Application of ARIMA-RTS optimal smoothing algorithm in gas well production prediction

• DOI: 10.1016/j.petlm.2021.09.001
• Publication Date: 2021-09-23T16:13:07Z
• Authors: Yonggang Duan; Huan Wang; Mingqiang Wei; Linjiang Tan; Tao Yue
• Keywords: Smoothing; Moving average

Gas field production forecast is an important basis for decision-making in the gas industry. How to accurately predict dynamic during development content of reservoir engineering research. Reservoir numerical simulation most common method predicting oil and production. However, it requires a lot data build accurate geological model which tedious time-consuming. At present, many scholars have used machine learning mining methods production, but they not considered whether use increasing measures will affect predicted results. Thus, ARIMA-RTS optimal smooth algorithm first applied establish prediction well According historical data, processed, differential autoregressive integral moving average (ARIMA) time series established, then ARIMA combined with RTS (Rauch Tung Striebel) smoothing, constructed. smoothing enhanced version Kalman filter. The measurements are firstly processed by forward filter, then, separate backward pass obtaining solution. correctness was verified actual data. results show that based on can reflect performance well. This effectively reduce error caused stimulation when predicting. When using ARIMA-Kalman same well, accuracy higher than wells stimulation. Compared model, mean relative fitted reduced 46.3%, square 56.48%. improves uses We therefore conclude help us better forecasting stimulation, as other fuels output.