With the vigorous development of the construction of new power systems, a large number of distributed power sources have been integrated into the distribution network, and the distribution network has become an important part of the power system, making the situational awareness of power distribution equipment more and more urgent. The load switch can interrupt the load current and short-circuit current, and is an important equipment in the medium voltage ring main unit. Once the mechanical mechanism fails, it will cause the switch to refuse, which will have a great impact on the distribution network scheduling. Therefore, the on-line detection of the mechanical status of the load switch is conducive to ensuring the reliability of the switch action and improving the power supply safety of the distribution network.
Statistics show that poor mechanical characteristics are the main cause of switch failures, among which the failure of operating mechanism accounts for the largest proportion. Commonly used detection methods include vibration signal analysis method, switch stroke detection method, infrared temperature measurement and motor current signal analysis method. Most of the studies on isolation switches and circuit breakers have been studied in the existing literature, and relatively few studies have been conducted on load switches.
The load switch is a momentary electrical equipment, and when receiving the opening and closing instructions, the energy storage motor is energized to drive the transmission gear to rotate, thereby driving the connecting rod mechanism to realize compression spring energy storage. The load switch may be jammed during the operation, and the current form of the energy storage motor can effectively reflect the mechanical state of the switch. Therefore, researchers such as Huang Haiyan and Gao Wei from the School of Electrical Engineering and Automation of Fuzhou University and Longyan Power Supply Company of State Grid Fujian Electric Power Co., Ltd. proposed a load switch jamming fault detection method based on improved dynamic time regularization.
Figure 1 Troubleshooting process for load switch jamming
Firstly, the researchers used the sliding mean filter to process the motor current signal in real time and filter out the interference signal. Secondly, the recording start and stop criteria of the action current are formulated to ensure that the complete motor action current is recorded. Subsequently, the dynamic time regularization (DTW) was improved by distance formula adjustment, algorithm acceleration and storage space optimization. Taking the current signal in the normal state of the switch as the standard waveform, the IDTW is used to calculate the standardized distance between the comparison waveform and the standard waveform, and the boundary threshold is formulated to realize the identification of the switching state. Finally, a set of online diagnostic terminals is designed to realize the engineering of the proposed algorithm.
Fig.2 Field experiments
The researchers further stated that the method needs to store a set of normal motor current waveforms, use IDTW to calculate the normalized distance between the comparison waveform and the normal motor current waveform, and set a boundary threshold to identify the switch jamming fault state. The proposed improvement measures speed up the execution speed of the algorithm, reduce the storage space required by the algorithm, and solve the problems of large memory space and long calculation time of the original algorithm.
Figure 3 Online diagnosis terminal
Through experiments, they found that the proposed algorithm can start accurately for the motor current signals of two different types of load switches, and the identification accuracy of the simulated jamming faults reaches 99% and 99.37% respectively, which fully shows that the proposed method has strong adaptability, and the designed online diagnostic terminal can accurately and quickly detect the jamming faults of different types of switches.
The results of this work were published in the 6th issue of Electrical Technology in 2024, with the title of "Identification of Stuck Faults of DC Motor Drive Load Switches Based on Improved Dynamic Time Regularization", and the authors are Haiyan Huang and Wei Gao.