| Safety technology for high-speed interruption includes safety technology and series of electrical equipments.When a short-circuit fault or leakage fault occurs in mine low-voltage power supply system,the short-circuit protection device or leakage protection device can cut off the fault current and prevent energy from being fed to the fault point before the fault current causes the surrounding medium to burn or explode.It is a kind of "all-round" explosion-proof safety and high-speed power-off safety technology for downhole coal mine low-voltage power supply systems.It includes high-speed short circuit protection technology,high-speed leakage protection technology and high-speed power off switch technology.In order to achieve the purpose of "all-round" fire-prevention and explosion-protection,the safety technology for high-speed interruption requires that the accurate identification time of short-circuit faults and leakage faults must be less than 2ms,and the action time of the high-speed power-off switch must be less than 1ms.It can guarantee the safety of the production of coal mine,and also improve the stability of low-voltage power supply systems.With the development of coal mining equipment technology,the stability requirements of coal mine downhole electrical equipment for power supply system are becoming more and more stringent.A more secure,stable and efficient coal mine downhole power supply system can ensure the safety,smooth and high-speed development of coal mine enterprises.At the same time,with the continuous maturity of ground DC transmission and distribution technology,and the supporting equipment is constantly improved,as a result,the DC power system became possible.However,considering the actual situation,in order to realize the full DC distribution,the coal mine needs to undergo a process of the AC-TO-DC conversion.The extensive using of the frequency converters and the consideration of the future application of downhole AC-DC hybrid power supply system makes the Variable-Speed System the most widely used power supply system in coal mine at present or in the future.Although the Variable-Speed System makes coal mine production more efficient and energy-saving,it also brings new problems and challenges to the safety of power supply in coal mine,such as fault protection during the variable frequency of the starting phase of the Variable-Speed System.At the same time,the 127 V power supply system in the coal mine,especially the 127 V electric-coal-drill(electric-rock-drill)system,is the power supply system with the most personal contact in the coal mine and the most prone to gas overrun conditions in the working environment system.Therefore,how to protect the127 V Variable-Speed System from faults is one of the important issues to ensure safety of the power supply in coal mine.This paper takes the safety technology for high-speed interruption as the technical goal,and takes the AC-DC hybrid power supply system in coal mine as the research background,and launches research and discussion on the 127 V Variable-Speed System.First,this paper proposes a DC power supply voltage class sequence by studying the existing ground voltage level sequence and formulating rules,combining the current AC-DC hybrid power supply and distribution level technology and the actual situation of the production equipment of the coal mine downhole power supply system."radiation type power supply" and "two-end power supply" hybrid coal mine downhole AC-DC hybrid power supply network topology.As the solid-state transformer will be used as the execution device of the high-speed power-off switch technology,it is also one of the most important electrical equipment in the construction of the coal mine AC-DC hybrid power supply.For the actual working conditions of the transformer,a solid-state transformer topology based on the modular cascade type is proposed,and the topology structure is analyzed in detail.Secondly,aiming at the problem of high-speed short-circuit protection algorithm of127 V variable-frequency-drive system in underground coal mines,using deep learning technology,a "non-tuning,self-adaptive" high-speed short-circuit protection algorithm based on deep learning is proposed.At the same time,in order to ensure the generalization and applicability of the high-speed short-circuit protection convolutional neural network model algorithm,a transfer-learning algorithm framework based on adversarial transduction is proposed,and the feature learning of the target domain is carried out through the sharing and transfer of the feature learning of the source domain.The reuse of the model enhances the accuracy of the model’s prediction of the target domain and the applicability of the model.Moreover,in order to ensure that the data used for model training has effective feature labels and sufficient quantity,through the transient analysis of the short-circuit fault of the 127 V variable-frequency-drive system in the coal mine,the sample feature quantity is obtained,and the high-speed short-circuit protection convolutional neural network model training is established.Using the data the dataset to train and adjust the parameters of the high-speed short-circuit protection convolutional neural network model.Finally,using the data of the verification set,the verification experiment proves that the time for determining the short-circuit fault(including the variable frequency start-up stage)of this method is less than 1ms,and the accuracy rate is higher than 99%.Electrical safety technology requirements for high-speed short-circuit protection.At the same time,through model performance evaluation and robustness analysis,the feasibility of the method applied to practical devices is proved.Third,using the Bessel function to establish the theoretical model of the common-mode voltage interference conducted by the Variable-Speed System,the mechanism of the failure of the traditional high-speed leakage protection caused by the Variable-Speed System is discussed.At the same time,through detailed research and discussion on the high-speed leakage protection method based on the additional high frequency square wave,the high-speed leakage protection method based on the zero sequence voltage component and the high-speed leakage protection method based on the zero sequence active power component,it is found that whether it is based on the time domain signal or the frequency domain signal,the high-speed leakage protection with domain signal as the criterion will have certain "dead zones",and these "dead zones" all appear when the system is in the starting-state of the frequency converter,which makes the above high-speed leakage protection unable to provides "all-round" protection of the 127 V variable frequency drive system.Applying deep learning technology,combined with the above theoretical analysis,established the sample feature quantity required for training the convolutional neural network model,and established and trained the convolutional neural network model of the high-speed leakage protection and proposed a deep learning-based “no-setting,self-adaption” adapts to the new method of high-speed leakage protection.Moreover,through the optimization of the NIN series model structure likes,the model volume and the response time of the CNN are reduced,and the deployability of the model is enhanced.Experiments have proved that the method for determining leakage faults(including the starting-state of the frequency converter)takes less than 1ms,and the accuracy rate is higher than 99%,which meets the requirements of safety technology for high-speed interruption for high-speed leakage protection.At the same time,the model performance evaluation and robustness analysis prove the feasibility of the method applied in practice.Finally,a safety device for high-speed power-off is designed.For the high-speed fault identification mechanism,an FPGA-based convolutional neural network model deployment platform is designed.At the same time,the two proposed models are optimized accordingly to ensure that the models are in reality.Aiming at the high-speed power-off actuator,a solid-state transformer-based inverter bridge module structure with high-speed shut-off capability is designed,and detailed theoretical analysis and experimental verification have been carried out for this design.It is proved that the longest time to cut off the fault is less than 1ms.The method proposed in this paper is a forward-looking study of the application of artificial intelligence to coal mine downhole relaying.On the basis of ensuring the“reliability,selectivity,sensitivity,and quick-action” of coal mine relaying,it also achieves the good balance the "fourths",On the other hand,this paper analyzes the mechanism of the corresponding failures and explains the high-speed short circuit protection CNN algorithm and the high-speed leakage protection CNN algorithm from the theoretical level,which enhances the interpretability of the CNN model;It provides certain help to ensure the safety of the power supply in coal mine downhole,and also provides new research direction and content for safety technology for high-speed interruption. |
PDF Full Text Request