A data mining approach to finding relationships between reservoir properties and oil production for CHOPS

Geovisualization for Association Rule Mining in Oil and Gas Well Data

“…To analyse the main features of production performance, a polynomial fitting curve of the daily production data that denoted production performance of the well was created. Three production parameters were used to characterize the fitting curve: peak value, effective life cycle and effective yield [3], as shown in Figure 6. The source data of the 118 studied CHOPS wells in the case study were collected and saved in the database [19].…”

“…The reservoir property parameters used for association rule mining were six basic petro-physical parameters of a reservoir that can be used to identify to describe characteristics of a reservoir: cumulative porosity, cumulative pore volume, cumulative shale content, cumulative oil saturation, cumulative fluid mobility factor, and effective thickness [3]. To analyse the main features of production performance, a polynomial fitting curve of the daily production data that denoted production performance of the well was created.…”

“…Recently various data mining methods have been successfully utilized to help understand the reservoirs and enhance oil recovery [1][2][3][4]. Association rule mining is one of the most popular data mining methods and was first proposed for extracting interesting correlations, frequent patterns, associations or casual structures among the item sets from transaction data or other data repositories [5].…”

“…Oil and gas well data generally includes well locations and other non-spatial attributes, such as reservoir property parameters (e.g., pore volume, porosity, permeability) and production performance parameters (e.g., peak value and effective yield) [3]. After the values of the non-spatial parameters are transformed into a set of sub-ranges, association rule mining algorithms such as Apriori [6] can be applied to the discretized data.…”

“…After the values of the non-spatial parameters are transformed into a set of sub-ranges, association rule mining algorithms such as Apriori [6] can be applied to the discretized data. The quantitative relationships between reservoir properties influencing oil recovery and oil production performance can then be discovered [3]. However, there are still some challenges in the understanding and identification of the interesting association rules from oil and gas well data.…”