Proportion optimization and strength prediction of CGS backfill materials based on GA-ELM mode

ABSTRACT

The strength of the backfill body is an important aspect of safe production in backfill mining technology. In order to quickly and accurately obtain the strength of coal gasification slag -based cemented backfill materials with different ratios, the orthogonal design of 4 factors and 5 levels and variance analysis method were used to study the mass concentration (A), coal gasification slag (CGS)/cement (B), bone glue ratio (C), activator content (D) on the compressive strength of backfill body and a genetic algorithm with 28-day compressive strength of CGS-based backfill material as the response value was constructed, and verify the applicability of the model. The results show that: the primary and secondary factors affecting the 28-day compressive strength are C > A > B > D, the optimal ratio is 80% mass concentration, CGS/cement ratio 4:1, bone glue ratio 3:7, and activator 3%, according to orthogonal variance analysis; the activity of CGS is gradually stimulated as the curing age grows, the hydration products Ettringite and C-S-H are continually generated, and the compressive strength of the backfill material increases in tandem; In the GA-ELM model test set, the intensity correlation coefficient between the prediction and the measured value reached an average of 0.99, and the prediction accuracy was 71.8% and 43.5% higher than that of ELM and PSO-ELM respectively; Laboratory studies have shown that the GA-ELM prediction model has a prediction accuracy of 98.2% and that it can accurately estimate the strength of CGS cemented backfill.

KEYWORDS:

• Solid waste recycling
• backfill material
• CGS
• strength prediction
• extreme learning machine
• orthogonal experiment

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The work was supported by the National Natural Science Foundation of China [51874229, 51974225, 51904224 51904225, 51704229]

Notes on contributors

Yonglu Suo

Yonglu Suo, born in 1960 in Baoji, Shaanxi Province, is a professor and doctoral supervisor. He has been committed to the research of coal mining methods for a long time, and has carried out a lot of research work in the fully mechanized caving mining technology of hard and thick coal seam, safe and efficient technology of working face, etc. He has won 1 provincial science and technology progress grand prize, 1 first prize, 1 second prize, 3 national invention patents, published a monograph and more than 90 academic papers.

Caixin Zhang

Caixin Zhang, born in 1996, is a master student at Xi 'an University of Science and Technology. Research direction: Solid waste treatment - material ratio optimization.

Lang Liu

LangLiu, Ph.D. (back), professor, doctoral supervisor of the project, co-trained by the University of Western Australia, has won the National Coal Youth May Fourth Medal, the Huo Yingdong Young Teacher Award of the Ministry of Education, the Shaanxi Youth Science and Technology Award, and was selected as the Shaanxi Youth ”100 Talent Plan” and the National University Outstanding Young scientific and technological talents in the field of Mining, petroleum and Safety engineering.

Huisheng Qu

Huisheng Qu, Xi 'an University of Science and Technology, Doctor, research direction solid waste treatment, has published one EI paper, two SCI papers.

Pan Yang

Pan Yang, Xi 'an University of Science and Technology, Doctor, research direction solid waste treatment, has published 6 SCI papers.

Geng Xie

Geng Xie, Xi 'an University of Science and Technology, Doctor, research direction solid waste treatment, has published 3 SCI papers.