WSP04999 - Laserfiche WebLink

Home Browse Search

prevents the typical hydrodynamic flow net with flow toward a sink being contributed from all .elevations (Koberg, 1965, and Harbeck 1958). The temperature drop due to outflow (TDEC) is subtracted from the temperature increase due to suppression (TPLS) and the net increase is added to the accumulated net increase from previous months to obtain the correction (TSUMA) to the mixed reservoir temperature (TMIX) from which the second iteration value of suppression is calculated. Since both TDEC and TPLS are functions of suppression, the corrected suppression value (SUPPB) implies a change in these secondary effects and a third iteralion is needed for all required parameter values. Since TPLS varies linearly with suppression volume the correction in this parameter is determined from the ratio SUPPB/SUPP(l). The revised TDEC value is calculated in the same manner as the original value but using the corrected mixed temperature (TMIXB). The final value of suppression (SUPPC) is determined from the corrected net increase in temperature for each month. The correction between SUPPB and SUPPC is normally small enough that no further iterations are required; however, certain conditions can result in convergence problems. thereby requiring additional iterations with a slightly modified algorithm. This portion of the program is discussed later. The final monthly values of suppression are weighted according to the assumed monthly distribution of evaporation and volumes of suppression and accumulated in order to produce the seasonal suppression volume. All pertinent monthly parameter values are included in the program output in order to allow a visual check on the validity of each step in the process. The program and output are listed in Appendix H. Because of the large number of variables used, a definition of all variables is also included in this appendix. Convergence of the algorithm. The basic algorithm described previously normally results in very rapid convergence (three iterations) to the proper mixed temperature and suppression values. Under some conditions, however, the algorithm became unsteble and did not converge at all. This problem occurred in situations with very high monthly outflow. The problem occurred when a very high outflow factor caused the net change in mixed temperature between iterations I and 2 to be significantly negative. This resulted in a second suppression estimate which was substantially higher than the first, with a corresponding large correction in secondary effects which gave a third suppression estimate slightly lower than the first and hence an unstable result for further iterations. In order to obtain convergence in this unusual situation a heuristic modification to the algorithm was added. The difference between the second and third estimates of monthly suppression are determined. If this difference (CB) exceeds 5 percent, additional iterations are performed. The fourth value of suppression is' computed as a function of the average between the first two values of mixed temperature. This results in a correction to lhe suppression estimate that is smaller (and more accurate) than thai of the basic algorithm. This resulted in convergence with only four iterations under all conditions for the available dala. The only assumption necessary for the validity of this perturbation of the algorithm is that the initial correction to mixed temperature is in the correct direction. This modification of the algorithm was not required in the 24 month model because the conditions which require it (high outflow to storage ratio) are never present in the types of high carryover storage reservoirs to which this model is applied. Twenty-four month model. This program is identical to the seasonal model during the first calendar year of operation except that it can be started at any month at which initial mixing is desired (NLAG) and it is continued past October to the end of December. The fact that artificial mix- ing would not normally be continued during November and December does not require a change in the model. The stop of artificial mixing implies that by this time the natural cooling process would have essentially eliminated the thermocline. As natural overturn occurs, the reservoir behaves for a period very much like it would if artificial mixing were still being induced (essentially a isothermal profile). The second year of the model simulates a situation in which no artificial mixing occurs; the natural thermocline develops except that it has superimposed upon it. the residual heat from the previous season's suppression. The purpose of this segment of the model is two-fold. (I) to determine the time required for the reservoir to recover its natural temperature and (2) to calculate the negative suppression (the above norrra! I,,""i:'apora- tion) during this heat decay period. Mixing may actually be continued for two or more consecutive seasons, but the proper combina- 30