to the ineffectiveness and non-conservatism of the UEP-based criteria. Simulation results demonstrate the effectiveness of the derived conclusions. With the increasing integration scale of CIG, the growing importance of UPO-based stability characteristic will become pronounced in future power systems.