Two software codes from NIOSH that are in the news. They were used to analyze conditions at the Crandall canyon Mine prior to the disaster that killed 6 miners and 3 rescue workers.

The first is Analysis of Retreat Mining Pillar Stability (ARMPS). The NIOSH site describes the software thus:

ARMPS calculates stability factors based on estimates of the loads applied to, and the load-bearing capacities of, pillars during retreat mining operations. The program can model the significant features of most retreat mining layouts, including angled crosscuts, varied spacing between entries, barrier pillars between the active section and old (side) gobs, and slab cuts in the barriers on retreat. It also features a pillar strength formula that considers the greater strength of rectangular pillars. The program may be used to evaluate bleeder designs as well as active workings.

The second is Stress and Displacement Calculations (LAMODEL). The NIOSH site describes the software thus:

LAMODEL is software that uses boundary-elements for calculating the stresses and displacements in coal mines or other thin, tabular seams or veins. It can be used to investigate and optimize pillar sizes and layout in relation to pillar stress, multiseam stress, or bump potential (energy release). LAMODEL simulates the overburden as a stack of homogeneous isotropic layers with frictionless interfaces, and with each layer having the identical elastic modulus, Poisson's Ratio, and thickness. This "homogeneous stratification" formulation does not require specific material properties for each individual layer, and yet it still provides a realistic suppleness to the overburden that is not possible with the classic, homogeneous isotropic elastic overburden used in previous boundary element formulations such as MULSIM or BESOL.

Clearly you have to know what you are doing to use these computer codes. Just consider what went wrong at the Crandall Canyon Mine. Here is an extract (edited for context) from the Report on the August 6, 2007 Disaster at the Crandall Canyon Mine—see pages 22 and following:

In performing geotechnical evaluations of Murray Energy’s plans to mine the barrier pillars, Agapito used the computer models LAMODEL and ARMPS to calculate mine stability factors. Agapito documented its work in two reports. The first report addressed the development of the initial four entries in the North Barrier and the second addressed retreat mining in the South Barrier. In an email Agapito detailed the ARMPS and LAMODEL results for planned retreat mining in the North and South Barriers. Together, these three documents were used by Murray Energy to support its proposals to develop and retreat mine the North and South barrier pillars of Crandall Canyon’s Main West section.

In a report entitled “Evaluation and Control of Coal Bumps,” mining engineers at NIOSH found serious flaws in Agapito’s analyses. NIOSH’s experience in this substantive area is vast. For “deep cover” mines like Crandall Canyon, NIOSH conducted a “special research project” in 1997 in which “97 panel design case histories were gathered at 29 mines located in 7 states….more than 40% of the case histories, including half of the bumps, were from coal mines in UT and CO.”

NIOSH’s primary critique of Agapito’s ARMPS analysis is that it substantially overstates the strength of the remnant barrier pillars left between the newly developed entries and the gob. NIOSH engineers wrote in their report that Agapito’s analysis failed to distinguish between barrier pillars – which remain between the mined-out longwall areas and the entries – and production pillars, which are pulled during retreat mining. This failure resulted in overstatement of the ability of the remnant barrier to support the load that would be transferred during Murray Energy’s retreat mining.

The result was tragedy. Lawyers will have to establish if this is stupidity, negligence, or fraud.

Agipito has stated their case. Again I quote (and edit) for the report (pages 34 and following.)

NIOSH took issue with Agapito’s use of a high coal strength quantity in their LAMODEL analysis, as opposed to the more conservative default coal strength quantity of 900 psi. In a submission to MSHA, Agapito states that the 1,640 psi coal strength value was not derived from tests of actual field samples, but was extrapolated from prior successful retreat mining operations. Agapito officials did not actually observe the areas of prior retreat mining, but rather relied upon company officials’ descriptions of them. In a recent letter to MSHA, Agapito further defends its use of the higher coal strength quantity by citing laboratory tests showing values as high as “4,512 psi.”

So it all boils down to a coal strength of 900 versus 1,640 versus 4,512 psi. Agipito used 1,640 and argues for 4,512. But we know it more likely was 900. Whatever happened to sensitivity analyses as a standard part of any calculation?

Seems that was left to MSHA. From pages 39 following of the report:

Using a certified map of the mine, Pete Del Duca [an engineer with MSHA] reran ARMPS to test Agapito’s work. Del Duca’s independent analysis found that complete pillar extraction of both Barrier Pillars, as the company proposed, was unsafe. Del Duca recommended that Murray Energy’s retreat mining plan (“pillaring” plan) be rejected. Del Duca’s analysis was prescient – his report found that retreat mining could be safely conducted only from crosscuts 163 to 149 in the North barrier and from crosscuts 149 to 142 in the South barrier.54 In March, the North barrier collapsed during retreat mining at crosscut 133, and, in August’s deadly tragedy, the South barrier collapsed at crosscut 139, both beyond the points Del Duca predicted would be safe.

Del Duca submitted his calculations to his boss, who discussed them with Murray Energy officials. Del Duca’s boss, Billy Owens, told Del Duca “that his analysis was flawed, that Agipito was correct and that mining would continue.” Which it did until everything fell down. Point is: do not rely on computer codes; and do not rely on regulators to check or approve your calculations. Particularly if you are pressuring them.