地下煤矿锚杆(索)过早失效腐蚀环境的水文地球化学模拟(1).pdf

J. Cent. South Univ. 2020 27 1599−1610 DOI https//doi.org/10.1007/s11771-020-4393-z Hydrogeochemical modelling of corrosive environment contributing to premature failure of anchor bolts in underground coal mines PENG Ya彭亚1, 2, Wendy TIMMS3 1. School of Civil and Resource Engineering, University of Science and Technology Beijing, Beijing 100083, China; 2. School of Minerals and Energy Resources Engineering, UNSW Sydney, Sydney 2052, Australia; 3. Faculty of Science Engineering hydrogeochemical modelling; rock bolts and cable bolts; stress corrosion cracking; water chemistry analysis Cite this article as PENG Ya, Wendy TIMMS. Hydrogeochemical modelling of corrosive environment contributing to premature failure of anchor bolts in underground coal mines [J]. Journal of Central South University, 2020, 275 1599−1610. DOI https//doi.org/ 10.1007/s11771-020-4393-z. 1 Introduction Anchor bolts including rock bolts and cable bolts are predominant components of the supporting system in underground mining projects [1−3]. However, frequent premature failures of these anchor systems in rock strata saturated with groundwater have been observed as a significant safety risk in mine operations around the world. These failures were not only restricted to rock bolts or cable bolts that had been in service for a long time. Bolts installed within one year or even less have also been suffering from premature failures [4−6]. Increasing interest has been shown in identifying these premature failures by considering stress corrosion cracking SCC mechanism [4, 7−13], which has also been widely recognized in other applications such as marine corrosion and petroleum pipeline corrosion [14−17]. According to Foundation item Project140100153 supported by Australian Research Council Linkage Grant Received date 2019-08-10; Accepted date 2020-02-17 Corresponding author Wendy TIMMS, PhD, Professor; Tel 61-3-5227-8692; E-mail wendy.timmsdeakin.edu.au; ORCID 0000- 0002-6114-5866 Open Science Identity OSID J. Cent. South Univ. 2020 27 1599−1610 1600 the SCC mechanism, there are three dominant factors that induce premature failures which include susceptible material, a corrosive environment and tensile stress. Mining-induced stress, the presence of groundwater and mineral environments are reported to be contributing factors to SCC failures of rock bolts and cable bolts [18]. There have been numerous laboratorial experiment studies investigating the influence of these factors on SCC of rock bolts and cable bolts by simulating underground in-situ conditions. Some experiments drew attention to the possible groundwater influence on inducing SCC of bolts under load conditions. Bolts were immersed in mine water alone, and these experiments found no SCC phenomena after one year [4] or after 3.5 years [19]. Other experiments were conducted focusing on the influence of both groundwater and surrounding mineral materials, and then SCC failures were identified during the test period [20], indicating that more complex factors might be involved in the SCC process. Additionally, to simulate SCC failures, accelerated s were adopted in laboratory testing by using more aggressive solutions or introducing microorganisms such as sulphate reducing bacteria SRB, and SCC failures were also observed in the experiment [10, 21]. In addition to the experimental study on SCC of rock bolts and cable bolts, there has already been some modelling work pered on SCC in other regions by using microstructural peridynamic models and finite element models [22, 23]. However, these models only focused on the mechanical behavior of the material during the stress cracking process. There would have been a relatively long period of corrosion process prior to the onset of stress cracking. To clarify the relation between these two processes, hydrogeochemical s can be a useful way to throw light on the research gap. Hydrogeochemical s have been adopted for research on a wide range of topics, such as mine drainage and water pollution [24, 25], aquifers and aquitards [26−28], and deep ation water [29, 30]. Only a few studies have applied hydrogeochemical knowledge to corrosion processes underground and including diffusion [30, 31]. However, they were dealing with corrosion problems under different circumstances than anchor bolt corrosion and failure. For example, PEA et al [31] built hydrogeochemical models to study the corrosion rate of carbon steel canisters used for the geological disposal of high-level radioactive waste. KLAPPER et al [30] focused on modelling the corrosion rate of scaling and corrosion influencing technical facilities during geothermal energy production. Considering this history, the current study is a valuable initial attempt to apply hydrogeochemical modelling in studying anchor bolt corrosion and failure issues. In this paper, a series of thermodynamic equilibrium models combined with laboratory experiments were developed for investigating factors that contribute to anchor bolt corrosion both in the laboratory experiment and underground coal mines. Water chemistry data of groundwater released by dripping rock bolts were used to simulate anchor bolt corrosion. Reactions under aerobic and anaerobic conditions were discussed, and pH and pE trends were analyzed in comparison with experimental data. In addition, it was identified that microbial activities could influence corrosion process that may further lead to SCC failures. This research is an interdisciplinary study with regard to the underground bolt corrosion environment under aerobic and anaerobic conditions, which ultimately aims to develop corrosion inhibitor strategies to prevent SCC failures of anchor bolts, and hence improving both mining safety and productivity. 2 s and setup 2.1 Rock and cable bolts corrosion experimental s and setup Testing of rock and cable bolts was conducted within a variety of material packings as shown in Table 1 and Figure 1 including saturated clay, saturated coal, mine water and grout/cement. Rock bolts and cable bolts used in these experiments were the same types utilized at a longwall underground coal mine in the Sydney Basin, Australia. The mill scale layer from every bolt was removed prior to the experiments. This mill scale, which is usually only 0.1 mm in thickness, is designed to act as a cathodic protection layer to provide an initial barrier to atmospheric corrosion. The work to remove it was pered by professionals at the Jennmar factory in Sydney using industrial sanding equipment. The type of rock bolt used was HSAC 840 ribbed bolt with a J. Cent. South Univ. 2020 27 1599−1610 1601 Table 1 Experimental design with different material packings Tube* Bolt type Packing 1-RBCY Rock bolt Clay b Frame with packed Perspex tubes J. Cent. South Univ. 2020 27 1599−1610 1602 Figure 2 Water sampling and testing s a Water sampling container; b Water filter and 45-micron filter paper; c HACH HQ40d water multi meter the default thermodynamic PHREEQC.DAT database, which includes essential thermodynamic data for aqueous species, gas and mineral phases. Microbial activity also contributes to SCC failures of anchor bolts, according to previous research showing the presence of microbial species including SRB such as Desulfovibrio in the bolt environment [20, 34]. In fact, SCC of anchor bolts may also be a special type of microbiologically influenced corrosion MIC, which is a process of electrochemical reactions [35−37]. Electron transfer must occur during the corrosion process. To achieve this electron transfer process, organic carbon and iron solid were represented by stepwise addition of carbon and iron elements with zero valences in the modelling, both of which were defined in the REACTION block. The ratio of carbon and iron addition was 31 by moles in each step, determined by a trial of different ratios. The value of elemental iron was set to approximate the mass loss of the bolts during experiments, and the distribution of total iron over Fe2 and Fe3 was calculated from pE values [38]. The redox reactions in the anaerobic condition would be driven by the of organic carbon and iron. The organic carbon was oxidized to inorganic carbon, increasing the electron activity in the solution, and then the electrons were utilized by other components such as sulphate. Iron was oxidized during corrosion of bolts, with the electrons available to reduce other species such as hydrogen ions. As oxygen was observed throughout the experiment, aerobic conditions positive DO were considered in the modelling. However, local anaerobic conditions could develop around the bolt surface due to biofilms and clay materials surrounding bolts. The initial chemical compositions in the models were specified based on water chemistry measurement of the initial mine water used in the experiment. To increase the reliability of water analytical data, the alkalinity value was calculated to achieve a zero-charge balance of ions in water. These data are shown in Table 2. The initial O2 value for mine water was 8.5 mg/L. However, to identify possible anaerobic process, it was also set as 0 or 3 mg/L in some models. The major elements involved in the bolt corrosion process, C, S, Fe, O and H, are affected by redox processes, dissolution equilibrium, and corrosion rust deposition [27, 31, 39−41]. In Table 3, equations of primary reactions Table 2 Initial chemistry of mine water from a dripping bolt Parameter Value pH 8.1 Temperature/C 24 Na/mg∙L−1 168 K/mg∙L−1 21.4 Mg/mg∙L−1 10.4 Ca/mg∙L−1 12.2 Cl/mg∙L−1 7 S/mg∙L−1 92 Fe/mg∙L−1 0.075 O2/mg∙L−1 8.5 Alkalinity as HCO3−/mg∙L−1 440 J. Cent. South Univ. 2020 27 1599−1610 1603 Table 3 Dominant reactions for main elements in solution Element Reaction Comment C* 2SO42−→2HCO3−HS−H HCO3−H→H2OCO2g Organics oxidized and sulphate reduced anaerobic S SO42−9H8e→HS−4H2O HS−H→H2Sg HS−Fe2→FeSsH Sulphate reduced anaerobic Fe Fes→Fe22e Fe2→Fe3e Fe2HS−→FeSsH Fe2HCO3−→FeCO3sH Fe22OH−→FeOH2s Iron oxidized at steel surface aerobic Rust deposit O O22H2O4e→4OH− Oxygen and water reduced aerobic H 2H2e→H2g Hydrogen ions reduced Note represents the available biodegradable organic matter, shown as carbon with the oxidation state of zero. that may occur during bolt corrosion process were listed, representing reaction types of organic matter degradation, sulphate reduction, iron oxidation, precipitation equilibrium, and cathodic reaction under aerobic or anaerobic conditions. The specific reaction process would be decided by thermodynamic calculations based on the PHREEQC.DAT database. 3 Results and discussion 3.1 Experimental results and discussion After 76 d of testing, visual uation showed general corrosion on almost all the bolt specimens. Corrosion was more obvious in the upper regions where more atmospheric oxygen could diffuse into the water of the packed sample. Pitting corrosion was found on bolt specimens surrounded by coal or coal and clay. Cable bolt specimens in particular showed crevice corrosion, which was probably as a result of the cable bolt specimen being composed of multiple cable bolt strands. It should be noted that no cracks were visible on any bolt specimens, which was attributed to the absence of tensile stress and the short period of testing. In other experiments as part of the broader ARCL research program, the influence of tensile stress on SCC has been identified and cracks have been identified [10, 42]. Water chemistry varied between each experimental sample, and over time during experiments. Water chemistry of the initial mine water and six final water samples are shown in Table 4. A piper diagram was plotted to show the relative abundance of major ions present and also to group the water samples Figure 3. Six samples were Na-HCO3 type water, while the sample 2-RBCL-b was Na-Cl type. The sulphate values declined in most of the samples, except the one in 2-RBCL-b showing an obvious increase in concentration. Concentrations of other ions including Ca2, Na, Mg2, Fe2 and Cl− in 2-RBCL-b were also higher than those from other measured samples. Additionally, the final pH of this sample was 8.0, being the lowest value among all seven water samples, while the EC value at 2.67 mS/cm was the highest. This water sample was collected from the lower region about 100 mm above the tube bottom, where coal was tightly packed around the bolt. The notable difference between the water chemistry of this sample and the others could result from the heterogeneous composition of coal, leading to substance Table 4 Water chemistry of water samples Sample K/ mg∙L−1 Ca2/ mg∙L−1 Na/ mg∙L−1 Mg2/ mg∙L−1 Fe2/ μg∙L−1 Cl−/ mg∙L−1 NO3−/ mg∙L−1 SO42−/ mg∙L−1 Alk as HCO3−/ mg∙L−1 TDS calculated/ mg∙L−1 Water type pH EC/ mS∙cm−1 DO/ mg∙L−1 0-initial 21.4 12.2 168 10.4 74.6 7 N/A 92 440 311.1 Na-HCO3 8.1 0.92 8.5 2-RBCL-a 37.1 2.6 203 0.9 12.2 30 4.7 70 467.6 348.3 Na-HCO3 10.6 1.11 8.7 2-RBCL-b 21.7 37.9 392 47.5 22.5 484 N/A 285 224.9 1268.1 Na-Cl 8.0 2.67 9.1 3-RBCC-a 31.9 4.3 143 3.2 170 32 1.9 88 290.8 304.4 Na-HCO3 10.0 0.93 7.3 3-RBCC-b 21.6 15.9 131 10.6 120 60 N/A 62 300.2 301.2 Na-HCO3 8.2 0.95 6.8 4-RBWG-b 28.1 9.3 122 13.2 9.5 11 N/A 104 311.6 287.6 Na-HCO3 8.2 0.77 9.2 6-CBCL-a 28.6 4.9 154 6.1 7.1 14 N/A 89 385.9 296.6 Na-HCO3 9.4 0.89 8.3 Note  “0-initial” was data for initial mine water; “a” represented water sample collected at the upper region while “b” was from the lower region.  “N/A” indicated the value was below the detection level.  The alkalinity value Alk was calculated given zero charge balance of the water.  Total dissolved solids TDS was calculated by the sum of cations and anions, not including the alkalinity value as HCO3−. J. Cent. South Univ. 2020 27 1599−1610 1604 Figure 3 Piper diagram of water samples dissolution including sulphur compounds and chlorides from coal material into limited amount of water [43]. There might also be more complex redox reactions here because of the lowered pH value and higher sulphate concentration. Concentration of iron ions in water samples from those tubes filled with coal or grout showed a decrease, while in contrast, iron ions clearly increased in clay filled samples. In tube 2-RBCL-a and tube 3-RBCC-a, distinct increases of nitrate occurred from below detection to 4.7 and 109 mg/L, respectively. These increases were probably attributable to dissolution and degradation of organic matter [27]. Continuous measurements showed pH of water samples from the upper tube generally increased during the experim