Sparse limited nonnegative matrix decomposition algorithm based ultrafiltration membrane water treatment prediction method
Sparse limited nonnegative matrix decomposition algorithm based ultrafiltration membrane water treatment prediction method
 CN 105,005,684 A
 Filed: 06/19/2015
 Published: 10/28/2015
 Est. Priority Date: 06/19/2015
 Status: Active Application
First Claim
1. , based on the ultra filtration membrane water treatment Forecasting Methodology of sparse constrained nonnegative matrix factorization algorithm, it is characterized in that, comprise the following steps:
 (1) sparse constrained nonnegative matrix factorization algorithm is utilized to set up the forecast model of ultra filtration membrane sewage disposal process;
(2) respectively under plan steady state (SS) and unsteady state, the parameter of the forecast model utilizing GA algorithm optimization step (1) to set up;
(3) respectively under plan steady state (SS) and unsteady state, the variation tendency of forecast model to membrane flux and membrane pollution resistance utilizing step (2) to optimize is predicted, and the basic operation parameter of analyzing film distillation is on the impact of membrane flux and membrane pollution resistance;
(4) sensitivity analysis calculating is carried out to predicting the outcome of obtaining of step (3), determine the leading factor affecting membrane flux and membrane pollution resistance.
Abstract
The invention discloses a sparse limited nonnegative matrix decomposition algorithm based ultrafiltration membrane water treatment prediction method. Firstly, a sparse limited nonnegative matrix decomposition algorithm is utilized to construct a prediction model for a membrane distillation wastewater treatment process; secondly, a GA algorithm is utilized to optimize parameters of the prediction model in a pseudosteady state and a nonsteady state respectively; thirdly, the optimized prediction model is utilized to predict the change trend of the membrane flux and the membrane pollution resistance and to analyze the influence of basic operation parameters of membrane distillation on the membrane flux and the membrane pollution resistance in the pseudosteady state and the nonsteady state respectively; and finally, a prediction result is subjected to sensitivity analysis calculation to determine leading factors which influence the membrane flux and the membrane pollution resistance. According to the method, the sparse limited nonnegative matrix decomposition algorithm is utilized to predict the change situations of the membrane flux and the membrane pollution resistance in real time, and the influence of the basic operation parameters of membrane distillation on the membrane flux and the membrane pollution resistance is clarified and quantified.

7 Citations
A kind of transformer temperature rise of hot spot prediction technique comparing optical fiber temperaturemeasurement  
Patent #
CN 109,344,559 A
Filed 12/03/2018

Current Assignee

Data processing method based on incremental locally convex nonnegative matrix factorization  
Patent #
CN 105,389,291 A
Filed 12/21/2015

Current Assignee

Difference expression gene discrimination method based on joint constrained nonnegative matrix decomposition  
Patent #
CN 107,016,261 A
Filed 04/11/2017

Current Assignee

Process for removing organic and inorganic contaminants from refinery wastewater streams employing ultrafiltration and reverse osmosis  
Patent #
US 6,054,050 A
Filed 07/21/1998

Current Assignee
N/A

Nonnegative matrix factorization in a relational database management system  
Patent #
US 20050246354A1
Filed 08/27/2004

Current Assignee
N/A

Video motion characteristic extracting method based on local sparse constraint nonnegative matrix factorization  
Patent #
CN 102,254,328 A
Filed 05/17/2011

Current Assignee

Image retrieval method based on robust nonnegative matrix factorization  
Patent #
CN 103,886,066 A
Filed 03/20/2014

Current Assignee

5 Claims

1. , based on the ultra filtration membrane water treatment Forecasting Methodology of sparse constrained nonnegative matrix factorization algorithm, it is characterized in that, comprise the following steps:

(1) sparse constrained nonnegative matrix factorization algorithm is utilized to set up the forecast model of ultra filtration membrane sewage disposal process; (2) respectively under plan steady state (SS) and unsteady state, the parameter of the forecast model utilizing GA algorithm optimization step (1) to set up; (3) respectively under plan steady state (SS) and unsteady state, the variation tendency of forecast model to membrane flux and membrane pollution resistance utilizing step (2) to optimize is predicted, and the basic operation parameter of analyzing film distillation is on the impact of membrane flux and membrane pollution resistance; (4) sensitivity analysis calculating is carried out to predicting the outcome of obtaining of step (3), determine the leading factor affecting membrane flux and membrane pollution resistance.


2.
2., according to claim 1 based on the ultra filtration membrane water treatment Forecasting Methodology of sparse constrained nonnegative matrix factorization algorithm, it is characterized in that, the detailed process of step (1) is as follows: 
First, the objective function F of sparse constrained nonnegative matrix factorization algorithm is defined; F(W,H)＝
VWH ^{2}＝
Σ
_{ij}[V _{ij}(WH) _{ij}] ^{2}(1)In formula (1), V is input sample of data collection, and W is eigenmatrix, and H is projection sparse matrix;
V _{ij}represent V ith row jth column element,Eigenmatrix W and projection sparse matrix H in initialized target function F, If W has sparse constraint, then W=Wθ
is first set _{w}(WHV) H ^{t}, then according to nonnegative sparse projection algorithm, each row of W are converted to nonnegative, keep its L simultaneously _{2}normal form is constant, arranges its L simultaneously _{1}normal form, to reach the degree of rarefication S specified _{w};
If W does not have sparse constraint, then interative computation


3. according to claim 1 based on the ultra filtration membrane water treatment Forecasting Methodology of sparse constrained nonnegative matrix factorization algorithm, it is characterized in that:
 the parameter of the described GA of the utilization algorithm optimization of step (2) comprises S _{w}, S _{h}, θ
_{w}and θ
_{h}.
 the parameter of the described GA of the utilization algorithm optimization of step (2) comprises S _{w}, S _{h}, θ

4. according to claim 1 based on the ultra filtration membrane water treatment Forecasting Methodology of sparse constrained nonnegative matrix factorization algorithm, it is characterized in that, the detailed process of step (3) is, proceeds as follows respectively under plan steady state (SS) and unsteady state:
Using temperature, crossflow velocity, inlet flowpatterm, transmembrane pressure as the input data of forecast model, using membrane flux and the membrane pollution resistance output data as forecast model, only change temperature difference, crossflow velocity, inlet flowpatterm, membrane aperture and a parameter in the Membrane Materials processing time at every turn, ensure other parameter constants, use the parameter of GA algorithm optimization forecast model, and the parameter substitution forecast model after optimizing is predicted, recycling rootmeansquare error RMSE and regression coefficient R ^{2}these two evaluatings are evaluated estimated performance, when RMSE gets over close to 0 and R ^{2}more close to 1 time, show that estimated performance is better.

5. according to claim 1 based on the ultra filtration membrane water treatment Forecasting Methodology of sparse constrained nonnegative matrix factorization algorithm, it is characterized in that:
 in step (4), adopting Spearman correlativity, gamma correlativity, Ken Deer correlativity and Pearson came correlativity to carry out sensitivity analysis calculating to predicting the outcome respectively.
Specification(s)