| 主权项 |
1. A method of performing real-time correction of a water stage forecast, the method comprising:
obtaining at least one predicted water stage of at least one time and a predicted water stage of a next time after the at least one time, wherein a previous time of the at least one time is a current time; obtaining at least one observed water stage of the at least one time; utilizing a Time Series method to generate a water stage error estimate of the next time according to a following equation:
εTSMt+1=f(Hpredt+1,Hpredt,Hpredt−1, . . . ,Hpredt−p, εt, εt−1, . . . , εt−p);wherein:εTSMt+1 is the water stage error estimate of the next time generated by a Time Series model f;εt=Hpredt−Hobst;Hpredt is a predicted water stage of the current time of the at least one time;Hpredt+1 is the predicted water stage of the next time;Hpredt−1 is a predicted water stage of a previous time of the current time;Hpredt−p is a predicted water stage of a previous pth time of the current time, wherein p is an integer greater than 1;Hobst is an observed water stage of the current time;εt is a water stage error of the current time;εt−1 is a water stage error of the previous time;εt−p is a water stage error of the previous pth time; utilizing an Average Deviation method to generate an average error of the next time according to a following equation:ɛADt+1=1n∑i=1n[Hpredt-n+i-Hobst-n+i],n=1,2,…,Nsp; wherein:εADt+1 is the average error of the next timeNSP is a predetermined interval number for calculating the average error of the next time;Hpredt−n+i is a predicted water stage of a previous (n-i)th time of the current time;Hobst−n+i is an observed water stage of the previous (n-i)th time of the current time, wherein (n-i) is an integer; generating a system error of the water stage forecast according to the water stage error estimate of the next time and the average error of the next time; utilizing a Kalman filter method to generate a random error of the water stage forecast; generating a water stage forecast correction of the next time using values of the system error and the random error at a same time according to a following equation:ɛcombt+1=1Nmodel∑m=1NmodelɛTSM,mt+1+ɛADt+1+ɛKF; wherein:εKF is the random error;εcombt+1 is the water stage forecast correction of the next time;Nmod el is a model number of the Time Series; andεTSM,mt+1 is the water stage error estimate of the next time generated by an mth mode of the Time Series model f; and correcting a predicted water stage of the next time according to the water stage forecast correction of the next time and the predicted water stage of the next time. |
