First, what are the important qualities for incoming ore? Mass recovery? Concentrate grade? Grindability? Metallic recovery? Controlling some minor constituent that affects concentrate quality in <<1% amounts? Keeping all of the above within an acceptable envelope? Remember the mine also has to supply a certain tonnage of acceptable ore while this is going on.

Second, what choice does the mine have? Do they have sufficient blasted inventory to allow corrections to ore delivered to the plant? Are all their loading units operating, or operable? They have operating cost constraints too, like stripping ratio, haul and lift distance, lousy ore on top of good ore that has to be moved, double handling of ore, and other things.

Third, are there tests that readily and consistently define good ore in the drill core or blast hole drilling stage? We are very fortunate with Magnetite that there are variations on the Davis tube test that readily and easily define what the plant will do with the ore. We can also measure some of the quality parameters on conveyor belts or even in the blast hole.

Fourth, there are people things involved. Is the mine manager associated with this a person who always promises more than he can deliver? Does he get along with mill management? Does he have authority to make changes in pit operation to keep the mill feed in line? Does he understand that the mill has a limited envelope of ore quality that the mill can massage into acceptable concentrate?

Fifth is the pit data base. Does it contain the information needed to make intelligent decisions about the mine plan? Is it up to date? One time where I worked a shovel dug over a quarter mile of bench in a cleanup cut over a weekend into a poor ore zone. Its location wasn't updated in the computer between Friday and Monday so the computer kept assigning Friday's ore quality parameters to the shovel's production. It took us a week to get out of that quality anomaly and figure out what happened.

Sixth is in process inventory. Does the mine feed the plant hand to mouth or is there a massive stockpile between them. By following quality anomalies we learned that our plant was making concentrate out of ore delivered 8 to 16 hours earlier. A problem with the ore showed up right away and could be readily traced back to the offending shovel. Other operations near here have massive stockpiles with up to a week of ore storage between the mine and plant. By the time a problem shows up in the plant there is a week of problem sitting in the feed stockpile and it's very hard to trace things back to the offending loading unit.

Seventh is the plant. Does it have the tools it needs to upgrade a reasonable variation in feed that the mine can consistently deliver into acceptable concentrate? Is it using these tools efficiently? Can it accept the ore available in the long term mine plan?

I used to help do this from the plant side at a very large iron ore operation. 

You have asked two great questions. I will start with the second question first and keep to the area that I am most familiar with - grinding. Best practice is so hard to achieve because we cannot agree on what is "best" practice. We are given client specifications to describe what a new plant should do. But we have focused so hard on keeping capital cost down that operating costs are allowed to sky-rocket and worse, in many cases the grinding mills chosen will not meet the specified tonnage. How can we even discuss best practice if the plant cannot achieve design tonnage on a daily, weekly and monthly basis? So why does this happen? In many cases, I believe it is due to a lack in accuracy and precision of the ore hardness measurements that are used to do the design and the lack of taking relevant samples on which the tests are done. To design proper grinding mills the measurements must be accurate (give the right answer), and be precise, (giving the same value on repeat measurements). Until the industry learns to focus on engineering fundamentals and give the design responsibility back to competent hands-on engineers, instead of relying on private software which only the developer understands, it will not be possible to achieve "best practice". In addition, the missing item in current software programs is the relationship of grinding performance to the economics of achieving the desired results. Not surprisingly, the operating savings in steel media cost, have been shown to actually pay for buying the proper grinding equipment, on an incremental basis when comparing to a bare bones design.

Going back to the first question, the barriers to best practice in mine to mill operations are both financial and behavioural. To even try to balance the mine result with the mill result and discuss optimization, we need to really understand both production units and the expected accuracy of predicted performance assessments. I think it fair to say that the effect on milling, of manipulating blasting in the mine, could be in the order of 5%. But if the measurements we use to assess the performance, (tonnes milled as a function of measured hardness) do not have an accuracy of 5% or less, there is not much chance of success. So why spend the money unless you know success can be achieved? Behavioural matters can be controlled with education, but only if there is "buy-in" from the mining and milling people who will need to work together to achieve best results. Buy-in can be achieved, but only when deadly accurate measurements are used to do the assessments and engineers are empowered to develop and understand the tools that will be required at their plant to achieve these improved results.

The engineers have developed tools to make this happen in the context discussed above (called SAG Design testing). Some are taking advantage of this. Many are not - and I can only surmise that this is also a behavioural phenomenon of using what our superiors have passed down as being the "best way". But if a better way has been discovered why must we wait for a whole generation to pass before we dare to call a spade a spade? If something is better, use it. Only in this mindset can we move forward in accomplishing what you have asked, and achieve "best performance" in mine to mill operations.

There are two recent issues that have recently crossed my desk.

•Miners have been brought up on "bonus" and the simplest mode of calculation is tonnage!

•Mine Conferences (who shall remain nameless- but it is a valid generalisation also) still plug "Bigger is cheaper" Big trucks, shovels etc save money, rather than the point that process engineers are making is that this is yesterday's technology.

Why "best practice" is so hard to achieve with the simple fact that experts in a field generally have difficulties to agree on what would constitute a set of "best practices". Throw-in some accountants, lawyers and "Qualified Persons" in the discussion (often arising from the need to prepare a NI 43-101 Compliant Technical Report) and what might have been an exercise of diplomacy and patience becomes impossible to achieve.

Labour is generally perceived as a major operating cost in a mine-mill facility and the obvious way to reduce labour costs is to produce more per person. Leading to the tonnage based bonus system for miners and the biggest is best approach for equipment selection for the mine and the mill.

I would go a bit further thatit is not only the industry needs to re-focus on engineering fundamentals albeit assisted by modelling and simulation software, but it must also be willing to expend the necessary efforts to generate the primary data - from mine to mill. It is however too frequent that precedence for data generation is in the area of defining the resources (how many tonnes and at what grades) with data about ore hardness, concentrate qualities and metal yields being an after-thought.

Unless a holistic approach to data generation and engineering is applied to the development of mineral resources, it will be extremely difficult to see successful mine-to-mill optimizations. There appears to be one exception - the low grade copper+/-molybdenum porphyries for which the margins are tiny and the mistakes of sending waste to the mill or ore to tailings cannot be tolerated for the site to be financially healthy. Note that the cut-off between ore and waste is hardness, concentrate grades, metal recoveries as well as feed grades.

All of the above comments put high values on communication between the stages especially mine to mill. The grinding circuits are all about getting the best result for costs incurred. To get that best result processing needs to know the changing characteristics of what future ore is coming down the feed belt. An ongoing communication problem does indeed make that tough.

Best practice, in many cases, is hampered by internal site communication.

The way a team plays as a whole determines its success. You may have the greatest bunch of individual stars in the world, but if they don't play together, the club won't be worth a dime.