Roxar Software Training: Advanced Petrophysical Modelling PDF

1 Reference Manual Advanced Petrophysical Modelling 2 Contents 1 INTRODUCTION ....................................................................................... 3 1.a Petrophysical Modelling dialog box ....................................................................... 4 2 GENERAL TA...

1 Reference Manual Advanced Petrophysical Modelling 2 Contents 1 INTRODUCTION ....................................................................................... 3 1.a Petrophysical Modelling dialog box ....................................................................... 4 2 GENERAL TAB ........................................................................................ 5 2.a 2.b 2.c 2.d 2.e 3 Using a Facies parameter ................................................................................ 5 Seismic Cosimulation ....................................................................................... 6 Algorithm .......................................................................................................... 7 Data analysis on input data .............................................................................. 8 Simulation settings ........................................................................................... 9 DISTRIBUTIONS TAB .............................................................................10 3.a 3.b 3.c 3.d 3.e User mode...................................................................................................... 11 Kriging Info ..................................................................................................... 11 Modelling mode .............................................................................................. 12 Declustering ................................................................................................... 12 Transformation Types..................................................................................... 12 3.e.i Basic ...................................................................................................................... 13 3.e.ii Skewness ............................................................................................................... 14 3.f Geological trends ........................................................................................... 15 3.f.i Intrabody trends ..................................................................................................... 17 3.f.ii 3D trends................................................................................................................ 18 3.g Transformation Sequence .............................................................................. 19 4 5 CORRELATION TAB ...............................................................................21 VARIOGRAMS TAB ................................................................................22 5.a The Variogram................................................................................................ 22 5.b Definition in the Variograms tab ...................................................................... 23 5.b.i Previewer ............................................................................................................... 23 5.b.ii Anisotropy .............................................................................................................. 24 5.b.iii Standard deviation (Variogram Sill) ....................................................................... 24 5.b.iv Variogram model .................................................................................................... 24 5.b.v Anisotropy directions .............................................................................................. 25 5.b.vi Range ..................................................................................................................... 27 5.c Using Data Analysis to define the variogram .................................................. 27 5.c.i Estimation of azimuth ............................................................................................. 28 5.c.ii Estimation settings ................................................................................................. 28 5.c.iii Variogram Modelling .............................................................................................. 29 6 LOCAL UPDATE TAB .............................................................................31 6.a Local Update .................................................................................................. 31 7 WATER SATURATION MODELLING .....................................................32 7.a General tab .................................................................................................... 32 7.b Variables tab .................................................................................................. 33 7.b.i Functions................................................................................................................ 33 7.b.ii Variables tab .......................................................................................................... 34 RMS Training Manual, Property Modelling 3 1 INTRODUCTION Petrophysical modelling concerns the distribution of reservoir the properties that are commonly used to calculate volumetrics (porosity & SW) and also as input for reservoir simulation (permeability). It is important to remember that a substantial amount of this work has probably already been done in the facies model. If that is geologically realistic and the facies themselves characterise good/medium/poor reservoir, then the job of the petrophysical model is to add detail to the facies model. This can be done in several ways  A deterministic or predictive approach  Stochastic simulation  By conditioning to wells, seismic data, facies  By incorporating trends and correlations This manual mainly concerns stochastic simulation of petrophysical properties, as that is generally considered to be best practice for their distribution, giving more realistic reservoir simulation results in the dynamic modelling phase. For quick look volumetrics, if a good facies model has been created, then simple interpolation or even averaging can give a good enough estimate, but for models that will move to simulation, the stochastic approach, with good data analysis and understanding, is the option to choose. The basic workflow for modelling a petrophysical parameter is as follows: 1. Analyse well logs and use geological concepts to define horizontal and vertical trends in the data 2. Remove the trends as part of the data transformation process, which aims to represent the data as a Gaussian distribution (also called the Residual component) 3. Define correlations between pairs of data, either logs being simulated together and/or parameters being conditioned to other (e.g. seismic) data 4. Define variogram models, which quantify the spatial continuity and variability of the data 5. The Residuals are then distributed using all input data, and automatically transformed back (step 2 in reverse) to give a parameter in original units The diagram below summarises this workflow: RMS Training Manual, Property Modelling 4 1.a Petrophysical Modelling dialog box The workflow is performed in a sequence of tabs which comprise the petrophysical modelling dialog:  General – general settings and algorithm  Distributions – transformation sequence, including geological trends  Correlations – if more than one parameter is modelled, or seismic data is to be conditioned to, correlations can be defined to link the parameters  Variograms – variogram settings to define the spatial continuity/variability  Local update – when new wells are input into a previous model The workflow is detailed in the remaining sections. RMS Training Manual, Property Modelling 5 2 GENERAL TAB Fits the data to the wells. If not ticked, the wells will be used to define the simulation values only. Select the log(s) to simulate. Add... is used to add parameters that are not related to a well log. Lists all continuous logs found in blocked wells Condition to seismic, or other parameter This allows you to use predefined DA objects – see section DA object created for each output parameter. Useful for QC Parameters will be modelled separately in each facies Fits the data to the wells. If not ticked, the wells will be used to define the simulation values only. 2.a Using a Facies parameter A Facies model is an essential element of any petrophysical model in a heterogeneous reservoir. Facies modelling allows geological concepts to be imparted in 3D space and this makes the subsequent distribution of petrophysical properties more robust. So if you have a facies parameter, it always makes sense to use it. RMS Training Manual, Property Modelling 6 The figure shows a channelised system, within which porosity is distributed with respect to distance from the channel margins. The inter-channel facies shows poor porosity in the log and thus in the final realisation Different Facies have different:  Mean values  Distributions (Trends)  Variability  Variograms As an alternative to a facies parameter, you can select a different discrete parameter, such as a region index parameter, for the current grid. This allows you to specify different model settings for (for example) different fault segments. If you want to use both a facies parameter and different settings for different fault segments, you need to use IPL to create a combined discrete parameter with one value for each facies/segment combination. 2.b Seismic Cosimulation Seismic cosimulation involves a seismic parameter (or other parameter indicating a relationship with the modelled parameter) being used to condition the petrophysical model. A correlation coefficient is set in the Correlation tab between the two variables, as in the example below: Acoustic impedance cube Cosimulated porosity This differs from classic cosimulation, in which more than one log parameter (but no 3D parameter) is chosen to be modelled, with a correlation (also set in the Correlation tab) set between them. Seismic cosimulation can also be used to allow a pre-defined parameter (e.g. porosity) to condition another (e.g. permeability). RMS Training Manual, Property Modelling 7 2.c Algorithm Two algorithms are available: Prediction (or kriging) is a mean value description, relying on a measure of spatial continuity encapsulated in a variogram. Mathematically, it is the expectation of the Gaussian Random Field. It is locally accurate and smooth in appearance. Simulation aims at a more realistic global representation of geological heterogeneity. It is stochastic (noisy) in nature, globally accurate, with many equally probable outcomes Both techniques can condition to well data Prediction (kriging) Simulation In practice, we generally use the simulation algorithm, with conditioning to wells handled by kriging as part of the algorithm. The 3 cross sections shows the difference in outcome between the 2 approaches RMS Training Manual, Property Modelling 8 2.d Data analysis on input data The settings in the Distributions, Correlations, and Variograms tabs are set based on a proper analysis of the input or on general geological knowledge gained from studies of reservoirs. You can specify the settings for the modelling in any of the following ways:  Generate or update the settings manually, using the buttons and fields in the appropriate tab; many settings can be estimated from the input data + Copy a previously defined specification in the appropriate tab  Use settings you have previously identified using data analysis. Either: o Update the whole job from an attached MVA using the option in the General tab (see Data Analysis on Input Data). The relevant data is copied to the other tabs. o Update the current tab only, for the current zone/facies/parameter selection (as appropriate), from the matching components of an attached MVA (see, for example, Defining the Transformations From Data Analysis) Ticking on this option allows you to use a pre-defined data analysis object to define the Distributions, Correlations & Variograms for the petrophysical simulation. There are pros and cons to this approach, but essentially, a simple project doesn’t usually merit the creation of DA objects, whereas the flexibility it grants to more complex cases is worth the effort. Pros:  Zones can be grouped together when working with a multizone grid to allow data to cover more than one zone. This cannot be done without a DA object.  DA from one zone can be used in another zone  Analogue data can be set up in a DA object and used as input  Variogram analysis can be performed in a DA object, but not directly in the petrophysical modelling job  Many have also used this option to include a DA object which is generated from the raw log data, with the requirement that the full data range is modelled. However, this is inadvisable as cells in a model should contain effective properties for the actual cell dimensions chosen for the grid. Using raw data imparts smaller scale heterogeneity that does not exist at grid cell scale Cons:  If a DA object is used, the Distributions, Correlations & Variograms can still be changed manually in their respective tabs. This may create confusion with project workflow/documentation  It takes slightly longer RMS Training Manual, Property Modelling 9 2.e Simulation settings Kriging neighbourhood : this is the number of data used in the kriging neighbourhood. Increasing the number increases the quality of the result, but slows down the simulation. The Seismic output option is available only if seismic cosimulation is selected. If the seismic parameter is incomplete (has holes or covers only a partial area) selecting this option will generate a new seismic parameter is generated in which undefined cells (i.e. have a value of 999) are simulated. You can choose to overwrite the existing parameter or create a new one. RMS Training Manual, Property Modelling 10 3 DISTRIBUTIONS TAB The main function of the Distributions tab is transform the input data into a Gaussian or ‘Normal’ distribution, prior to simulation, then transform the Gaussian distribution back to the input data after simulation. This is because simulation algorithms run in the Gaussian domain. So what does Gaussian mean? A Gaussian distribution has a mean of 0 and a Standard Deviation of 1 Of course, it would be rare to have a completely normally distributed set of data, particularly where we have few wells. In addition, trends in the data (spatial, depth etc) also affect the distribution and these have to be removed during the transformation. Original data RMS Training Manual, Property Modelling Transformed data 11 3.a User mode There are 3 user modes, the selection of which should depend on the extent of the input data and the complexity of the reservoir. The following table serves as a guide. The input facies model also has to be taken into account. If it is fairly detailed, geologically realistic, and with facies classifications that honour differentiation on petrophysical properties, then much of the data analysis and spatial distribution work has in fact already been done. The remainder of this section will focus on ‘Advanced mode’, which gives you the full range of options. 3.b Kriging Info The Kriging info... button will detail the type of kriging used, which is selected automatically by RMS depending on the User mode and whether trends and/or seismic is used. The panel below RMS Training Manual, Property Modelling 12 shows the Kriging selection for Advanced mode for Automated Parameter Estimation. Information for all modes can be accessed in the same panel by toggling the relevant options. 3.c Modelling mode The mode can be switched to ‘Constant’ if you want to specify a constant value for e.g. one facies. Otherwise, prediction or simulation of the input data (whatever has been selected in the General tab) will be used. 3.d Declustering Declustering is used when the data (i.e. wells) are concentrated in particular areas, which is a common scenario. The data effectively represents sub-samples of the whole data set and spatial correlation assumes that nearby well data will not be independent. In practical terms, wells are purposely drilled in areas of better reservoir quality, so it is likely that the well data will give a better than average summary compared to the whole reservoir. Declustering is a means to account for this clustering of the data and allows a weighting to be attached to each data point. 3.e Transformation Types There are 6 transformation categories RMS Training Manual, Property Modelling 13 3.e.i Basic Transformations Truncate data This will truncate the input data and is useful if there are some spurious max or min values which are not commensurate with the facies character. Truncate realisation All input data is used as input to the simulation, but the resulting parameter is truncated (i.e. values lower than min set to min and values above max set to max) Truncate data and real. Both the input data and realisation are truncated using the same cut-off(s) Mean (Constant) Selecting both is equivalent to ‘Truncate data and real.’ Set cut-off for low end, high end or both. Same cut-offs used for input data and truncate realisation. The transformation is shown in the summary as: RMS Training Manual, Property Modelling 14 3.e.ii Skewness If the data is roughly symmetrical, then no skewness transformation need be applied. If skewness is a factor, the following are available: Transformations Logarithm Use this option if the distribution has a lognormal skew. Permeability often follows a lognormal distribution. Log Transform Square root ? Box-Cox ? Box-Cox Transform Normal score The data is mapped onto a Gaussian cumulative probability function, by subtracting the mean and normalising by the standard deviation. Data should contain at least 200 values and have no large peaks or holes. If normal score is used it should be the last RMS Training Manual, Property Modelling 15 step in the transformation sequence 3.f Geological trends Transformations Compactional depth trend Petrophysical properties which change with depth due to diagenetic processes related to TVD. Commonly, porosity and permeability tend to decrease with depth due to compaction and cementation Compactional depth Transform Depositional depth trend Trends resulting from deposition tend to be related to stratigraphic depth (simbox), e.g. fining upward sequences 1D Lateral trend Trend in world coordinates (e.g. UTM) 2D Lateral trend General surface trend (map) RMS Training Manual, Property Modelling 16 The Edit dialog for the two depth trends allows you to change the slope and datum, as indicated below. RMS Training Manual, Property Modelling 17 3.f.i Intrabody trends An intrabody trend defines a change in petrophysical property within an individual facies body. This requires that a body parameter is produced as output during the facies modelling job. Porosity simulation using a lateral intrabody trend to honour good porosity close to sediment source and poorer porosity with increasing distance from source RMS Training Manual, Property Modelling 18 Reference is available for lateral trends only 3.f.ii 3D trends Transformations General 3D trend This uses an existing continuous parameter to describe the ‘expected’ value in each cell. A scaling coefficient can be used to re-scale the parameter to match the value range of the parameter to be modelled . RMS Training Manual, Property Modelling 19 Cloud transform The 2D (bivariate) probability density function (pdf) for the 3D parameter and the well log parameter is calculated using the Blocked Well values for the well log parameter and the co-located values of the 3D parameter (at the well locations)  Input = well data + 3D parameter  Output = 3D parameter  Y axis = the parameter to be modelled (e.g. permeability)  X axis = the explanatory parameter which has data (e.g. porosity, AI)  The relationship between the parameters is not straight forward to establish. Cloud transform is an easy to use solution 3.g Transformation Sequence Workflow. The general sequence in which data is transformed is as follows: 1. Truncate the input data (max and/or min) to remove outliers or extreme values that may be erroneous 2. Incorporate geological trends 3. Incorporate skewness 4. Correct the residuals to give mean = zero The residual component is what cannot be explained by geological features (i.e. a stationary field). Note: it is common not to have a perfect histogram after this process! In Advanced mode, a default sequence is automatically given which truncates the data (Truncate), shifts it (Mean) and removes skewness. When working with transformations, especially when using manual mode, the Distribution residuals (data distribution after listed transforms have been applied) should be shown to check final distribution to Normal Score. It is also a useful tool to learn how each transform affects the data. RMS Training Manual, Property Modelling 20 Estimated residual std. Dev. Is used as input in the Variograms tab. This should approximate to 1 if a good normal score distribution is achieved after all transformations. In Advanced mode, full flexibility is given to edit, add and remove elements. If you turn off Transformation sequence = Automated, some Estimate transformations options appear, allowing further refinement of the sequence. To add transformations to the sequence:  In ‘Automated’ mode, simply click the Append/Insert icon  With ‘Automated’ mode off, either o Add and estimate the transformation directly by clicking the ‘Estimate’ icon, or o Add the transformation by clicking the ‘Append/Insert and Estimate’ icon to estimate, or specify settings in the Edit dialog box RMS Training Manual, Property Modelling 21 4 CORRELATION TAB A correlation specified in the Correlation tab is the linear correlation between two residual distributions (i.e. after transformation). Correlations can be set in two ways in the Correlations tab:  If modelling more than one petrophysical parameter, they can be cosimulated by specifying a correlation factor which indicates the strength of the relationship between them.  If seismic cosimulation is selected, a correlation factor is specified to indicate the strength of relationship between the seismic data and parameter(s) to be modelled. The correlations are defined in the correlation matrix, which shows all variables plotted against themselves (correlation = 1) and against each other (default value = 0). The value indicates the degree to which, for a given position, the value of one variable is related to another. If the correlation is positive, the values of the two variables tend to be small or high at the same time; if the correlation is negative, the value of one variable tends to be small if that of the other variable is high. The closer the correlation is to either 1 or -1, the stronger these tendencies become. Zero correlation means that the variables are independent. Correlation matrix Correlations can be estimated directly from the data, or entered manually in the matrix RMS Training Manual, Property Modelling 22 5 VARIOGRAMS TAB The Variograms tab is used to define the 3D spatial distribution of the parameters to be modelled. 5.a The Variogram A variogram model specifies the continuity of the parameter. It indicates to what degree the residual value in one position is related to the residual value in a position nearby, as a function of their separation distance. That is, the variability increases with increasing separation distance between two observations. Variograms are used in both facies and petrophysical modelling to determine the likely value of a particular property. They can be generated without well data, but are more often estimated from the well data itself. The main terminology in a variogram is discussed in the diagram below. Variance (2) – degree of dissimilarity between points Sill – maximum dissimilarity. If there are no trends in the data, this is simply equal to the variance of the data. When there is a trend, there is usually no sill (i.e. the data continues to change with distance from a given point). h – separation distance (i.e. distance between points) Range – correlation length, which is the distance at which variance = sill RMS Training Manual, Property Modelling 23 Nugget – degree of dissimilarity at zero distance, which is normally due to microstructure or to instrument noise/measurement error in sampling Variograms can be defined via data analysis or directly through the Variograms tab. The following sections describe how to use the Variograms tab: 5.b Definition in the Variograms tab 5.b.i Previewer The Previewer can be launched anytime from the bottom of the tab and gives a very useful visual guide to what the final parameter will look like. It’s a good place to experiment with settings, remembering to always ask yourself “Does this look geological?” If you are using a facies model also use the Previewer with this in mind. Remember the rough sizes of facies bodies to see whether the variogram is a realistic complement to the facies model Clicking the dice will show a different realisation using the same variogram X-Z will show you a cross section to assess vertical distribution RMS Training Manual, Property Modelling 24 5.b.ii Anisotropy Two anisotropy settings are available Geometric anisotropy involves definition of a single variogram for each zone/facies etc. Zonal anisotropy involves the definition of 2 variograms, one in the lateral direction and one in the vertical direction. 5.b.iii Standard deviation (Variogram Sill) The standard deviation is the square root of the variogram sill, which is the highest point of the variogram curve. A higher sill equates to greater variability in the residuals. The SD is defined for each parameter to be modelled, so it can be different for each variable. Automatic: The SD will be calculated automatically from the transformed data Constant: Either enter the constant value manually or click the Estimate button to calculate the value from the transformed data 3D Par.: Select a 3D continuous parameter from the list. You can use this option to specify different sills in different parts of the reservoir. 5.b.iv Variogram model Two basic types of variogram model are available: Texture. A value for Roughness is specified, where 0 indicates a rough variogram and 1 a smooth variogram RMS Training Manual, Property Modelling 25 Standard. A variogram type is selected from the drop down list. The most common ones used are Spherical, Exponential & Gaussian. If the 3 are plotted together, it shows that the Exponential variogram shows the highest variability, giving a noisier result. The Gaussian variogram is often too smooth (low variability), and the Spherical variogram lies between the two. Exponential Spherical Gaussian 5.b.v Anisotropy directions This defines the directions of the axes of the variogram ellipsoid using azimuth and dip values. The ellipsoid should be aligned with the plane of maximum continuity. RMS Training Manual, Property Modelling 26 There are 2 options for Azimuth. Constant allows you to enter a direction and dip: If a Body Facies Parameter is used as input (General tab), the Azimuth instead can be set to follow the orientation of the bodies: World coordinates – All bodies of that facies have the constant azimuth and dip specified Body direction – Each body has a different parameter azimuth, depending on the body orientation Body curvature – The anisotropy will follow the local body direction, so, for channels, it will curve with the channels For Body direction and Body curvature, set the Azimuth constant to zero. If it is not zero, the variogram ellipsoid will be rotated with respect to the local body orientation. The 2D option for Azimuth uses a 2D trend surface or vector field, to very the azimuth. This cannot be used in conjunction with intrabody options, so should be used when a general variation in azimuth is seen across the reservoir/zone/facies etc. RMS Training Manual, Property Modelling 27 Vector field generated from polygons. The yellow line shows the gradual change interpolated across the polygons Resulting 3D parameter 5.b.vi Range Ranges are specified for X, Y & Z. Using the previewer for this is very helpful. 5.c Using Data Analysis to define the variogram The Data Analysis tool allows inspection of well data to create a variogram model based on the analysis, which can subsequently be used in the petrophysical modelling job. To successfully employ this method requires adequate data points (at least 100 wells as a general guide). Workflow 1. Create a univariate DA object for the wells/blocked wells 2. From the DA object, Create... Variogram 3. Follow the 3 steps from the Vario object RMS Training Manual, Property Modelling 28 5.c.i Estimation of azimuth This panel can be used to estimate the variogram azimuth 5.c.ii Estimation settings Wells are plotted to enable the search cone to be defined. RMS Training Manual, Property Modelling 29 Max lag: Lag length: Max. width: Width angle: Max height: Height angle: max. length of search area for pairs separation distance between pairs (bin size) max. width of search area bandwidth angle of the search cone (horizontal) max. height of the search area bandwidth of the search cone (vertical) Max lag distance is the maximum distance at which data will be paired. When it is set to the largest distance between any two wells, the maximum number of data pairs is captured, and increasing this value further will have no effect. It is often useful to set the maximum lag in the parallel direction to half the length of the field. Lag length should be set to the typical inter-data spacing, that is, the interwell spacing in the parallell dirction and the layer thickness in the vertical direction. This affects the number of points plotted for variogram modelling. It will depend on how many wells you have, but iti si difficult to fit a variogram curve unless you have data points for at least 5 (preferably 10) data bins. 5.c.iii Variogram Modelling Variogram modelling involves fitting the model to an experimental variogram (commonly spherical, gaussian or exponential). It is an iterative process and should take into account RMS Training Manual, Property Modelling 30 geological knowledge (and remembering, as ever, that a good facies model will take care of much of the petrophysical distribution). The amount of data is key. We often have plenty of vertical data, so the variogram is easily estimated and reliable. The horizontal variogram often lacks enough data points to be valid. 2 1 1. Estimate all directions will plot the variogram points (red) based on the well data, then draw the model (blue line) according to the varigram type chosen). 2. The model line can then be manually shifted by dragging the blue dot. RMS Training Manual, Property Modelling 31 6 LOCAL UPDATE TAB 6.a Local Update Local Update is used to update existing 3D models of reservoir properties (like porosity and permeability) with new well data. A typical scenario is where an asset team has a historymatched 3D model of (for example, permeability), and where a newly drilled well shows some differences along the wellbore. The aim is then to update the model locally around the new well, keeping the model equal in areas outside the well neighbourhood. The final result is created as a weighted average of the initial petro model and a new intermediate model: new = weight_grid * intermediate + (1-weight_grid) * initial The Weight Grid is a continuous parameter with values from [0,1]. A unit value in the Weight Grid means that the value in the initial model is completely disregarded in the final model, while a zero value in the Weight Grid means that the initial model is perfectly preserved. Typically, the Weight Grid will be “1” around the new wells, while “0” around existing wells. Different options for defining the Weight Grid exist, from manually to fully automated. RMS Training Manual, Property Modelling 32 7 WATER SATURATION MODELLING The role of water saturation modelling is to produce an accurate representation of fluid proportions as a continuous 3D parameter, which will subsequently be used in volumetrics calculations. Water saturation modelling is performed via the Water Saturation modelling job. 7.a General tab Regions can be fault blocks or any other discrete parameter The property classifier is used when different input is required for different porosity or permeability ranges. A new parameter with discrete values per interval will be created when ‘Map to discrete parameter’ is clicked. RMS Training Manual, Property Modelling 33 7.b Variables tab The variables tab defines the input variables for the water saturation calculation. 7.b.i Functions There are 4 types of functions to select from. The function will be applied across the whole model, but the values within the function can be varied per zone, facies etc. For each function, two different Height calculations are available: Basic Height options Height options when integrating into the formula for J-functions Centre of cell : The centre of the cell is extracted and truncated at the FWL Centre of top/bottom cell face Centre of cell above FWL : Only the part of the cell above the FWL is taken into account Centre of top/bottom cell face above FWL 7.b.i.1 Look up functions This is the simplest method involving use of a user defined trend. The Sw calculated is a function of height in the reservoir. The height (H) used is the distance between the value/surface specified as FWL and the cell centre. The look up trend must be placed in the trends container, with SW represented on the Y axis and height on the X axis. RMS Training Manual, Property Modelling 34 7.b.i.2 J-function (SCAL based) Using this function, specify the density for water (Rho_w) and oil/gas (Rho_hc), the gravity acceleration constant (g), interfacial tension (gamma), contact angle/wettability (theta), permeability (Perm), porosity (Poro) and two petrophysical constants, a and b. In SCAL-based J-function a default conversion factor (0.017453) is set for converting input angles in degrees to radians (as IPL takes radians only). Function 7.b.i.3 Function when integrated J-function (simplified) The input for this function is permeability (Perm) and porosity (Poro) taken from the grid and the petrophysical constants a and b. Function Function when integrated 7.b.ii Variables tab Once the function is selected, the variables can be specified per zone, sub-grid, facies & property classifier RMS Training Manual, Property Modelling 35 Define values for each zone/facies etc allows you to entre values in a table RMS Training Manual, Property Modelling 36 7.b.ii.1 Definition of variables The table shows the full list of variables, along with a tick list of what is required per function. Look up function J-function (SCAL) J-function (simplified) User defined Variable Definition FWL Free water level     SWirr Irreducible water saturation. Low values are truncated to this value     SWmax Maximum Sw value. High values are truncated to this value     Rho_w Water density  Rho-hc Hydrocarbon density, oil or gas (non-wetting phase)  g Gravity acceleration constant  gamma Interfacial tension  theta Contact angle/wettability (default of 0.017453 used to convert from degrees to radians)  Perm Permeability    Poro Porosity    a Petrophysical constant    b Petrophysical constant    Cpc Unit conversion constant. Use if Pc input unit is different to Pc unit in function  Cj Unit conversion constant. Use if J input unit is different to J unit in function  lookup The trend function residing in the Trends container (y-axis=SW, xaxis=height) RMS Training Manual, Property Modelling   