1. Question B, fourth question: Use your model to obtain 325 data samples (see Appendix 4 "325 data sample data.xlsx"), here we talk about "325 data samples" before and after, the actual corresponding should be in Appendix 1. Instead of Annex IV.
2. Question 1 of question B: You can compare whether the data obtained by preprocessing is different from the data given by the manufacturer. And because the sampling processing methods are not exactly the same, there may be slight differences that do not affect the results.
3. The level unit of the raw material buffer tank (C50 in the standard table). The three excel tables are inconsistent. There are %, more than 100, and more than 100,000?
This variable can be ignored
4. The meaning of some operating variables is not clear, and it is impossible to know the specific physical meaning of the
operating variables. The meaning of the operating variables is not important, as long as there is a Chinese name or code name.
5. Data processing before and after 2: To make an analogy: the octane number is measured in 8 points, then the octane number is actually the result of the influence (produced) of various variables in the first two hours (2 hours have passed), that is The value of each variable cannot be a value of 8 points
6. Decrease rate: For example, the octane number of the gasoline raw material is 90. After the refined treatment, the octane number of the product is 89, and the octane number has lost 1 unit. If you optimize the octane number of the gasoline product to 89.5, then the octane number The alkane number loss is only 0.5 units, which is equivalent to a 50% reduction in the octane number loss.
7. Is the screening variable reduced or combined?
The variables required to be screened in the question are representative and independent
8. Δ value: For example, the optimized reaction temperature should be increased by 10 degrees, but in actual production, if it is directly increased by 10 degrees, there will be large fluctuations in production, which is very taboo, so it is often slightly increased , Increase the temperature by 2 degrees at a time, and adjust the temperature 5 times in this way, so that the production is relatively stable. The value of Δ is the magnitude of each adjustment of a certain variable.
9. Negative Δ value: for example, from -10 to 10 (from small to large), Δ is added. If the variable is to be adjusted from large to small, Δ is reduced.
10. Null value: The data setting method is for all samples (data), not every sample encounters this kind of null value phenomenon. If you don’t encounter it, you don’t have to consider
11. Null value and 0: Check whether the variable value range exceeds 0.
12. Attachment 3 contains the specific time data of sample No. 313, and the fifth question mentioned "gradual change". What kind of picture should be drawn in the fifth question?
According to the optimization, combined with the allowable adjustment range of each main operating variable, the corresponding octane number and sulfur content change trajectory of gasoline is given.
13. Location: It is the data sampling point of the instrument, in fact, each location is a variable
14. Attachment 4: The collated sample has nothing to do with Attachment 4. Attachment 4 is only used for optimization
15. "According to the process requirements and operating experience, summarize the operating range of the original data variables, and then use the maximum and minimum limiting method to remove some samples that are not in this range", the operating range refers to the range of data in the 325 sets of samples;
16. The 3σ criterion is used for the treatment of non-manipulated variables.