秦可 卜仁祥 李鐵山 劉勇 鄭力銘



摘要:為提高船舶風(fēng)壓差的預(yù)測(cè)精度,使船舶能夠更快穩(wěn)定在計(jì)劃航線上以保障航行安全,提出一種基于主成分分析(principal component analysis, PCA)法和自適應(yīng)粒子群優(yōu)化(self-adaptive particle swarm optimization, SAPSO)算法的船舶風(fēng)壓差神經(jīng)網(wǎng)絡(luò)預(yù)測(cè)模型。該方法采用PCA法對(duì)航行數(shù)據(jù)進(jìn)行預(yù)處理,然后將數(shù)據(jù)輸入由SAPSO算法優(yōu)化的BP神經(jīng)網(wǎng)絡(luò)中,改變以往通過(guò)復(fù)雜的數(shù)學(xué)建模計(jì)算風(fēng)壓差的方法,提高預(yù)測(cè)的時(shí)效性和準(zhǔn)確性。利用實(shí)船數(shù)據(jù)對(duì)模型進(jìn)行船舶風(fēng)壓差的實(shí)時(shí)預(yù)測(cè)仿真,結(jié)果驗(yàn)證了該預(yù)測(cè)模型具有較高的可靠性。
關(guān)鍵詞: 船舶;風(fēng)壓差預(yù)測(cè);主成分分析(PCA);自適應(yīng);粒子群優(yōu)化
中圖分類(lèi)號(hào): U675.79 ? ?文獻(xiàn)標(biāo)志碼: A
Neural network prediction model of ship leeway angle
based on PCA and SAPSO
QIN Ke, BU Renxiang, LI Tieshan, LIU Yong, ZHENG Liming
(Navigation College, Dalian Maritime University, Dalian 116026, Liaoning, China)
Abstract: In order to improve the prediction accuracy of ship leeway angle so that the ship can be stabilized on the planned route more quickly to ensure navigation safety, a neural network prediction model of ship leeway angle is proposed based on the principal component analysis (PCA) method and the self-adaptive particle swarm optimization (SAPSO) algorithm. In the method, the navigation data are preprocessed by the PCA method, and the processed data are input into the BP neural network optimized by the SAPSO algorithm. It changes the previous method of calculating the leeway angle through complex mathematical modeling, and improves the timeliness and accuracy of prediction. The real-time prediction simulation of ship leeway angle by the model is carried out with the real ship data, and the results show that the prediction model has high reliability.
Key words: ship; leeway angle prediction; principal component analysis (PCA); self-adaption; particle swarm optimization
0 引 言
隨著航運(yùn)業(yè)的快速發(fā)展,船舶的航行安全和效率問(wèn)題日益得到重視。船舶風(fēng)壓差預(yù)測(cè)可以使船舶快速穩(wěn)定在計(jì)劃航線上,減小航跡帶寬度,對(duì)于船舶航行安全和效率都有重要的理論與現(xiàn)實(shí)意義。然而,船舶在海上的運(yùn)動(dòng)不但具有非線性、時(shí)滯性,而且受風(fēng)、浪、流等因素影響較大,這些都增加了船舶風(fēng)壓差預(yù)測(cè)的難度[1]。
在利用船舶運(yùn)動(dòng)模型預(yù)測(cè)風(fēng)壓差的方法中,由于船舶裝載量、航速和外界環(huán)境等的變化對(duì)船舶風(fēng)壓差的影響很大,難以建立船舶實(shí)時(shí)運(yùn)動(dòng)數(shù)學(xué)模型,且這種方法對(duì)數(shù)據(jù)源的要求較高,所以研究的船舶大多處于理想狀態(tài)[2]。張潤(rùn)濤[3]通過(guò)建立風(fēng)、流干擾下的船舶運(yùn)動(dòng)數(shù)學(xué)模型對(duì)船舶的風(fēng)流壓差進(jìn)行了研究,但運(yùn)算過(guò)程復(fù)雜且受多種不確定干擾因素影響;余力等[4]提出運(yùn)用風(fēng)流壓差表,通過(guò)別爾舍茨方法求取船舶風(fēng)壓差系數(shù);SUN等[5]基于經(jīng)驗(yàn)公式,利用轉(zhuǎn)向前后風(fēng)舷角的變化預(yù)測(cè)轉(zhuǎn)向后的風(fēng)流壓差。……