农机化研究
農機化研究
농궤화연구
JOURNAL OF AGRICULTURAL MECHANIZATION RESEARCH
2015年
8期
20-25
,共6页
孙启新%张瑞宏%陈书法%杨进%芦新春
孫啟新%張瑞宏%陳書法%楊進%蘆新春
손계신%장서굉%진서법%양진%호신춘
水田耕整机%功耗%模糊系统%粒子群算法%回归分析
水田耕整機%功耗%模糊繫統%粒子群算法%迴歸分析
수전경정궤%공모%모호계통%입자군산법%회귀분석
tillage machine in paddy%power consumption%fuzzy system%particle swarm algorithm%regression analysis
针对耕整机功耗影响因素多且相互之间为复杂非线性关系的问题,提出了用模糊系统进行功耗建模的方法。以水田耕耙平地机为例,设计多功能耕整机功率消耗试验,由试验得到功耗的训练数据和测试数据。根据训练数据分别运用基本粒子群算法和带惯性因子的粒子群算法训练模糊系统,得到两种耕整机功耗预测模型。为验证模糊系统的可行性,同时用回归分析法对功耗进行建模,采用线性全因子多项式形式的回归函数,得到功耗经验公式。运用测试数据对建立的3种功耗模型进行测试,结果显示:采用带惯性因子的粒子群算法建立的模糊系统模型预测结果相对误差平均值小于11%,优于另外两种模型,用该方法进行水田耕整机功率消耗建模是可行的。
針對耕整機功耗影響因素多且相互之間為複雜非線性關繫的問題,提齣瞭用模糊繫統進行功耗建模的方法。以水田耕耙平地機為例,設計多功能耕整機功率消耗試驗,由試驗得到功耗的訓練數據和測試數據。根據訓練數據分彆運用基本粒子群算法和帶慣性因子的粒子群算法訓練模糊繫統,得到兩種耕整機功耗預測模型。為驗證模糊繫統的可行性,同時用迴歸分析法對功耗進行建模,採用線性全因子多項式形式的迴歸函數,得到功耗經驗公式。運用測試數據對建立的3種功耗模型進行測試,結果顯示:採用帶慣性因子的粒子群算法建立的模糊繫統模型預測結果相對誤差平均值小于11%,優于另外兩種模型,用該方法進行水田耕整機功率消耗建模是可行的。
침대경정궤공모영향인소다차상호지간위복잡비선성관계적문제,제출료용모호계통진행공모건모적방법。이수전경파평지궤위례,설계다공능경정궤공솔소모시험,유시험득도공모적훈련수거화측시수거。근거훈련수거분별운용기본입자군산법화대관성인자적입자군산법훈련모호계통,득도량충경정궤공모예측모형。위험증모호계통적가행성,동시용회귀분석법대공모진행건모,채용선성전인자다항식형식적회귀함수,득도공모경험공식。운용측시수거대건립적3충공모모형진행측시,결과현시:채용대관성인자적입자군산법건립적모호계통모형예측결과상대오차평균치소우11%,우우령외량충모형,용해방법진행수전경정궤공솔소모건모시가행적。
On the basis of the characteristics of multifunction tillage machine, fuzzy system upon particle swarm algorithm ( PSO) was adopted to model power consumption.Power consumption test was designed and carried out on a new type of multifunction tillage machine in paddy field.From the orthogonal test, training data and test data were obtained.By training data, fuzzy system based on basic PSO algorithm and PSO algorithm with inertia factor was trained respectively. In order to verify fuzzy system, one kind of experience formula of power consumption was obtained through regression a-nalysis.By testing data, the three models obtained by each method were tested respectively.The prediction effect of the fuzzy system trained by PSO algorithm with inertia factor was better than other methods ,whose average value of relative error was less than 11%.The prediction results showed that the fuzzy system trained by PSO algorithm with inertia factor was reliable to model power consumption.