Understanding the impact of process parameters on the crystallization process within an integrated suspension melt crystallization pilot plant

<b>Project:</b>
Developing environmentally-friendly and efficient purification technologies is an inevitable trend in the chemical industry.
Freeze concentration is gaining attention due to its high separation efficiency and low energy demands, relying only on electricity, which allows for integration with renewable energy sources and could be an alternative for dewatering processes like desalination and wastewater treatment.
To utilize the full potential of freeze concentration, the solid-liquid separation and purification of ice crystals are essential, and they are efficiently combined in continuously operated wash columns.
However, the complex operation of these columns and their connection to the suspension crystallization unit requires a fundamental understanding of the effects of different process parameters on the operating window and the quality attributes of the suspension, especially the particle size distribution.
Using a simple binary aqueous substance system with sodium chloride, the operating window of the scraped cooling crystallizer with a forced circulation loop is determined regarding the three main process parameters: the scraper rotational rate, the volume flow rate in the circulation loop, and the cooling rate.
A vibration measurement is implemented and validated as a suitable tool to detect unstable process conditions, such as a crystal layer formation on equipment walls. The effects of the process parameters on the particle size distribution are quantified and optimized to achieve a desired large median particle diameter and a narrow size distribution.
Here, the volume flow rate highly significantly affects the particle size distribution and the interaction with the scraper rotational and cooling rate shows a significant effect. Prioritizing the median particle diameter results in a diameter of 553 µm with a low volume flow, low scraper rotational, and low cooling rate. However, the stirred tank behavior results in a relatively broad distribution of 510 µm.

<br><br>
<b>Data set:</b>
The dataset provides a comprehensive collection of raw experimental data, processed analyses, and visual representations of crystallization experiments. The focus is on understanding the influence of scraper rotational rate, volume flow rate, and cooling rate on layer formation, particle size distribution, and vibration behavior at various operating points.

General Structure:
The dataset consists of three main categories:
<ul>
<li>Figures: opju files and PNG files (OriginLab project files containing experimental data, plots, and regressions and the corresponding pictures)
</li>

<li>Raw_data: Excel files with raw data and analyses</li>

<li>Processed: data: Excel files with processed data</li>
</ul>

Operating Parameters (A, B, C):

The experiments were conducted with different combinations of the following parameters and four center point experiments:
<ul>
<li>A: Scraper rotational rate → 140 rpm or 70 rpm</li>

<li>B: Volume flow rate → 35 Hz (20 m³/h) or 15 Hz (8 m³/h)</li>

<li>C: Cooling rate → 3 K/10 min or 1 K/10 min</li>
</ul>

File Contents and Data Description:
Figures: opju Files (OriginLab Project Files)
<ul>
<li>Figure 1 (Sheet: SLE NaCl)
Experimental data on the solubility of sodium chloride in water
Quadratic regression with equation in comments</li>

<li>Figure 5 (Sheet: CP1-4)
Raw data from center point experiments CP 1-4
Mean values and standard deviation</li>

<li>Figure 6 (Sheet: Data points for plotting)
Graph with red squares (layer formation) and green spheres (no layer formation)</li>

<li>Figure 7 (Four different sheets: OPx and OPy)
Operating parameters A (scraper rotational rate), B (volume flow rate), and C (cooling rate)
Time data in seconds/minutes, energy input, temperature measurements</li>

<li>Figure 8 (Sheets: Effects & Regression)
Effects of operating parameters on particle size distribution
Regression values for d50 and d90-d10, standardized between 0 and 1</li>
</ul>
Image Files (PNG)
<ul>
<li>Figure 1: Experimental data on NaCl solubility with quadratic regression</li>

<li>Figure 5: Temperature trajectories, scraper blade drive current, and standard deviation of crystallizer acceleration</li>

<li>Figure 6a: Effects on layer formation with confidence intervals</li>

<li>Figure 6b: Operating window (stable/unstable conditions)</li>

<li>Figure 7a-d: Temperature and scraper current at different operating points</li>

<li>Figure 8 & 9: Influence of process parameters on particle size distribution</li>
</ul>
Processed data: Excel Files
Layer_new_X_Y and Layer_new_CP
<ul>
<li>Excel tables with raw vibration data of the operating points (X, Y, or CP = center points) and acceleration evaluation</li>

<li>Two sheets:
High scraper rotational rate (excluding center point experiments)
Low scraper rotational rate (excluding center point experiments)</li>

<li>Key columns:
Time data (seconds): Start time when process medium reaches 1°C
Acceleration values (xyz-direction): Raw data and derived metrics
Standard deviation & mean values over different time spans</li>
</ul>
SI: Evaluation DoE (Design of Experiments)
<ul>
<li>Summary of conducted experiments with respective operating points and parameters</li>
<li>Abbreviation: SD = Standard Deviation</li>
</ul>
Raw Data: Excel Files and videos
<ul>
<li>File name format: OperatingPoint_ScraperRotationalRate_VolumeFlow_CoolingRate_date</li>

<ul>
<li>Contains:
Concentrations:
Measured sample weights in feed and concentrate containers
Calculation of NaCl weight fraction</li>
<li> Results:
Raw data from the human-machine interface (HMI)
Temperature data (process medium, melt loop, cooling medium)
Scraper blade energy input</li>
</ul>

<li>For experiments without layer formation:
Additional vibration data (acceleration in x, y, z directions)</li>
</ul>
Particle Size Distribution (PSD) Results
<ul>
<li>File name format: PSDResults_OperatingPoint_ScraperRotationalRate_VolumeFlow_CoolingRate_date</li>
<li>Contains:
Cumulative distribution function (Q_i) → Characteristic variables and summation function
<ul>
<li>Characteristic parameters:
X_10, X_50 (median), X_90 (percentile values)
Agglomeration degree (Ag)</li>
<li>Probability density function (q_i):
Size distribution data for single crystals and agglomerates</li>
</ul>

<li>Video Folder: QICPICVideos_OperatingPoint_ScraperRotationalRate_VolumeFlow_CoolingRate_date
Contains particle videos (only for experiments without layer formation)</li>
</ul>

<b>项目：</b>
研发环保高效的提纯技术是化工行业的必然发展趋势。
冷冻浓缩技术因分离效率高、能耗低而备受关注——其仅需电力驱动，可与可再生能源整合，有望成为脱盐、废水处理等脱水工艺的替代方案。
为充分发挥冷冻浓缩的技术潜力，冰晶的固液分离与提纯环节至关重要，而连续运行的洗涤柱可高效整合该两项流程。
然而，此类洗涤柱的复杂操作逻辑，以及其与悬浮结晶单元的衔接方式，要求我们深入掌握各类工艺参数对操作窗口及悬浮液品质属性（尤其是粒径分布）的影响规律。
本研究以氯化钠-水二元简单水相体系为研究对象，针对带强制循环回路的刮刀式冷却结晶器，围绕三大核心工艺参数——刮刀转速、循环回路体积流量及冷却速率，确定其操作窗口。
本研究采用振动检测手段，并验证其可有效识别工艺不稳定工况，例如设备壁面的晶层结垢现象。
研究对工艺参数影响粒径分布的规律进行量化与优化，以实现目标大粒径中值与窄分布的悬浮液品质。
其中，体积流量对粒径分布具有极显著影响，而刮刀转速与冷却速率的交互作用同样存在显著影响。
当以粒径中值为优化目标时，在低体积流量、低刮刀转速及低冷却速率工况下可获得553 µm的粒径中值；但若出现搅拌槽式流态，则粒径中值为510 µm，且分布相对较宽。

<br><br>
<b>数据集：</b>
本数据集收录了结晶实验的原始实验数据、处理后分析结果及可视化表征结果，核心目标是揭示刮刀转速、体积流量及冷却速率在不同操作工况下对晶层形成、粒径分布及振动特性的影响规律。

数据集整体结构：
数据集包含三大核心类别：
<ul>
<li>图像文件：opju格式文件与PNG图像文件（均为OriginLab项目文件，包含实验数据、绘图及回归分析结果与对应可视化图片）
</li>

<li>原始数据：存储原始实验数据与分析结果的Excel文件</li>

<li>处理后数据：存储经处理的实验数据的Excel文件</li>
</ul>

操作参数（A、B、C）：
本实验采用以下参数的不同组合开展，并设置4组中心点实验：
<ul>
<li>A：刮刀转速 → 140 rpm或70 rpm</li>

<li>B：体积流量 → 35 Hz（对应20 m³/h）或15 Hz（对应8 m³/h）</li>

<li>C：冷却速率 → 3 K/10 min或1 K/10 min</li>
</ul>

文件内容与数据说明：
图像文件：opju格式文件（OriginLab项目文件）
<ul>
<li>图1（工作表：SLE NaCl）：氯化钠在水中的溶解度实验数据，附带带注释的二次回归方程</li>

<li>图5（工作表：CP1-4）：中心点实验CP1至CP4的原始实验数据，包含平均值与标准差</li>

<li>图6（工作表：绘图数据点）：以红色方块表示晶层形成工况、绿色圆点表示无晶层工况的绘图数据</li>

<li>图7（共4个工作表：OPx与OPy）：包含操作参数A（刮刀转速）、B（体积流量）、C（冷却速率），以及以秒/分钟为单位的时间数据、能耗输入与温度测量数据</li>

<li>图8（工作表：影响因素与回归分析）：操作参数对粒径分布的影响规律，包含标准化至0~1区间的d50与d90-d10回归值</li>
</ul>

PNG图像文件：
<ul>
<li>图1：氯化钠溶解度实验数据与二次回归拟合曲线</li>

<li>图5：温度变化曲线、刮刀驱动电流与结晶器加速度标准差</li>

<li>图6a：带置信区间的晶层形成影响因素分析</li>

<li>图6b：操作窗口（稳定/不稳定工况）</li>

<li>图7a-d：不同操作工况下的温度与刮刀电流变化</li>

<li>图8与图9：工艺参数对粒径分布的影响规律</li>
</ul>

处理后数据：Excel文件
文件命名格式：Layer_new_X_Y与Layer_new_CP
<ul>
<li>包含各操作工况（X、Y或CP=中心点）的原始振动数据与加速度评估结果的Excel表格</li>

<li>包含2个工作表：
- 高刮刀转速工况（不含中心点实验）
- 低刮刀转速工况（不含中心点实验）</li>

<li>核心列字段：
时间数据（秒）：工艺介质达到1℃的起始时刻
加速度值（x/y/z方向）：原始数据与衍生指标
不同时间跨度下的标准差与平均值</li>
</ul>

SI：试验设计（Design of Experiments, DoE）评估表
<ul>
<li>已开展实验的汇总表，包含对应操作工况与参数信息</li>

<li>缩写说明：SD = 标准差（Standard Deviation）</li>
</ul>

原始数据：Excel文件与视频文件
<ul>
<li>文件命名格式：OperatingPoint_刮刀转速_体积流量_冷却速率_日期

<ul>
<li>包含以下内容：
浓度数据：进料与浓缩液容器内的样品称重结果，以及氯化钠质量分数的计算结果
实验结果：人机界面（Human-Machine Interface, HMI）采集的原始数据、工艺介质/熔体回路/冷却介质的温度数据，以及刮刀能耗输入数据</li>
</ul>
</li>

<li>针对无晶层形成的实验，额外包含x/y/z方向的加速度振动数据</li>
</ul>

粒径分布（Particle Size Distribution, PSD）结果
<ul>
<li>文件命名格式：PSDResults_操作工况_刮刀转速_体积流量_冷却速率_日期</li>

<li>包含以下内容：
累积分布函数（Cumulative Distribution Function, Q_i）：特征变量与求和函数
<ul>
<li>特征参数：X₁₀、X₅₀（粒径中值）、X₉₀（百分位值），以及团聚度（Agglomeration Degree, Ag）</li>

<li>概率密度函数（Probability Density Function, q_i）：单晶与团聚体的粒径分布数据</li>
</ul>
</li>

<li>视频文件夹：QICPICVideos_操作工况_刮刀转速_体积流量_冷却速率_日期
存储颗粒观测视频（仅针对无晶层形成的实验）</li>
</ul>