This post presents an approach for modeling and building requirements for a MES / Production Management Stockpile module for continuous processes using iron ore stockpiling as reference.
Ore mining and beneficiation is a continuous process by nature. In several points of this kind of process it is usual the usage of large stockpiles (ROM and final product). Make to stock strategies are more common than make to order.
ANSI/ISA-95 oriented MES production, storage and inventory solutions modules are more often used in batch and discrete processes such as drugs, shampoos and cars production.
Does this fact restrain an approach for iron ore stockpiling?
An yard can be modeled in two or three dimensions.
Considering three dimensions, yard length is generally split by landmarks along it at each X meters and it is the first dimension to take account. Yard height is commonly split in 3 or 5 positions and it is the second dimension considered and yard depth as third.
Two dimensions modeling is nothing else than third dimension (depth) fixed with a unique value of Z meters.
A landmark (X position), height (Y position) and depth (Z position) set is then a discrete stockpile position and its content material modeled as lot (material lot as ANSI/ISA-95 definition).
This lot may have all modeling features of an ANSI/ISA-95 oriented MES software / solution such as amount, material definition (SFXX – Sinter Feed XX as an example) with more detailed properties like %FE and %SIO2.
Besides lot concept itself it is possible to perform a relational modeling of each position as a storage spot and a spot can be composed as a set of others (children spots). A huge storage region A can contain several yards, as an AA yard which can contain a subset of positions as described above. A specific position is then defined as Region A, Yard AA, Landmark 10, Height 2 and Depth 1 as an example.
What are this approach benefits?
A relational database and modeling and ANSI/ISA-95 oriented approach brings inventory management typical benefits such as fine grained inventory control at each spot (specific position as explained) which settles close to reality features making it possible to simulate commonly used ore storage operations as product dispatching and blending for specific customer orders / requests of certain products with well defined properties.
Does this approach always brings benefits?
Relational modeling for discretization does not always bring benefits. In many circumstances or real world processes this kind of modeling only adds more bureocracy for process fitting to system and not the way round.
A clear example can be a beneficiation process where there is few resulting ROM property variation as iron ore mining in north of Brazil which produces final iron ore with extremely high iron rate. Final product is almost uniform.
Want to know more?
Next parts of this series will present link between more mining processes from beneficiation to stockpiling, typical stockpile operations details and modeling as blending for instance as well as intermediate buffer piles, some features of out of the box tools / products for mining, last generation mining tools or features (e.g. augmented reality) and much more.
Watch for next releases.
Portuguese version: http://goo.gl/GJDpPD