Optimal Shunt Capacitor Allocation by Nonlinear Programming

experiment performed to measure the shielding effect of cages as soiling of the surfaces by animal excretions progressed with the length of time the cages had been occupied and at relative humidity levels of 30 to 90 percent. Data obtained from the experiment is pre¬ sented as a percentage of the induced current levels mea¬ sured with clean cages at low relative humidity. A composite of the averages induced under all test conditions, shown in Fig. 1, indicates that soiling of the cage surfaces had only a slight effect ( < 2 percent) on the currents induced at relative humidity levels below 40 percent. How¬ ever, at relative humidity levels above 40 percent, the average Se through the models decreased while the variabil¬ ity in Se between individual models increased. These effects were detectable after only 0.54 days of cage occupation time and accounted for an average Se reduction of 2.5 percent, standard deviation 2.5 percent at 50 percent relative humidity. At 70 percent relative humidity, the recom¬ mended maximum under the National Institutes of Health guidelines [2], the average Se reduction was 5 percent with a standard deviation of 5 percent. After 2.71 days of occupa¬ tion, a near typical length of time between cage washings in an exposure study, the average reduction in Se was 4 percent, standard deviation 2 percent, at 50 percent relative humidity and increased rapidly to 17 percent, standard deviation 6 percent, at 70 percent relative humidity. After 5.42 days of cage occupation, a longer than normal length of time between cage washings, the average Se reduction at 50 percent relative humidity was 6 percent, standard deviation 3 percent. At 70 percent relative humidity, the average Se reduction was a dramatic change of 23 percent with a standard deviation of 9 percent. However, these shielding effects were virtually eliminated by washing the cages. With the exception of measurements made at relative humidity levels above 80 percent, the average Se levels and variabili¬ ties between individual hemiellipsoids measured after the cages were washed were nearly indistinguishable from those measured at the beginning of the experiment.