自动化仪表
自動化儀錶
자동화의표
PROCESS AUTOMATION INSTRUMENTATION
2014年
3期
28-31
,共4页
标校周期%经验模态分解( EMD)%支持向量机( SVM)%灰色模型( GM)%组合预测
標校週期%經驗模態分解( EMD)%支持嚮量機( SVM)%灰色模型( GM)%組閤預測
표교주기%경험모태분해( EMD)%지지향량궤( SVM)%회색모형( GM)%조합예측
Calibration cycle%Empirical model decomposition ( EMD)%Support vector machine( SVM)%Grey model ( GM)%Combined forecasting
标校周期的确定是测量系统校准工作的重要内容。针对某武器系统测量设备标校周期的确定缺少理论支撑,提出EMD-SVM-GM组合预测优化方法。采用经验模态分解EMD方法分离原始校准数据的随机波动和趋势项。将预测分为两部分：针对随机波动的特性,采用支持向量机SVM进行处理；针对趋势项,采用灰色模型GM(1,1)进行预测,综合两者得到最后的预测结果。仿真实例表明,组合预测方法比单独的预测具有更高的精度,可以应用于标校周期的优化过程。
標校週期的確定是測量繫統校準工作的重要內容。針對某武器繫統測量設備標校週期的確定缺少理論支撐,提齣EMD-SVM-GM組閤預測優化方法。採用經驗模態分解EMD方法分離原始校準數據的隨機波動和趨勢項。將預測分為兩部分：針對隨機波動的特性,採用支持嚮量機SVM進行處理；針對趨勢項,採用灰色模型GM(1,1)進行預測,綜閤兩者得到最後的預測結果。倣真實例錶明,組閤預測方法比單獨的預測具有更高的精度,可以應用于標校週期的優化過程。
표교주기적학정시측량계통교준공작적중요내용。침대모무기계통측량설비표교주기적학정결소이론지탱,제출EMD-SVM-GM조합예측우화방법。채용경험모태분해EMD방법분리원시교준수거적수궤파동화추세항。장예측분위량부분：침대수궤파동적특성,채용지지향량궤SVM진행처리；침대추세항,채용회색모형GM(1,1)진행예측,종합량자득도최후적예측결과。방진실례표명,조합예측방법비단독적예측구유경고적정도,가이응용우표교주기적우화과정。
It is important to define the calibration cycle for calibration of the measuring system. Aiming at the difficulty of lack of theoretical support to determine the calibration cycle for the measuring device of a weapon system, the combined forecasting optimization method of EMD-SVM-GM is proposed. The stochastic volatility and trend term in original calibration data are separated by adopting empirical model decomposition. The forecasting is divided into two parts, the stochastic volatility is processed by using SVM, and the trend term is forecasted by adopting grey model GM(1,1), and the final forecast result is obtained by integrating these two parts. The simulation examples show that the combined forecasting method offers better accuracy and can be applied in optimization of the calibration cycle.
