OR/17/037 Appendix 2 - Ince Marshes 1 (SJ47NE/100)

Hannis, S, and Gent, C. 2017. Petrophysical interpretation of selected wells near Liverpool for the UK Geoenergy Observatories project. British Geological Survey Internal Report, OR/17/037.

This appendix explains the well data available for the Ince Marshes 1 well and the details of, or additions to the general method explained in the section Method.

Figure 8    Well sketch and data acquired in the Ince Marshes 1 well.

The Ince Marshes 1 well was drilled in 2 main sections and wireline log suites run over the lower section. Almost 200 m of core was drilled over part of this section and 98 sidewall cores were recovered from the section beneath the full core. XRD analysis samples appear to be from depths corresponding to some of the sidewall cores and also depths beneath that, suggesting that the analysis must have also been on cuttings (rock chips brought up in the mud during the drilling).

Wireline services acronyms in Figure 8:
FMI = Formation micro-imaging
DSI = dipole sonic
ECS = elemental capture spectroscopy
HRLA = laterolog resistivity
PEX=platform express (neutron, density, gamma ray)
HNGS = spectral gamma ray

Data availability

(Figure 9) shows a summary plot of the data available for this well. Much more data is available for Ince Marshes 1 than for Kemira 1 (Appendix 1 - Kemira 1 (SJ47NE/101)). See Table 2 for comparison.

Digital PDF scans of company field prints are available for each of the wireline logging services. Digital data is also available for each tool suite as *.DLIS files.

A PDF scan of the company composite logs is also available, along with various reports. Excel sheets tabulating sample analysis data is also available.

Data loading and quality checks

The data was loaded and checked according to the summary in Data preparation.

Data load and input data quality summary notes
Data loaded from BGS Petrel–project *.LAS file and DECC *.DLIS file. Selected appropriate curves to use for the interpretation and ensured data available over all expected intervals. Ran TVD computation using reformatted, digitised PDF scan of TVD survey. Loaded sample analysis data after converting sample depths to metres and reformatting for IP load. Compared to core and sidewall core depths, also digitised and loaded from PDF report/log scans. Loaded Fault depths extracted from Petrel project from previous BGS FMI interpretation.

Data quality appears to be not too bad from CALI and HDRA curve evidence and curves responding as expected. A few washouts, rugosity (CALI spikes) and areas where HDRA is out of tolerance, in some cases corresponding to coals (e.g. around 620 m), but not always. CALI not open to maximum extent anywhere, i.e. pad tools not ‘floating’. Hole ovalised in top section above 434 m to casing shoe (378 m) (from FMI dual calipers, C1, C2) but data generally OK.

Interpreted output data quality check notes
ECS log and XRD, XRF sample data used to cross check input and interpreted curves and for parameter selection. V_COAL created porosity spikes where porosity is nulled over the coal bed.

1. Comparisons of core sample data with ECS log results and interpreted lithological volume curves (V_clay etc)

User calculations were written and implemented in IP to add XRD sample values together to make them more comparable to the ECS curve outputs and the interpreted lithology curves. The groupings and main comparisons are indicated in Table 11 by boxes and arrows respectively.

Table 11    Showing interpreted curves and corresponding XRD and ECS data used for comparison. (Arrows and boxes indicate where some minerals were grouped for comparison with combined curves).

The comparisons were made by eye, and just used as a cross-check for the data quality, precision and accuracy. When the interpreted curves were produced these were also used visually as a cross check to help calibrate the outputs, particularly for determining V_CLAY and V_COAL parameters. A plot of the comparisons is shown in Table 11. Some of the parameters selected are tabulated in Table 9.

2. High porosity spikes created adjacent to identified coal beds

In the case of the V_COAL parameters in the IP software these are either ‘coal’ (and all other curves are made zero) or ‘not coal’, rather than the wt% coal provided by the ECS tool processed outputs. Therefore, coal ‘kill’ parameters were adjusted to allow for this. Unfortunately this creates spike artefacts in the porosity around the coal picks, where the rock either side of a coal bed may also have a high percentage of coal and therefore appear to have a high porosity, whereas the adjacent coal bed is assigned a porosity of zero. This creates ‘spikes’ in the porosity log around the coal beds. Coal and salt ‘kill’ parameters are recorded in Table 9.

Petrophysical interpretation and results

The data was interpreted according to the summary in Data preparation and Petrophysical interpretation of lithology and porosity. Interpretation parameter sets are saved in the IP database and exported as *.set files (Appendix 5 - Technical information for BGS internal use).

Vclay interpretation notes
Used GR, Neutron-Density for the interpretation. Also displayed (but did not use) Son-Den, and XRD data points. XRD vol% data for all clays was summed (using user formula) i.e. the data points for chlorite, illite+mica, illite/smectite and kaolinite were summed.

The log was zoned initially according to the stratigraphic options and Neutron–Density cross plots were produced for each to select clay points. Max and min GR parameters were also selected for the same zones based on curve responses. Son-Dens cross plots were also selected for comparison. The HDRA curve was used to indicate intervals of poor or suspect density data and remove Vcl from the Neutron-Density output Vclay. Once best possible parameters were selected, all output curves and clay XRD data points matched satisfactorily by eye (i.e. ‘quick-look’). Deviations between N-D and GR derived Vclay curves towards the base of the log are caused by U content of basal sands (by examination of HGNS data) e.g. deeper than ~1418 m.

PHIE interpretation notes
InterpMultiMin set — density porosity model, U/Rho 3 mineral plot. Requires Pef, Rxo etc. OK match/cross check of log interpreted curves with sample & ECS data for clay (sample data used to help guide input params for clay), pyrite and coal. Sand appears underestimated in MultiMin interp cf ECS and sample data, and carbonates appear to be generally over estimated (although some samples have higher carbonate values than log interp predicts. Limestone-dolomite split not examined closely due to time constraints). See Figure 10.

InterpSonic set — sonic porosity model, to calibrate/cross check if SPOR ok to use where HDRA suggests density quality poor.

InterpND set — neutron-density porosity model. ‘Preferred’/best IP porosity output, plus most consistent to use between the two wells.

The presence of the XRD, XRF and ECS data, together with HGNS data means that the Ince Marshes 1 dataset could lend itself to probabilistic petrophysics methods, if a more detailed lithological study is required with uncertainty analysis. (Deterministic petrophyics means 1 possible solution for each set of parameters selected).

Figure 9    Ince Marshes 1 compressed well plot to broadly indicate data availability.

Figure 10    Ince Marshes 1 compressed well plot to show data comparisons: (XRD sample analysis vs ECS curves vs interpreted lithological curves).

Figure 11    Histograms of calculated TOC for the Westphalian A (grey, left) and Millstone Grit (green, right) for Ince Marshes 1.

Figure 12    A cross-plot of measured TOC against calculated TOC in Ince Marshes 1.