MODELLING RATE OF PENETRATION AND WEAR OF PDC BITS

Development of mathematical models for calculation of the Rate of Penetration (ROP) including prediction of bit wear is typically fulfilled by collection and processing data obtained during numerous laboratory experiments. Such models cover influence of a small set of geological factors, some of which cannot be measured under bottomhole conditions or require additional expensive operations. Besides that, cost of PDC bits is considerably higher than of roller-cone ones necessitating the search for more efficient solutions in order to model their behavior.Abovementioned reasons lead to that performance of traditional mathematical models for ROP prediction significantly diminishes while testing under field conditions. The latter, in turn, prevents their widespread use in practice for tackling a number of problems such as optimization of the drilling regime parameters, selection of the best of competing bits, estimation of the recommended bit operating time, determination of the bit lift moment for further replacement, and etc.In the current article, an alternative approach is considered based on the neural networks and embeddings which allows modelling a PDC bit condition and the impact of geological factors through hidden features. The distinctive characteristics in this case are the high adaptability of the model to specific well drilling conditions, as well as the ability to take into account an unlimited number of parameters that affect the process.Reliability of the research results was estimated both qualitatively and quantitatively. The nature of the calculated ROP change corresponds to the actual one. The statistical indicators of the mathematical model performance based on the trained neural network also demonstrate a high convergence of the calculated data with the actual ones. In particular, the determination coefficient exceeds 0.80.Based on the results of the conducted research, it was possible to establish previously unknown features of PDC bits; to develop new equations describing the relationship of ROP as well as a bit wear on its progress while drilling under constant geological and technological conditions.