ANN-based urban and rural mixed garbage aerobic fermentation humic degree prediction method
A technology of aerobic fermentation and prediction method, applied in neural learning methods, chemical process analysis/design, bio-organic part treatment, etc., can solve the difficulty of predicting the humification degree of garbage aerobic fermentation and the inability to predict the aerobic fermentation process of garbage Problems such as humification trends and laws can achieve the effects of improving prediction accuracy, realizing harmlessness, and making up for uncontrollable effects
- Summary
- Abstract
- Description
- Claims
- Application Information
AI Technical Summary
Problems solved by technology
Method used
Image
Examples
specific Embodiment approach 1
[0011] Embodiment 1: An ANN-based method for predicting the degree of humification by aerobic fermentation of urban and rural mixed garbage in this embodiment is as follows: obtaining environmental information of the garbage to be predicted and information on the physical and chemical properties of the garbage itself, and analyzing the environmental information and the physical and chemical properties of the garbage itself. At least one data in the property information is input into the trained optimal BP artificial neural network to obtain the humus production amount of the garbage to be predicted, and the humus degree of the garbage to be predicted is determined according to the humus production amount.
[0012] Input the environmental information of the garbage to be predicted and the number of parameters of the physical and chemical properties of the garbage itself to be set according to experience;
[0013] The environmental information of the garbage to be predicted inclu...
Embodiment
[0036] Embodiment: In order to verify the feasibility of the neural network structure of the present invention, the following experiments are carried out in this embodiment:
[0037] First, based on the same optimization method, comparing the convergence effects of models with different learning rates, when the optimization algorithm is Adam, the optimal learning rate is 0.1, and when the optimization algorithm is SGD, the optimal learning rate is 1×10 -5 . According to this result, different random seeds are used for random initialization. The results show that the use of different random seeds has a certain influence on the learning convergence of the model, and they all show a convergence trend. Compared with seeds 84 and 41, the convergence effect is better. Based on the above analysis, the convergence effects of the two model structures were compared, and the optimal model was obtained, that is, the model structure was 7-14-1, the optimization algorithm was Adam, the rand...
PUM
Login to View More Abstract
Description
Claims
Application Information
Login to View More 


