Benchmark Sampling Systems for Material Balance

To run flawless processing plants, developing an unbiased metallurgical accounting is essential. However, most sampling equipment in the market is biased by design as manufacturers are rarely educated in the Theory of Sampling.

One prominent actor in the mining industry is Promi-Mex, a company that has much experience selecting the right equipment from their portfolio, by installing it and operating it. The company is also capable of installing all necessary infrastructure.

Promi-Mex´s ally Francis Pitard, President of Francis Pitard Sampling Consultants, and an expert on the Theory of Sampling explained to MBN that this theory provides the foundation for the basics of sampling, for instance:  what to look for, what can go wrong, and to what extent. Pitard says that choosing the right equipment is the core of success, and its primary recommendation is to educate the team to choose equipment from the leading manufacturers in the world.

According to Pitard, there are several things to take into consideration when obtaining samplings. “Achieving a correct sampling practice in a mining project is an iterated process. Analysis is required every step of the way, and the right information must be produced in each stage for the sampling to work.”

Pitard stresses how important it is for geologists to collect information that sampling experts require in order to understand the material’s sampling characteristics. All deposits are different, therefore, geologists drill, and study the core sample and carry out the logging. They focus on geological information. Nevertheless, they often forget to log what is important for the sampling experts, for instance: the size of the particles, where and how they are formed, among others. It is very important that all details are collected from the beginning, so the sampling process can be a success.

Pitard shared with MBN some of his research results published in his book called the Theory of Sampling. Pitard first shared certain things that can go wrong when sampling. Thus, optimizing sampling protocol is essential. “We need to stablish a protocol for where the sample is collected, how much material is collected, how can small scale segregation be avoided and minimize the effect of segregation.”

Pitard describes that the implementing sampling protocols is critical, therefore avoiding process where trouble can occur should be a priority. After its implementation, preserving the sample integrity to not change the chemical or physical composition of the material is fundamental. And lastly, avoiding analytical errors is considered the last step for success.

Optimizing sample protocols for precious metals could face poison processes. “For instance, when you look at copper you look at percent copper, for gold is PPM. When the sample is too small in magnitude, you do not have probability of distribution and everything becomes obscure and misses the goal, in rare occasions one gets too much gold for sampling,” said Pitard.

There are certain things to consider when creating a protocol and Pitard explained a couple fundamentals. The first one is in situ nugget effect, “if we do not know what is happening on the site, we might find nothing in our research, when it is actually a rich block, we need to understand it at very early stages to get most of it.”  Grouping and segregation error is another fundamental, this aims to reduce segregation as well as its negative effects.

When implementing a sampling protocol, one must know the sampling bias generations, said Pitard. “The world market today has a range of sample equipment available to use on your plant and 70 percent is ruled by design, thus it would never do the job of taking the sample because their designs are based on experience or imagination, but they do no fill good sampling jobs.”

Pitard said he and other colleagues realized of the lack of proper equipment and founded the Conference on Sampling and Blending, which takes place every two years in a different continent. “Our goal is to attract manufacturers to build better machines.”  

Implementing sampling protocols can also meet bias generators. One being the incrementing delimitation error. “In order to have a representative element coming from a stream for a sample, have a cross string sample a constant speed across the stream.” Another bias generator is incrementing extraction error and Pitard explained that sampling tools will impact with foreign solids, and when the material hits the cutter, it can become selective on what is taking, which is negative. It might be for the blades, speed, etc. “The latter is also known as an incrementing recovery error.” Increment weighting error is also a bias generator.

For preserving the sample integrity, which can also be another bias generation, one must be aware of preparation errors, contamination error, loss errors, alteration error, human error and fraud and sabotage. Preparation stages can alter physical and chemical alteration, said Pitard.

In regards to analytical errors, these can be scope versus principle, which implies an in-depth understanding of particles. It can also imply additive interferences, proportional interferences, drying temperature baseline, dissolution technique, composition of dissolution residues and contamination and losses, variability on a large scale is another challenge for sampling, this is based on an interpolation error of time and space, that can repeat periodically and increment weighting error.

In order for a company to obtain the true benefits of sampling, companies would have to rely on three fundamental pillars. The first one is effective management to identify structural problems, invest in solutions to solve structural problems, identify circumstantial problems, understand and act on all sources of variability and be proactive instead of reactive.

The second one is correct sampling. Based on In Situ Nugget Effect, identifying fundamental sampling error, grouping and segregation error, increment delimitation error, increment extraction error, analytical error, fundamental sampling error, increment preparation errors, interpolation error, periodic error and increment weighting error.

The third one is understanding variability. Sampling Variability generates visible and invisible financial losses, Pitard explained that variability presents itself as a problem and as an opportunity as well, and that there are many kinds of variability and all sampling errors magnify variability. “Geostatistics analyzes variability at the mine and chrono statistics analyzes variability at the plant and to minimize variability requires of a constant improvement strategy.”

Bias Generators

For a better understanding of processes, Pitard exposed that the increment delimitation error introduced the largest biases encountered in sampling. Thus, a sample must be equiprobabilistic, a process stream is always segregated, a process stream is 3-dimensional. But it is up to the company to perform 3-dimensional, 2-dimensional, or 1-dimensional sampling. For instance, for accurate metallurgical accounting, it is a must to perform 1-dimensional sampling. For effective process control, it is recommended to perform 1-dimensional sampling.

Pitard explained that stream sampling requires a constant stream, however, the delimitation does not take into account the fragments. “Another key aspect for a correct sampling is the cutters of the equipment,” said Pitard. The cutter and cutter edges must be cleaned every working shift to achieve a good sample gathering. The drive system of the cutter must not be under-designed, the parking position of the cutter must be far away from the stream: the drive of the cutter must be electric. Also, the rotating vezin sampler needs to be in a correct radio, the cutter must not collect any material when stopped in a parked position, as it can collect dust.

Other conditions to consider for extraction correctness relative to the cutter edge are the straightens, thickness, shape of the cutter edge and length. As for the cutter opening, it is important to consider its inclination and to be aware of the overall cutter width, cutter speed, depth, capacity and general design of the cutter. Also, consider the stream turbulences and stream velocity.  “Adjustable cutter edges and design of the cutter edges to avoid selectivity, I recommend sip slow outside 70 degrees to avoid misplacing,” said Pitard.