Rock mass classification for predicting environmental impact of blasting on tropically weathered rock
Tropical climate and post tectonic impact on the rock mass cause severe and deep weathering in complex rock formations. The uniqueness of tropical influence on the geoengineering properties of rock mass leads to significant effects on blast performance especially in the developmental stage. Differen...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2020
|
| Subjects: | |
| Online Access: | http://eprints.utm.my/id/eprint/92368/1/BhatawdekarRameshMurlidarPSKA2020.pdf.pdf |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | Tropical climate and post tectonic impact on the rock mass cause severe and deep weathering in complex rock formations. The uniqueness of tropical influence on the geoengineering properties of rock mass leads to significant effects on blast performance especially in the developmental stage. Different rock types such as limestone and granite exhibit different weathering effects which require special attention for classifying rock mass for blastability purpose. Rock mass classification systems have been implemented for last century for various applications to simplify complexity of rock mass. Several research studies have been carried out on rock mass and material properties for five classes of weathered rock- fresh, slightly, moderately, highly and completely weathered rock. There is wide variation in rock mass properties- heterogeneity and strength of weathered rocks in different weathering zones which cause environmental effects due to blasting. Several researchers have developed different techniques for prediction of air overpressure (AOp), peak particle velocity (PPV) and flyrock primarily for production blast. These techniques may not be suitable for prediction of blast performance in development benches in tropically weathered rock mass. In this research, blast monitoring program were carried out from a limestone quarry and two granite quarries. Due to different nature of properties, tropically weathered rock mass was classified as massive, blocky and fractured rock for simpler evaluation of development blast performance. Weathering Index (WI) is introduced based on porosity, water absorption and Point Load Index (PLI) strength properties of rock. Weathering index, porosity index, water absorption index and point load index ratio showed decreasing trend from massive to fractured tropically weathered rock. On the other hand, Block Weathering Index (BWI) was developed based on hypothetical values of exploration data and computational model. Ten blasting data sets were collected for analysis with blasting data varying from 105 to 166 per data set for AOp, PPV and flyrock. For granite, one data set each was analyzed for AOp and PPV and balance five data sets were analyzed for flyrock in granite by variation in input parameters. For prediction of blasting performance, varied techniques such as empirical equations, multivariable regression analysis (MVRA), hypothetical model, computational techniques (artificial intelligence-AI, machine learning- ML) and graphical charts. Measured values of blast performance was also compared with prediction techniques used by previous researchers. Blastability Index (BI), powder factor, WI are found suitable for prediction of all blast performance. Maximum charge per delay, distance of monitoring point are found to be critical factors for prediction of AOp and PPV. Stiffness ratio is found to be a crucial factor for flyrock especially during developmental blast. Empirical equations developed for prediction of PPV in fractured, blocky, and massive limestone showed R2 (0.82, 0.54, and 0.23) respectively confirming that there is an impact of weathering on blasting performance. Best fit equation was developed with multivariable regression analysis (MVRA) with measured blast performance values and input parameters. Prediction of flyrock for granite with MVRA for massive, blocky and fractured demonstrated R2 (0.8843, 0.86, 0.9782) respectively. WI and BWI were interchangeably used and results showed comparable results. For limestone, AOp analysed with model PSO-ANN showed R2(0.961); PPV evaluated with model FA-ANN produced R2 (0.966). For flyrock in granite with prediction model GWO-ANFIS showed R2 (1) The same data set was analysed by replacing WI with BWI showed equivalent results. Model ANFIS produced R2 (1). It is found the best performing models were PSO-ANN for AOp, FA-ANN for PPV and GWO-ANFIS for flyrock. Prediction charts were developed for AOp, PPV and flyrock for simple in use by site personnel. Blastability index and weathering index showed variation with reclassified weathering zones – massive, blocky and fractured and they are useful input parameters for prediction of blast performance in tropically weathered rock. |
|---|