Decision-making method of comprehensive alumina production indexes based on multi-scale deep convolutional network

Patent metadata

Field	Value
Publication number	US-11487962-B2
Application number	US-201916955490-A
Country	US
Kind code	B2
Filing date	Jul 17, 2019
Priority date	Jul 16, 2019
Publication date	Nov 1, 2022
Grant date	Nov 1, 2022

Abstract

The invention provides a decision-making method of comprehensive alumina production indexes based on a multi-scale deep convolutional network. The method mainly consists of several sub-models: a multi-scale deep splicing convolutional neural network prediction sub-model reflecting the influence of bottom-layer production process indexes on the comprehensive alumina production indexes, a full connecting neural network prediction sub-model reflecting the influence of upper-layer dispatching indexes on the comprehensive alumina production indexes, a full connecting neural network prediction sub-model reflecting the influence of the comprehensive alumina production indexes at a past time on current comprehensive alumina production indexes, and a multi-scale information neural network integrated model for collaborative optimization of sub-model parameters. According to the method, through an integrated prediction model structure, a memory capacity of a superficial-layer network and a feature extraction capacity of a deep-layer network, a precise decision-making for the comprehensive alumina production indexes is realized.

First claim

What is claimed is:

1. A decision-making method of comprehensive alumina production indexes based on a multi-scale deep convolutional network, comprising the following steps:
step 1: collecting production index data generated in an alumina production process, partitioning the collected production index data into a training dataset, a validation dataset and a test dataset by using a sample partition algorithm, and pre-processing the data by using a data pre-processing algorithm so as to obtain data which can be used for modelling;
step 2: constructing a multi-scale deep splicing convolutional network prediction sub-model which reflects influence of bottom-layer production process indexes on the comprehensive alumina production indexes;
forming an information matrix by a sampling value of alumina production process indexes closely related to final comprehensive alumina production indexes for a period of time, which is expressed as follows:
Y a =[ v ( k v ), v ( k v −1), v ( k v −2), . . . v ( k v −n k )],

wherein Y d is an input of the multi-scale deep splicing convolutional network prediction sub-model reflecting the influence of the bottom-layer production process indexes, and contains characteristic information of the influence of the bottom-layer production process indexes on the final comprehensive alumina production indexes, and a size thereof is l p ×n k , l p , is a number of the bottom-layer production process indexes, n k is a sampling frequency of the bottom-layer production process indexes in a period of time, v(k v ) is any one of the bottom-layer production process indexes, and k v represents a certain sampling time,
wherein the multi-scale deep splicing convolutional network prediction sub-model reflecting the influence of the bottom-layer production process indexes on the comprehensive alumina production indexes comprises 3 multi-scale convolutional layers, 3 pooling layers, 1 common convolutional layer, 1 full connected layer and an output transformation layer;
step 3: constructing a full connected neural network prediction sub-model which reflects influence of alumina upper-layer dispatching indexes on the comprehensive alumina production indexes;
forming an information matrix by a sampling value of the upper-layer dispatching indexes in a period of time, which is expressed as follows:
X d =[ q ( k q ), v ( k q −1), v ( k q −2), . . . , v ( k q −n h )],

wherein X d is an input of the prediction sub-model reflecting the influence of the upper-layer dispatching indexes on the comprehensive alumina production indexes, a size thereof is l h ×n h , l h is a number of the upper-layer dispatching indexes, n h is a sampling frequency of the upper-layer dispatching indexes in a period of time, q(k q ) is any one of the upper-layer dispatching indexes, k q represents a certain sampling time,
wherein the full connected neural network prediction sub-model uses a single-layer full connected neural network, a number of nodes of a full connected network is consistent with the number of the upper-layer dispatching indexes, and a Sigmoid activation function is selected as an activation function;
step 4: constructing the full connected neural network prediction sub-model which reflects influence of the comprehensive alumina production indexes at a past time on current comprehensive alumina production indexes:
step 4.1: defining a sample set of historical working conditions: simply expressing working conditions {X d , Y d } through initial production conditions and X d of the upper-layer dispatching indexes, and besides, forming the sample set T′={(X d ,Z d+1 )}⊂T of the historical working conditions from corresponding comprehensive alumina production indexes, wherein d=1,2, . . . , n d ;
step 4.2: grouping the historical working conditions:
firstly, adopting an automatic clustering method based on a Gaussian mixed model so as to obtain classification of the historical working conditions; expressing a clustering result as C={c 1 , . . . c l c }, wherein l c is a number of the obtained historical working conditions;
step 4.3: searching for the comprehensive alumina production indexes at the past time related to current working conditions so as to obtain comprehensive alumina production index information not only including correlation information of the comprehensive alumina production indexes at the past time, but also including correlation information of the historical working conditions and the current working conditions of the comprehensive alumina production indexes,
wherein the sub-model adopts a single-layer full connected neural network, the number of nodes of the full connected network is consistent with a size of input variables, and the Sigmoid activation function is selected as an activation function;
step 5: building a multi-scale information neural network integrated model for collaborative optimization of sub-model parameters,
the integrated model consists of a single-layer neural network, wherein a number of input source variables is 3, corresponding to outputs of the three prediction sub-models established in the steps 2 to 4, and a number of output variables is 1, which indicates a prediction value of the comprehensive alumina production indexes; the Sigmoid activation function is selected as a nonlinear activation function of output nodes; the integrated model trains network parameters of the three prediction sub-models at the same time according to gradient information of a prediction error loss function of the comprehensive alumina production indexes, i.e. training errors of the integrated model can be reversely propagated to an input layer of the prediction sub-model reflecting influence of various types of information on the comprehensive alumina production indexes at the same time, and a weight of each type of input information is subjected to common influence of other input information on the training errors of the integrated model at the same time, so that the collaborative optimization of influence weight of different time scale information on the comprehensive alumina production indexes is realized, and besides, complexity of the prediction sub-models is reduced; and
step 6: performing optimization decision-making on the comprehensive alumina production indexes through the established models:
according to one or more models established for the comprehensive alumina production indexes, performing single-objective or multi-objective optimization decision-making; and giving boundary conditions of decision-making variables, and performing the optimization decision-making through a single-objective or multi-objective optimization algorithm, so as to obtain an optimization decision-making result of the comprehensive alumina production indexes.

2. The decision-making method according to claim 1 , wherein the bottom-layer production process indexes in the step 2 comprise: grinding AO and A/S, ore adjustment Nk, ore adjustment solid content, lime effective calcium, bauxite slurry solid content, bauxite slurry fineness, digestion red mud A/S, digestion red mud N/S, discharge A/S, digestion ak, digestion solid content, green liquor Nk, green liquor ak, green liquor seston, pregnant liquor seston, circulating spent liquor Nc/Nt and Nk, circulation efficiency, seed precipitation spent liquor seston, seed precipitation end tank ak, decomposition rate, flat plate filter cakes with water and alkali, flat spent seston, decomposition spent liquor ak and water content of red mud filter cakes.

3. The decision-making method according to claim 1 , wherein the multi-scale convolutional layer in the step 2 uses convolutional kernels of 3 sizes at the same time to perform convolution operations in parallel, and splices multi-scale features obtained together as an input of next layer; in a convoluti

Assignees

• Univ Northeastern

Inventors

• Liu Changxin
• Xu Depeng
• Ding Jinliang
• Chai Tianyou

Classifications

• G06N20/00 Primary

  Machine learning · CPC title

• G06F18/23 Primary

  Clustering techniques · CPC title

• G06N3/048

  Activation functions · CPC title

• G06N7/01

  Probabilistic graphical models, e.g. probabilistic networks · CPC title

• G06F18/214

  Generating training patterns; Bootstrap methods, e.g. bagging or boosting · CPC title