[發明專利]一種油水井行為分析預測裝置及方法

有效

• 申請號: 201410355231.4
• 申請日: 2014-07-24
• 公開（公告）號: CN104200059B
• 公開（公告）日: 2017-08-01
• 發明（設計）人: 黃克;周宏明;黃旺森
• 申請（專利權）人: 溫州大學
• 主分類號: G06F19/00
• 分類號: G06F19/00
• 代理公司: 杭州浙科專利事務所(普通合伙)33213
• 代理人: 吳秉中

1.一種油水井行為分析預測裝置，包括被測對象油井（1）及布置在油井周圍的水井（2），其特征在于：所述油井（1）、水井（2）上分別設有原油含水率傳感器（3）和注水量傳感器（4），分別用于采集油井的含水率和水井的注水量，并將采集的模擬量經A/D轉換器（5）轉換成數字量，再通過數據電纜傳輸至上位機（6），所述上位機對采集數據進行穩定性辨識；

所述上位機對采集數據進行穩定性辨識包括如下步驟：

設采集$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/145640dest\_path\_image001.GIF$個數據量$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/179455dest\_path\_image002.GIF$，采集到足夠的樣本并劃分為$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/891059dest\_path\_image003.GIF$個$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/693930dest\_path\_image004.GIF$矩陣，$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/770470dest\_path\_image003.GIF$為自然數；

（1）令$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/353899dest\_path\_image005.GIF$時的$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/541297dest\_path\_image001.GIF$階矩陣$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/526571dest\_path\_image006.GIF$為穩定狀態，取其協方差矩陣$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/774013dest\_path\_image006.GIF$，則特征方程為：$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/782420dest\_path\_image007.GIF$；

（2）認為余下的$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/835827dest\_path\_image008.GIF$個$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/613290dest\_path\_image001.GIF$階矩陣的穩定性判斷在正常過程數據$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/93950dest\_path\_image006.GIF$基礎上加擾動變化而成，若擾動對原過程影響不大，則認為是良性擾動，否則是惡性擾動，前者用于構建預測模型或者用于更新模型，用數學的語言表示為：第$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/324074dest\_path\_image009.GIF$個$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/112995dest\_path\_image001.GIF$階矩陣協方差為$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/807282dest\_path\_image010.GIF$，相應的特征根和特征向量變化為：$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/396526dest\_path\_image011.GIF$，$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/379525dest\_path\_image012.GIF$，則第$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/774735dest\_path\_image009.GIF$個$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/261211dest\_path\_image001.GIF$階矩陣協方差的特征方程為：$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/349253dest\_path\_image013.GIF$，對上式化解得到$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/553969dest\_path\_image014.GIF$；

（3）對上式取2范數得到$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/752869dest\_path\_image015.GIF$，即為第$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/93852dest\_path\_image009.GIF$個$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/290478dest\_path\_image001.GIF$階矩陣的協方差矩陣對正常過程數據矩陣的相對誤差，由于協方差矩陣是對稱陣，所以$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/920173dest\_path\_image016.GIF$，由于該式中矩陣屬于實對稱陣具有完備的特征向量系，采用冪法迭代求取矩陣的最大特征值；

（4）重復（2）、（3）直到$說明:\; D:/Program\; Files/gwssi/CPC客戶端/cases/inventions/c0b0dd8e-3457-42f7-a38a-4ef1695ce721/new/100001/594868dest\_path\_image017.GIF$，范數值較小的認為正常過程數據，否則認為是初始故障數據或故障數據，應丟棄。