This paper presents a drilling and coring device for the lunar exploration, which is possibly utilized to acquire the lunar regolith with a certain depth. The drilling device is composed of three components: rotary unit, percussive unit and penetrating unit. The rotary-percussion drill can work in two different operating modes: rotary mode and rotary-percussive mode, depending on the properties of cut object. In the relatively loose regolith, rotation and penetration can make the drill work in a well state. However, once rock is encountered in the drilling process, besides rotation and penetration, percussion must be launched to reduce the drilling power and the required penetrating force. Due to the indetermination of the lunar environment, it is not easy to control the coring drill to adapt to the encountered conditions. To obtain a high coring ratio with relatively low power, an intelligent drilling strategy is inevitably proposed to accomplish the drilling process control. Considering the lunar soil simulant should cover the possible composition of real lunar soil, simulant are classified into several levels based on the generalized drillability. For each level of drillability of lunar soil simulant, experiments are conducted to get the characteristics in frequency-domain of rotary torque output. The sampled characteristics of rotary torque output are utilized to train the object-recognition system based on Support Vector Machine (SVM). Information in all the levels of drillability of lunar soil simulant is stored in the object-recognition system as an expert system. To understand the properties of the drilling object, rotary torque is selected to identify the level of drillability of simulant in drilling process. Subsequently, once the level is obtained, drilling strategy is adjusted to adapt to the current level correspondingly in real time. Experiments are conducted to verify the intelligent drilling strategy successfully.

This content is only available via PDF.
You do not currently have access to this content.