Water control method for loosening and conditioning process based on neural network model and double parameter correction

A control method for water supply during loosening and conditioning based on a neural network model and a double parameter correction, comprising:

• S1, establishing a neural network-based model for predicting the amount of water supplied during loosening and conditioning;
• using historical production data, taking an inlet moisture and a set value of an outlet moisture in a loosening and conditioning process as input values, and the amount of water supplied as an output value, training a neural network to establish the neural network-based model for predicting the amount of water supplied during loosening and conditioning;
• S2, predicting and distributing the total water supplied;
• collecting parameters under the current production conditions, and based on the neural network-based model obtained in S1, obtaining an optimal total water supplied corresponding to the set value of the outlet moisture;
• then, according to distribution coefficients of water supply at an inlet and an outlet, distributing the amount of water supplied at the inlet and the amount of water supplied at the outlet;
• S3, correcting the model based on a material balance calculation correction and a deviation correction
• according to the deviation that the actual value of the outlet moisture deviates from the set value of the outlet moisture, performing the material balance calculation correction or the deviation correction;
• when the deviation of the outlet moisture is greater or equal to 0.5%, performing correcting by using the value from the material balance calculation correction, which is fed back to the total water supplied for correction;
• when the deviation of the outlet moisture is less than 0.5%, performing correcting by using the value from the deviation correction, which is fed back to the amount of water supplied at the outlet for correction.