Hi, please help me to find the correct sampling method. For my experiment, I have to sample the waste products for proximate analysis (in terms of carbohydrates, protein and fats). However, the amount of waste product produce (per tray) daily is inconsistent. For example: Monday I have 2 trays; Tues 5 trays, Wed 1 tray. Is there an recommended sampling amount (%) I should use to have a better representative of the nutritional content of waste produced each day?
the sampling depends on the aim of the ultimate aim of the analysis.
types of waste and composition. However, since your target seems to get a representative sample. you need a historical data about the amount since you can decides . For example, if you have a low variation you choose any day. But if you have a high variations... you may select three samples . one from a low genration, medium and high generation tray rate.
Hello Rui En.
First question is: what should your sample represent. If I understand you right, it should represent the daily waste, independent of its amount, rather then for example the week. In that case you can consider the days independently, and - if I got that right - want to make sure the accuracy of the sampling is consistent by changing sample mass.
Let's first look at the other case: if you want to represent a week for example, you should not sample the trays differently depending on the number of trays per day, since this would violate the fundamental sampling principle: "Each particle must have the same probability of ending up in your sample".
So let's assume you keep the sample mass per day consistent, although you want so sample the week's waste. Now the amount per tray on a 2-trays day is half the amount as on a 1-tray-day. Hence the particles from the latter have twice the probability of ending up in your sample.
Now if you want values for each day independently the case is different. If you sample correct you have two remaining sampling "errors" the correct sampling errors. One of them, the fundamental sampling error (FSE), is influenced by sample mass. There is an equation by Pierre Gy for it, and a simplification for "simple particles" in Wavrer (2018). What they show is that as long as the weight of the lot to be sampled is much higher than the sample mass, the weight of the lot hardly influences the FSE. So unless you are usually taking let's say 10% of the tray I would keep the total sample mass per day more or less consistent independent of the number of trays. If you want to be more exact, you should still be able to calculate the sample mass for 3 trays that produces the same FSE as your reference sample mass for 1 tray, without inserting all the unknowns for particle masses or heterogeinit invariant (see equations), since you always refer to the same kind of material with a similar heterogeinity invariant.
You can find further explanation on sampling errors for example in the attached paper. There you also find the equationsentioned (equations 5-7)
- Badges
- Science topic
- Similar topics
- Citations