From exploration to production

The whole module in the order you would actually do it: a drillhole database on Monday, a published plan the shifts file against by the end of it.

The other pages describe a screen each, because that is how you look something up. This one is the road: seven parts, in order, each ending with something the next part needs. Follow it once end to end on a small deposit you already know the answer to, and everything else on this site becomes reference rather than reading.

The road, in one table

Part	You start with	You end with	Screens
1. Ground to stand on	A licence and a mine grid	A Tracker Group and an item to file against	Tracked Groups, Tracked Items
2. Exploration	Collar, survey and interval files	A validated drillhole database	Mining → Drillholes
3. Interpretation	Holes and sections	Closed wireframes of the ore	Mining → Wireframes
4. Resource	Wireframes and composites	A block model with a grade on it, and a curve	Mining → Block models
5. What is worth mining	The model and the topography	A pit shell — or a stope layout — filed as a design	Mining → Block models, Mine designs
6. The mine as built	The shell	Pit, dumps, roads, ramps, water, stages	Mining → Mine designs
7. The plan	A design and today's ground	Published period plans the dashboards judge	Mining → Plans

Then it loops: the shifts file, the surveyor measures what moved, the plan is reconciled against it, and the next period is planned from ground that is a month newer.

You do not have to start at part 1

An operation that is already running usually starts at part 7 — plan against the pit as it is — and fills the earlier parts in behind. See Bringing an existing operation across for the order to do that in.

Part 1 — Ground to stand on

Nothing in this module floats free. Every hole, surface, wireframe, model, design and plan belongs to a Tracker Group, and the group's whitelist is who may see it. Most of the second week's confusion — "the model is not in the picker" — is a whitelist, not a fault.

Step 1. Make the group, and say what it is used for

Tracked Groups → Add Group

Under Used by, tick the technical work this group carries: surveys, geology, designs, plans. The Mining pickers narrow to groups that say they do that job, so a group set up for weighbridge tickets does not clutter a wireframe picker.

A group that says nothing is offered everywhere

Used by left blank is not yet said, not none. The group is offered in every picker with a hint that nobody has said what it is for. Tick the boxes once and the pickers stay short forever after.

Step 2. Make the item

Tracked Items

An item is the piece of ground: the pit, a named dump, a stockpile. It is what ties a survey to the survey before it, what a drillhole database is filed against, and what a plan is a plan of. Get this wrong and nothing is comparable to anything — the fix is to import again against the right item.

One pit, one item. Do not make an item per survey.

Step 3. Name the grid, once

Every file that leaves this module carries the account's mine grid, and every file read in is read on it. State it once and every export, every desurveyed interval and every plan document agrees. See The mine grid, named once.

Check the easting and the northing before anything else

This is the one mistake that costs a month. A collar file mapped the wrong way round loads perfectly, validates cleanly, builds a plausible database and puts every hole mirrored about the diagonal — which on a north-south orebody looks almost right. The same trap sits on survey import, where the wrong instrument does not fail, it mirrors.

At the end of part 1 you have a group you are whitelisted on, an item, and a grid. Everything below files against those three.

Part 2 — Exploration: getting the holes in

Mining → Drillholes

A drillhole database is three files joined on the hole name. Trackapt reads them as they come out of acQuire, Surpac, Datamine or a spreadsheet somebody keeps by hand, as .csv, .txt or .dat.

Step 1. Get a blank to fill in

If you are collecting the data rather than exporting it, take the blank first: each import screen offers a template with the columns it expects. See A blank to start from.

Step 2. Import the three tables

Import drillholes takes:

File	What it holds	Without it
Collars	One row per hole: identifier, easting, northing, level, total depth, and the collar dip and azimuth	Nothing can be placed
Downhole surveys	Depth, dip and azimuth down each hole	Every hole holds its collar attitude all the way down — right for a short hole, optimistic for a long one
Interval tables	From, to and values: assay, lithology, recovery. One file each, and you name each one	There is nothing to estimate from

Because no two packages agree on column order, the reader proposes a mapping from the header row and you confirm it. Read the proposal rather than clicking through it — this is where the easting and the northing get swapped.

Step 3. Assemble before you file

Assemble reads the files and shows what they hold without filing anything: how many holes, how many intervals, the extent, the depth range. Check the extent against where you know the drilling is. File then commits the holes against the item.

Step 4. Validate the whole register

Validate is the step people skip and then pay for. Run it on the whole database, not on a hole.

A duplicate hole identifier cannot be seen from inside a single hole, and it is the worst fault available: two holes' intervals silently become one hole's, at two different places in space. Validation reports against named holes and depths — "DDH014 at 45 m" — and each hole in the register carries its own error and warning count.

It checks collars (missing or duplicate identifiers, a dip pointing upward, an azimuth out of range), downhole surveys (a station past the end of hole, duplicated depths, an implausible dogleg), intervals (overlaps, inverted from-to, gaps beyond tolerance, suspiciously long intervals) and values (negative grades or recoveries where you have said a field may not be negative).

Everything found is in the export, not just what fits on the screen. Fix at source, re-import, validate again — and only then move on.

Step 5. Read one hole, and check the desurveying

Open a hole for where it went, what was in it, and what is wrong with it.

Depth along the hole is not elevation

At −60°, 100 m down the hole is 50 m away from the collar and 86.6 m below it; at −90° it is directly underneath. Treating depth as elevation is the most expensive mistake in the module, because it produces a model wrong by tens of metres that looks entirely reasonable. The desurveyed position is shown beside every interval, on the mine grid. Read that, against a plan, before you believe an intersection.

Step 6. Look at the composites

Composite by a fixed length to see what an estimate would see; composite by what the ground is, broken on lithology or ore type, to read a contact and the thickness of a band. Domain composites are shown, never handed to the estimate — the estimate keeps its own compositing.

At the end of part 2 you have a database that validates clean, and you have looked at a hole and believed its position. See Drillholes for the full detail.

Part 3 — Interpretation: wireframes

Mining → Wireframes

A wireframe is the geologist's opinion made measurable: outlines drawn section by section against the holes that pierce them, tied into a closed body.

Step 1. Let the holes draw the first pass

On an item that carries drilling, the editor offers Build from drillholes. Give it the interval table, the field, a cut-off, the section bearing (normally the strike), the section spacing (usually the drill spacing) and a minimum half-width, and Build the sections puts a convex hull round the intercepts on each section. It reports what it drew and what it cannot be trusted about.

That is a shape to start from in seconds instead of an afternoon of pasting coordinates. It is not an interpretation — see A first pass from the drilling for what a hull gets wrong, and in which direction.

Step 2. Edit it into an interpretation

One outline per section, against the holes on it, using the structure and the logging the assay table knows nothing about. The skin between the sections is tied each time the solid is drawn rather than stored, so what you see is always the shape the sections currently make — revise a section and the body follows.

Step 3. Make it close, and only then read a tonnage

A volume comes from the divergence theorem, which is exact only over a closed boundary. On an open mesh it still returns a number — plausible, stable between runs, and wrong.

So the register shows a tonnage only where there is one. An unclosed wireframe says so instead, and names what is wrong: an edge used by only one triangle (an uncapped end), an edge used twice in the same direction (a triangle inside out, subtracting its part of the body), an edge joining more than two triangles, a triangle with no area, a triangle referring to a point the mesh does not have.

A missing figure is a question somebody asks; a wrong one is a figure somebody reports

That is the whole reason the tonnage is withheld rather than estimated.

A tonnage also needs a density on the wireframe. Where either is missing, the volume stands alone.

Step 4. Turn it

View in 3D and look at it against the holes. A solid that closes and is still wrong usually looks wrong here first.

At the end of part 3 you have closed solids with densities, ready to decide which blocks of a model are ore. See Wireframes.

Part 4 — Resource: the block model

Mining → Block models

Step 1. Define the lattice

Origin, bearing, block size and counts. Two decisions made here bite later:

• Block height should divide the bench height you will mine at, or every bench's reserve borrows from the one below. The plan reports it; it does not refuse.
• Block size against the drill spacing. Blocks far smaller than the data produce a model that is mostly interpolation, and a model past a million blocks is refused by the pit optimiser rather than part-solved.

Step 2. Fit the variogram before you estimate

Under How grades behave, Plot what the samples say draws the experimental variogram from the composited samples, with the model you typed drawn over it — using the very curve the estimator will use.

Read the two against each other: a curve well above the points overstates how far grades carry and smooths the estimate; one below chases noise. Bins with few pairs are drawn small and hollow, so you do not fit to them. Plot with Along the axis on and off — the two curves reaching their sill at different distances is the anisotropy, measured rather than assumed.

A variogram from the wrong deposit gives weights that look exactly like the right ones

Kriging derives every weight from the nugget and the sill. Nothing downstream will tell you the model was wrong. This plot is the check.

See Checking the variogram against the samples.

Step 3. Estimate

Estimate interpolates from the drillholes around each block.

Method	When you would use it
Ordinary kriging	Weights from the fitted variogram, with a variance alongside. The standard for a reported resource
Inverse distance	One over distance to a power. No assumptions, defensible to anybody, and knows nothing about how the deposit behaves
Nearest neighbour	The closest sample and nothing else. The reference a kriged model is judged against, not a block value

You also set the composite length, the search ellipsoid, the minimum and maximum samples and a maximum per hole. Every estimated field records which method and which search produced it — a grade nobody can state the provenance of is a grade nobody can defend.

Step 4. Read the curve

The curve is the point. Four million numbers is not a resource; how much ore there is above each cut-off, and at what grade is.

Set the grade field and the cut-offs, and set two more fields deliberately:

• Ore fraction — how much of each block is ore. Without it every edge block counts whole, and the tonnage is overstated all round the orebody.
• Density — a field where it varies, otherwise one figure. Everything scales directly with it.

The average grade above a cut-off is weighted by tonnes, not block count, and a block with no estimate is left out rather than counted as zero. Both are handled for you, and both are how a curve gets quietly overstated elsewhere.

At the end of part 4 you have a model with a defensible grade on it and a curve you can show. See Block models.

Part 5 — What is worth mining

Two things have to be true before this step means anything: the model carries an estimated field, and there is a survey of the ground. Blocks above the ground are air, and without a topography the optimiser pays to move the sky above the deposit.

Step 1. Survey the topography first

Mining → Pickups and Mining → Surveys

If the ground has never been surveyed, do it now — it is the same import you will run every month from part 7 onwards.

1. Import the file and say which instrument wrote it. The wrong choice does not fail, it mirrors: total stations, GPS rovers and office packages all emit comma-separated points and then disagree about which of the first two numbers is the northing. Work it out from the file is a genuine answer on a national grid and a guess on a small local one, and the reader says which.
2. Check what the reader understood: how many shots, how many feature lines, the extent, the level range.
3. Say what each code means, so breaklines are honoured.
4. Name it, date it the day the ground looked like this — not the day you filed it — and file it against the group and the item.

Consider starting a running surface for the pit at the same time (the layers icon on Surveys). It is the ground as it stands, kept up to date by every survey imported into it, and it is what a plan reads as of a date. See A surface that keeps up to date.

Step 2. Optimise a pit out of the model

The trending icon on Block models opens Optimise a pit: which blocks are worth mining once the waste above each one is paid for, and therefore what shape the pit is.

Under	You give it
What the rock is worth	The grade field, the price per unit of grade, recovery, and the selling cost per unit
What it costs to get at	Density, mining cost per tonne (every tonne), processing cost per tonne (ore only — a block that cannot pay it is mined as waste, which is where the cut-off comes from), and the wall slope from the geotechnical report
Prices to run	Fractions of the price above. Each gives a pit; the small ones sit inside the big ones, and that nesting is the order to mine in
Surveyed ground	From step 1

The ultimate pit reports value, ore tonnes, stripping ratio and the floor level, then a row per price.

Whole blocks quantise a wall

A slope built out of whole blocks can only ever steepen a wall. The dialog says what slope the pit actually stands at, and how narrow the blocks would need to be for the one you asked for.

The optimisation runs on the server, so this is the one screen in the module that needs a connection.

Step 3. Make it a design

Make it a design files the shell's deepest outline as a pit floor under Mine designs, with the bench height, batter angle and berm width you give. That is the handover from part 5 to part 6.

The underground branch

Underground, the equivalent question is Lay out stopes: give it the grade field, density, plant recovery, product price and the underground mining and treatment costs, then the shape a stope may be — sublevel interval, panel length, narrowest and widest stope, wall slough and its grade, mining recovery.

It reports how many stopes pay for themselves, the tonnes fed, how much is slough, the grade the mill sees, and what the layout is worth before development.

It files nothing, deliberately

A pit shell is a design you can plan against. A stope layout is what a frame and a set of costs imply — and both live on the plan, where they are read again every time it computes. Filing the stopes would freeze an answer that is meant to move. To schedule them, open a plan and set its shape to Stopes and headings.

See Optimising a pit out of the model and Laying stopes out of the model.

Part 6 — The mine as built

Mining → Mine designs

The parameters are the record; the geometry is generated. A pit's benches are not stored — they come from its floor outline and its slope rules, and opening the design runs it. So when a geotechnical report is revised, every pit designed to the old one is visibly out of date rather than quietly still drawn the old way.

Step 1. The pit

A pit is designed upward and outward from its floor, because that is what an engineer actually decides: the bottom is placed on the ore, and the walls are whatever the rock demands on the way up.

Parameter	What it is
Floor outline, Floor level	The pit bottom in plan, and the level it sits at
Bench height, Batter angle, Berm width	The slope rules
Maximum benches, and either a topography or a maximum level	Where it has to stop

The result says why it ended: Surface (it reached the topography — the normal outcome), Closed (the offset ran out; the floor is too narrow for the slope rules asked of it), or Benches (it hit the limit first).

Warnings are outcomes, not errors

"The floor is too narrow for a 65 degree batter" is the design telling you something true. The remedy is a decision — a switchback, a flatter grade, a different floor — and the design does not make it for you.

A pit designed in another package comes in whole under As drawn elsewhere, as toe and crest strings; the drawn benches are then what every plan uses. See A pit as drawn elsewhere.

Step 2. Water in the pit

Three numbers, and the plan reads them so the pit is dewatered as it was designed rather than as somebody remembered it: sump depth and sump area (the box below the floor where the water collects and the pumps stand), and the diversion drain — how far beyond the top crest the cut-off drain runs. Ground beyond the drain sheds away; the strip between the crest and the drain still drains in, and its area is reported. Leave all three blank for a design with no water on it.

Where the water table is, and how much the pit takes from it, belongs to the plan, not the design. See Water in the pit.

Step 3. Stages, where the sequence matters

A stage records a pushback or a cutback the mine wants taken before or after another part of the same pit: a label, an outline in plan, a From level and a To level, and After — the other stages that must be finished first. After is the whole of the ordering; the list order is display only.

A pit with no stages is mined bench by bench, top-down, exactly as drawn. Add a stage only where a real sequencing decision needs to be seen — and the schedule optimiser in part 7 will use it.

Step 4. Somewhere for the waste to go

A dump is the pit turned inside out: designed upward and inward from its footprint, lift by lift, until the top closes on itself or the limit is reached. Design the dumps now, because part 7 checks whether the waste actually fits in them.

Step 5. Roads, ramps and rings

Haul roads and ramps are designed here too, and a haulage process in the plan can read its route straight off one rather than taking a tonnes-a-month figure on trust. Underground, headings reach the stopes and drill rings are laid out on them.

At the end of part 6 you have the mine as it will be built, and the plan has something to walk. See Mine designs.

Part 7 — The plan

Mining → Plans

This is where the technical work becomes something the mine is run against.

Step 1. Choose the question

Mode	The question
To a product target	We need this much product. What has to move?
To capacity	We can move this much. What does that buy?
To the best order	Given the whole horizon, what order gets the most out of an objective?
Imported	The schedule was decided elsewhere. Read it against this plan's own reserve and publish it here

The first two walk the pit top-down, bench by bench, the last bench of a period pro rata — an answer you can check by hand. The third plans every period at once. See Optimising the schedule and A schedule decided elsewhere.

Step 2. Point it at the sources

Everything from parts 2 to 6 arrives here:

Field	What you choose
Pit	The pit being planned. Its surveys, models, designs and boundaries fill every picker below
Ground, Ground as of	The surveyed ground and the date it is read as of. A running surface is read as it stood then, so everything already mined drops out on its own
Block model, Grade field, Grade unit	The deposit, which attribute is the grade, and what one unit of it means per tonne
Shape	A pit design (part 6), or slices of a boundary — a column from the highest ground inside an outline down to a floor level, in benches. No batter, no berm: a first answer, not a design
Ore is decided by	A cut-off grade, or an ore-fraction field
Recovery, Density, Unestimated blocks are, Swell factor	Product is in situ × recovery; a block with no grade is never ore; swell is what turns waste tonnes into the volume the dumps must hold

Step 3. Say what the mine can actually do

A new plan starts with nine processes in the order the ground is worked: pre-stripping, topsoil removal, waste drilling, waste blasting, waste haulage, ore drilling, ore blasting, ore haulage and explosives. Add any others. Switch off what the mine does not do — a switched-off process places no demand and never binds.

Then, process by process, Capacity from:

Capacity from	You give it
Declared, one figure	A million tonnes a month
Declared, varying by period	A figure per period, or a column pasted from a spreadsheet
Built up from equipment	Units with a count and a rate per hour, shifts, hours, availability and utilisation. Where the account holds Maintenance, Read availability from maintenance replaces the typed figure with what was measured
Demonstrated by a parameter	What the process already records, over a look-back, as the mean, the best period or the 80th percentile
From a truck cycle	The trucks, the haul route as segments of length and grade, loading and dumping time, and the shifts

This is the step that keeps a plan honest

A fleet build-up that promises double what the same fleet demonstrated last quarter is visible as such. Switch the process between the two sources and watch the walk stop in a different place. And a fleet build-up says what the fleet could do; a truck cycle says what the road lets it do — the same ten trucks give a different capacity on a 2 km flat haul and a 5 km ramp at 10 %.

Step 4. Set the horizon

A cadence from daily to annual, a start date, and up to 24 consecutive periods. Each period continues where the last stopped, with every capacity refilled. The screen says how many calendar days the first period has, because that is what a fleet build-up is multiplied by.

Step 5. Compute, and read it properly

Compute saves the plan, loads the sources, reads the reserve bench by bench and schedules the horizon. Per period you get:

• the benches taken and the fraction of each, ore and waste tonnes, grade and product;
• every enabled process's demand against its capacity, with the binding one marked and any exceeded one in red;
• the stripping ratio, the ground-as-of date, and the warnings;
• on haulage costed from a truck cycle, the trips, truck hours, diesel and CO₂;
• whether the waste fits in the dumps as loose cubic metres, and the pit's water balance where the plan carries rain and groundwater.

The binding process is the answer to what is stopping us. Everything else is slack.

Step 6. Try the alternative before you commit

New scenario makes a draft copy beside the plan, ready for one thing to be changed — a second excavator, a lower cut-off, the schedule the consultant sent. Compare scenarios lays them side by side, summed over the horizon and then period by period. A scenario is a plan in every other respect, so the one that wins is simply the one that is published.

Step 7. Approve, then publish

Where the pit's group has an approval chain, Publish parks the plan as Awaiting approval and names the desk it is waiting on; every approver at that stage is told on their devices. Nothing reaches the parameters until the last desk signs. See Approval before publishing.

Publish then lists every figure the plan produced — ore, waste, total rock, product, drilled, blasted, truck hours, diesel, dump volumes, pumping hours — and against each a Publish onto picker: the monitored parameter that figure becomes the plan for. One period plan is written per figure per period.

This is the moment the module stops being a technical department

From then on nothing on the dashboard says mine plan. The published figure is simply the Plan the dial, the variance and the status use, and the period summary over a form's submissions judges the filed readings against it. A dashboard already open picks it up without a refresh.

A published plan is locked. Unpublish withdraws its period plans and makes it editable again.

And then it loops

The shifts file against it

The people filing carry on as they always do. The shift form they already fill in is judged against the plan, period by period, on the parameters you published onto.

The surveyor measures what moved

Every month, import the survey against the same item (part 5, step 1). Read cut and fill between two dates inside the boundary you actually report on, and check the triangulated figure against the grid and end-area methods — the second opinion is the point. See Change since an earlier survey.

Reconcile the plan against the ground

Reconcile on a published plan shows each figure four ways: planned, surveyed (the ground lost over the period), filed on the parameter, and weighed across the weighbridge. Nothing is stored — it is read from the survey, the readings and the tickets each time.

Four columns that disagree are the most useful screen in the module. See Reconciling and Reconciling against the weighbridge.

Re-plan the rest of the horizon

Two months into a six-month plan, Re-plan keeps the periods already worked and plans the rest afresh from the ground as it stands now. See Re-planning part-way through.

And back to part 2

Infill drilling goes into the same database, re-estimates the same model, moves the same curve, and the next pit shell is optimised against a resource that knows more than it did. That is the loop the whole module is shaped around.

If you are doing this for the first time

A fortnight, in the order that builds confidence fastest:

1. Import one survey of a dump you already know, and check the plan against what you remember of the ground.
2. Draw a boundary round the block you actually report on, and name it.
3. Import last month's survey of the same item, add it as a layer, and read cut and fill inside that boundary. Check the two methods agree.
4. Reconcile that against the weighbridge, and draw a plan sheet at 1:1000.
5. Import a small drillhole database and run Validate over it. Fix what it finds at source and import again.
6. Open one hole and read a desurveyed interval against the plan.
7. Draw one wireframe from two or three sections and confirm it closes.
8. Define a small block model over it, plot the variogram, estimate one field, and read the curve.
9. Optimise a pit out of that model against the survey, and file the shell as a design.
10. Design one waste dump for it.
11. Start a plan against that design and today's ground, compute it to capacity, and see which process binds.
12. Publish its waste onto the parameter the shift form already fills in, and watch the dial change.

Step 12 is the one to reach. Everything before it is technical work; that is the step where the mine starts being run from it.

Where to go next

For	Read
The screens, one at a time	Surveys, Volumes, Drillholes, wireframes and block models, Mine designs, Plans
Getting existing data in, or handing work out	Working with other packages
A printed copy for a desk or a cab	Mining user guide (PDF)
A question somebody already asked	FAQ