Raw tobacco stack internal temperature prediction algorithm based on improved LSTM

Abstract

The invention discloses a raw tobacco stack internal temperature prediction algorithm based on an improved LSTM, and belongs to the field of raw tobacco maintenance, an SOM algorithm and the LSTM are combined, firstly, data are preprocessed, time sequence data are normalized, then the normalized data are divided into a training set and a test set, the training set and the test set are input into an SOM neural network, the data are clustered, and then an LSTM neural network structure is constructed; and inputting temperature training data for training, predicting the indoor temperature in the nearby time period in the future, then predicting and outputting, and measuring the algorithm effect through actual data. Experimental results prove that the algorithm can realize relatively accurate prediction of the internal temperature of the crude tobacco stack.

Description

technical field [0001] The invention belongs to the field of raw smoke maintenance, and more particularly relates to an improved LSTM-based internal temperature prediction algorithm for raw smoke stacks Background technique [0002] At present, cigarettes have become the main storage form of raw cigarettes. The temperature and humidity environment inside the stacks where cigarettes are stored directly affects the alcoholization effect of cigarettes, and the alcoholization effect of cigarettes is related to the quality of the finished cigarettes. The alcoholization effect of cigarettes is closely related to the temperature and humidity of the storage environment. Therefore, in the natural alcoholization process of cigarettes, it is necessary to monitor the temperature and humidity environment inside the cigarette stacks, which can guide relevant personnel to take timely measures to avoid mildew of the preserved tobacco leaves and reduce losses. Therefore, it is very necessar...

AI Technical Summary

Problems solved by technology

[0006] In order to solve the shortcomings of RNN's gradient disappearance and gradient explosion, long-term memory capacity is insufficient, and to improve the training speed and calculation accuracy of LSTM, the present invention proposes an internal temperature prediction algorithm for raw cigarette stacks based on improved LSTM, which combines SOM (Self -Organizing Maps (SOM) algorithm combined with LSTM, first preprocess the data, normalize the time series data, then divide the normalized data into training set and test set, input the SOM neural network, cluster the data, and then Construct the LSTM neural network structure, input temperature training data for training, and at the same time predict the indoor temperature in the near future, and then predict the output, and measure the effect of the algorithm through actual data, and measure the experiment through indicators such as root mean square error (RMSE) result

Images