System and method for on-line analysis of structure of dried shredded tobacco

A system for on-line analysis of a structure of dried shredded tobacco, comprising:

• a shredded tobacco structure on-line analysis module, wherein the shredded tobacco structure on-line analysis module comprises a sample extraction unit, a sample analysis unit and a shredded tobacco filling capacity data acquisition unit, wherein
• the sample extraction unit is configured to extract 5 g±0.5 g of the dried shredded tobacco online and then acquire images of the 5 g±0.5 g of the dried shredded tobacco through an image acquisition device for detection by the sample analysis unit;
• the sample analysis unit is configured to automatically perform statistics on length data of the 5 g±0.5 g of the dried shredded tobacco through an image processing software and a data statistical software to obtain statistical data, and then perform interval grouping and storage on the statistical data for a construction of a prediction model;
• and
• the shredded tobacco filling capacity data acquisition unit is configured to extract 10.0 g to 20.0 g of the dried shredded tobacco online and simultaneously put the 10.0 g to 20.0 g of the dried shredded tobacco and the 5 g±0.5 g of the dried shredded tobacco extracted by the sample extraction unit into a shredded tobacco filling capacity detector for detection for the construction of the prediction model;
• a shredded tobacco filling capacity prediction model module, wherein the shredded tobacco filling capacity prediction model module comprises a model parameter screening unit and a model construction unit, wherein
• the model construction unit constructs a first model and a second model through an artificial neural network (ANN) model, and performs a third construction through the ANN model based on an output parameter of the first model and an output parameter of the second model to obtain a third model;
• and
• the third model is a shredded tobacco filling capacity prediction model;
• and
• an early-warning module, wherein
• an early-warning threshold is provided in the early-warning module, and when a deviation of a predicted shredded tobacco filling capacity from a standard value exceeds the early-warning threshold, an alarm is given;
• wherein the model parameter screening unit screens out model parameters of:
• device parameters, wherein the device parameters comprise a cylinder wall temperature of a thin-plate shredded tobacco dryer, an inlet moisture amount of the thin-plate shredded tobacco dryer, an outlet moisture amount of the thin-plate shredded tobacco dryer, a hot air temperature of the thin-plate shredded tobacco dryer, and an outlet temperature of the thin-plate shredded tobacco dryer;
• and
• shredded tobacco structure parameters, wherein the shredded tobacco structure parameters comprise a long and short shredded tobacco rate, a broken shredded tobacco rate, and a measured shredded tobacco filling capacity;
• the first model is a three-layer ANN model, and has three input values, two output values and two neurons;
• the first model is configured to predict the outlet moisture amount of the thin-plate shredded tobacco dryer and the outlet temperature of the thin-plate shredded tobacco dryer;
• the three input values of the first model are the cylinder wall temperature of the thin-plate shredded tobacco dryer, the inlet moisture amount of the thin-plate shredded tobacco dryer, and the hot air temperature of the thin-plate shredded tobacco dryer;
• the two output values of the first model are the outlet moisture amount of the thin-plate shredded tobacco dryer and the outlet temperature of the thin-plate shredded tobacco dryer;
• and
• the second model is a three-layer ANN model, and has three input values, one output value and two neurons;
• the second model is configured to predict the predicted shredded tobacco filling capacity;
• the three input values of the second model are the long and short shredded tobacco rate, the broken shredded tobacco rate, and the measured shredded tobacco filling capacity;
• the one output value of the second model is the predicted shredded tobacco filling capacity.