Results And Discussion - Magnetic Sludge Studies - Magnetic Sludge Factorial Experiment

The initial results from the magnetic sludge factorial experiment are summarized in Table 2. For ease in interpretation, the four separate variables are identified with the high and low level of each variable being represented by + and - signs, respectively. Several important relationships are evident from an inspection of the data in Table 2. For example, magnetic sludges were obtained only during the high temperature tests. Secondly, with the exception of the last three tests, there is an apparent correlation between settled sludge volume (or sludge settling rate) and pH. Data on residual dissolved Fee+ indicate that precipitation of iron is not complete at pH 8 and room temperature (Tests l, 3, 5, and 7). Although this should not have any confounding effect on the interpretation of the results, it is an important observation and an aspect of the overall conceptual process that should be considered carefully during any future developmental work. The data in Table 2 were employed more fully in the calculation of main effects and variable interactions for each of the three responses. For this purpose, a codified design and calculation matrix, shown in Table 3, was used. The derivation of this matrix is described by Davies.(9) The positive or negative signs for variable interactions are obtained simply by algebraic multiplication of the signs for each single variable involved in the interaction. Main effects and interactions are calculated by multiplying the response values (Table 2) by the corresponding effect elements (Table 3), summing the results, and dividing by 8. The resulting value for single variables is also sometimes designated as an average effect, since (for tyro levels of the variable only) it is simply the difference between the average response of all trials carried out at the first level of the variable (factor) and that of all trials at the second level. If the effect of one variable changes at different levels of another variable, the two variables are said to interact. The relative importance of the variable or variables in question is reflected by the absolute magnitude of the calculated main effect or interaction. First-order (two variable) interactions are usually easy to interpret in terms of the physical behavior of the system. Second- and higher-order interactions are often less readily understood, and are often of minor significance in any case. The main effects and interactions derived from the present study in the manner described above are listed in Table 4. It should be noted that