Quantifying the relative importance of soil moisture, nitrogen, and temperature on the urea hydrolysis rate

Abstract

Urea hydrolysis is a complicated process influenced by multiple factors. Most previous studies only examined the causative factors of the process without ranking these factors’ relative importance quantitatively. In this work, the experimental analysis method, ANOVA method, and backpropagation (BP) neural network method were used to rank the effects of moisture content (W), nitrogen amount (F), and temperature (T) on the urea hydrolysis rate. A group of 22 artificial neural network structures with different numbers of neurons in the hidden layer (2 ≤ L ≤ 12), different initial connection weights, and different learning algorithms was trained and validated using the data set. The relative importance of the factors that affect urea hydrolysis was studied on the basis of information on the weight and threshold value. Results showed that the individual factors and the interaction between any two factors exerted an extremely significant effect (p < 0.01) on the hydrolysis rate, except for the interaction among the three factors. The ANOVA method achieved the importance ranking of T > F > W > T*F > T* W > F*W > T* F*W for the hydrolysis rate. The incremental backpropagation-BP-K model with the topological structure of 3-6-1 expressed the relationship between the input variables and the output variable accurately. All of the methods indicated that the degree of influence of the factors on urea hydrolysis followed the descending order of T > F > W. Temperature was the key factor in the hydrolytic process. The results of the importance analysis of the three methods were consistent.