Mixing uniformity is the most important index to evaluate the performance of the mixer.

1. The meaning of powder mixing Powder mixing is a process in which more than two components are mixed in a dry state or in the presence of a small amount of liquid, and the non-uniformity is continuously reduced by external force. The so-called two or more components can be different substances or the same substance with different physical characteristics: such as different water content, different particle diameters, different colors, etc.

Powder mixing is a complicated random process, and the evaluation and determination method of mixing quality has always been a thorny problem that puzzles people. With the development of the times, the days of judging the uniformity of mixing based on human senses are over. The scientific and quantitative form to determine the mixing uniformity is the quantitative analysis of powder mixing. To achieve quantitative analysis, there must be several processes of sampling, testing, and statistical analysis (data processing), so as to obtain a single quantity to express the uniformity of the mixture.

2. Sampling Take a small amount of material from a certain position in the mixture and call it “sampling”. This small amount of material is called “observing sample” and is also called “point sample”, and the sampling position is called “sampling point”.

In the same container, at the same time level, samples obtained from different sampling points constitute the "sample" at this time level. The number of spot samples is the size of the sample.

Regarding the size of the sample: on the premise of satisfying the amount required for the test and possibly representative of the materials around the sampling point, the smaller the sample, the better. If the sample is too large, it not only wastes materials, but also is detrimental to the accuracy of quantitative analysis.

Regarding the number of samples (that is, the size of the sample): the more samples, that is, the larger the sample, the more reliable the quantitative analysis results and the smaller the error rate. But so far, the determination of the optimal number of samples has not been studied. According to the recommendations of the American Chemical Engineering Association, 5 to 15 samples are required, and some people think that at least 20 or even 50 samples are required. Our country customarily takes 5 ~ 10 samples. Regarding the location of the sampling point, when the material is sampled in a stationary state, the sampling point should be distributed as uniformly as possible at each position of the material. If the sample can be sampled in the motion flow of the mixture, the analysis result is more accurate than the sample obtained at rest. Therefore, when conditions permit (such as the optimal mixing time has been determined), it is good to sample in the exit stream of the mixer.

3. Detection The samples will be taken, and the content of each component (especially the key component-tracer) will be determined by chemical or physical methods: X. If the number of samples is 5, then 5 results are detected: X1; X2; X3; X4; X5. Regarding the detection method, it should be determined by the nature of the components, the purpose of mixing and the actual conditions.

Fourth, statistical analysis The above test results are calculated by statistical methods to obtain a single value to evaluate the mixture quality of the mixture, called statistical analysis method.

1. The arithmetic mean, also known as the sample mean: X

X = — Σxi

In the formula: n——the number of sample points (sample size)

Xi——parameter measured by the i-th sample. Such as quality, water content, particle number, etc.

2. Variance: σ2

σ2 = —— Σ (Xi — X) 2

3. Standard deviation, also known as SD (short for Standard Deviation): σ

σ = √σ2 = √——Σ (Xi – X) 2

Before the 1940s, σ was used to define (assess) the quality of powder mixing. The smaller the σ value, the more uniform the mixing. However, practice shows that this definition is very inaccurate in some cases, and the error is large. British powder engineering authority N. Hahnbey said in his monograph "Mixing Process in Industry" (China Petrochemical Press 1991): "... the practical variance obtained from the sample

(Σ2) The value has little value unless it can be correlated with the limit variance value ".

4. Relative standard deviation, also known as RSD (Relative Standard Deviatio): V V has many kinds of algorithms, only two commonly used algorithms are provided below.

Algorithm 1: V = × 100%

Algorithm 2: V = (× 100%)