Mississippi Chronicle

Prof. Chongchong Qi, A/Prof Qiusong Chen and other collaborators announce the idea of “Machine-learning aided design for cemented paste backfill”

 Breaking News
  • No posts were found

Prof. Chongchong Qi, A/Prof Qiusong Chen and other collaborators announce the idea of “Machine-learning aided design for cemented paste backfill”

September 16
22:28 2019
Prof. Chongchong Qi, A/Prof Qiusong Chen and other collaborators announce the idea of “Machine-learning aided design for cemented paste backfill”

Photos of Prof. Chongchong Qi (left) and A/Prof. Qiusong Chen (right).

Prof. Chongchong Qi and A/Prof Qiusong Chen are experts in resources and environmental engineering at Central South University. Along with other collaborators, especially Prof. Andy Fourie at The University of Western Australia, they proposed the idea of machine-learning aided design for cemented paste backfill (also known as intelligent mining for backfill). A series of papers have been written and published in leading journals like Journal of Cleaner Production, Minerals Engineering, and Powder Technology, to promote this novel idea. Mineral processing tailings (MPT) are inevitable by-products of hard rock mining. They are considered a major source of contaminants to the environment, especially in global arid and semiarid regions. A study has shown that more than 25 billion tons of mineral processing tailings have been produced in China. Poor disposal of the same can be observed in other nations like Australia, Brazil, etc. The right steps must be taken for its disposal as the surface-disposed tailings destroy mining land resources and pose other environmental hazards, which eventually leads to limited cleaner production of the mining industry. Hence, recognized researchers provided a solution. They suggested recycling MPT as cemented paste backfill (CPB), which proved to be an ideal measure for the safe and environmentally friendly disposal of MPT.

CPB is a mine composite material produced using powder tailings, a hydraulic binder, and mixing water. It typically comprises of dewatered tailings (70-85% solids by weight), a hydraulic binder (3-7% by dry paste weight), and mixing water (fresh or mine processed). CPB will cater to the safe disposal of mine tailings. “It eliminates the need for constructing tailing dams at the surface, enabling the waste tailings to be effectively used to fill underground voids”. Other benefits of CPB include reduced surface subsidence and rehabilitation costs. Further, it can provide secondary ground support for mining operations to improve the underground working environment. Therefore, based on the varied advantages of CPB, it is being increasingly used in underground mining as it has technical, economic and environmental benefits. Based on a series of papers, the machine-learning aided design for cemented paste backfill has been validated on three major processes during the application of CPB.

The research paper titled, “Data-driven modeling of the flocculation process on mineral processing tailings treatment” aims to propose a data-driven method for modeling the flocculation process on mineral processing tailings treatment. The model was composed of gradient boosting machine (GBM) and firefly algorithm (FA), in which GBM was used for non-linear relationship modeling whereas FA was used for GBM hyper-parameters tuning. The modeling performance was evaluated and the relative importance of influencing variables was investigated. The supplementary study in this paper showed that there was room for improvement in the predictive performance of the GBM with the chemical characteristics of MPT being considered. Furthermore, the research team is hopeful for more accurate findings and a better representation of the MPT. This could be made possible by using more influencing variables such as particle size distribution (PSD) from laser scattering methods, the chemical compositions, and the temperature. In addition to the above, the need for more investigation of other types of flocculants and the dosing method on the flocculation process has been suggested too. Finally, the work is concluded by recommending the use of an enlarged dataset with more influencing variables. Further, a proposal to use more advanced machine learning (ML) algorithms to attain more accuracy has been suggested as well.

The research work, “An intelligent modeling framework for mechanical properties of cemented paste backfill” proposed an intelligent modeling framework for the mechanical properties prediction of cemented paste backfill (CPB).

Research work in the past was observed and it was seen that only a small number of studies have used machine learning (ML) algorithms to estimate the mechanical properties of CPB. For example, a study has proposed the uniaxial compressive strength (UCS) prediction model for CPB using boosted regression trees and particle swarm optimization. Although such studies have shown great promise, some shortcomings cannot be neglected. To compensate for the limitations in the literature, an intelligent modeling framework was proposed to accurately predict the mechanical properties of CPB. This was able to provide a fast estimation of CPB mechanical properties. Further, it reduced the time and cost allocated for the mechanical properties in the minerals industry to a great extent. Experiments were conducted and “based on the results, a user-friendly software package, named the intelligent mining for backfill (IMB) was developed in python programming for a wider application in the minerals industry”. It was also concluded that the recommended modeling framework and the IMB might be of great help for CPB design by saving time, reducing trial tests and cutting costs in the future.

The research study titled, “Pressure drop in pipe flow of cemented paste backfill: Experimental and modeling study” presents a framework for investigating and modeling the pressure drop of the CPB during pipe transportation with complex circuit shapes. The idea behind the study was to accelerate the efficiency of mining operations, which could be made possible if the pressure drop during the pipe transport of CPB can be predicted accurately. The need to design the transportation of fresh CPB from the surface plant to the underground voids was important for the mining industry. After trying multiple transport methods, it was found that the hydraulic transport of CPB through pipes is becoming more and more popular. This was because of the year-round availability, low technical maintenance, and its environment-friendly property. More studies revealed the constant need to investigate the pressure drop of fresh CPB during pipe transportation. In the present study, the team of researchers used a test loop system to examine the pressure drop of CPB under the influence of solids content, cement-tailings ratio, inlet pressure, and circuit shape. The pressure drop during the pipe transport of CPB was studied. In addition to this, it signifies “a fundamental change in the process of CPB pipe transportation by teaming researchers and practitioners with gradient boosting regression tree (GBRT) modeling for a reliable prediction of the pressure drop”.

The findings indicated that the pressure drop had a positive correlation with the solids content, cement-tailings ratio, and inlet velocity. It was also seen that the GBRT technique had great potential for pressure drop modeling. The combination of both loop test experiments and GBRT modeling proved to be an effective method to determine pressure drop, which could eventually be of monumental significance in engineering applications of pipe transportation. Observations also pointed out that such a modeling approach could provide fruitful benefits and provide the research and practitioner community with a reliable and cost-effective way for analyzing the pressure drop during engineering applications. However, the team has suggested continuing the quest for more influencing variables for the pressure drop. Besides, “other advanced AI techniques should be investigated “as well.

The three brilliant research papers, together with other papers about the machine-learning aided design for cemented paste backfill, advocate the potential of artificial intelligence (AI) in the sphere of mineral engineering. Prof. Chongchong, A/Prof. Qiusong and other team members are now calling for international collaborations on this topic and a database project has been initiated.

Moreover, Prof. Chongchong, A/Prof. Qiusong, and other collaborators also proposed innovative findings in recycling solid wastes, such as phosphogypsum, construction demolition waste, different ore tailings, smelting waste, etc., as materials for CPB. Research papers have been written and published in related journals like Journal of Cleaner Production, Construction and Building Materials, and Journal of Environmental Management. They are now running several special issues of SCI journals about CPB materials (details can be obtained upon contact).

Detailed information about the above contributions can be found at www.chongchongqi.com, which must be highly appreciated by the research community.

Media Contact
Company Name: Index of Sciences Ltd.
Contact Person: Jessica Simpson
Email: Send Email
Address:Kemp House, 160 City Road
City: London
Country: United Kingdom
Website: https://indexofsciences.com