基于油氣識(shí)別的概率神經(jīng)網(wǎng)絡(luò)算法

隨著我國(guó)對(duì)石油需求的逐年增加，對(duì)石油勘探技術(shù)的要求也越來(lái)越高，而測(cè)井信
號(hào)的準(zhǔn)確識(shí)別對(duì)石油的勘探技術(shù)有著重大的意義。本文通過(guò)對(duì)測(cè)井信號(hào)的預(yù)處理，利用徑向基神經(jīng)網(wǎng)絡(luò)的一種延伸--概率神經(jīng)網(wǎng)絡(luò)，對(duì)測(cè)井信號(hào)進(jìn)行網(wǎng)絡(luò)訓(xùn)練，達(dá)到識(shí)別的效果。研究和實(shí)驗(yàn)結(jié)果表明，本文方法對(duì)測(cè)井信號(hào)的識(shí)別有很高的準(zhǔn)確性，實(shí)用性，運(yùn)行速度快。
關(guān)鍵詞：模式識(shí)別；徑向基；概率神經(jīng)網(wǎng)絡(luò)；儲(chǔ)層
Abstract：As the oil demand increased year by year, oil exploration technology requirements are also getting higher and higher, so identify the exact signal for oil exploration technology has a great significance. Based on the logging signal preconditioning, Used a kind of RBF network extension—PNN to train the logging signal to attach the clustering effect. Research and experimental results show that this method of identification on logging signals has high accuracy, practicality, run faster.
Key words：Pattern Recognition; RBF; PNN; Reservoir