可解释的人工神经网络作为软传感器,用于预测连续颗粒线中的水分含量
Petra Záhonyi1, Dániel Fekete1, Edina Szabó1
1Department of Organic Chemistry and Technology, Faculty of Chemical Technology and Biotechnology, Budapest University of Technology and Economics (BME), Műegyetem rkp. 3, H-1111 Budapest, Hungary.
概括
可解释的人工神经网络 (ANN) 可以作为可靠的软传感器,用于连续制药制造. 这项研究表明,ANN可以准确估计水分含量,提高过程理解,并为传统方法提供具有成本效益的替代方案.
科学领域:
- 制药制造业 制药制造业 制药制造业
- 化学工程是化学工程的重要组成部分.
- 数据科学数据科学数据科学
背景情况:
- 人工神经网络 (ANN) 为制药制造提供了潜力,但由于它们的黑子性质,它们面临信任问题.
- 软传感器对于实时过程监控至关重要,但传统方法可能受到限制.
研究的目的:
- 开发和评估可解释的人工神经网络 (ANN) 作为软传感器,用于监控连续的制造过程.
- 为了估计颗粒的水分含量,仅使用工艺参数,没有直接测量.
- 将ANN性能与传统近红外 (NIR) 光谱法进行比较.
主要方法:
- 开发了两个ANN:一个多层感知子 (MLP) 和一个带有外源输入的非线性自回归 (NARX) 模型.
- 使用过程参数作为输入来预测水分含量.
- 使用SHAP分析来研究MLP模型的解释性.
- 使用离线干燥损失测量和与NIR光谱学进行比较的验证.
主要成果:
- 对于湿度含量预测,ANN模型 (MLP和NARX) 实现了与NIR光谱学相似的准确性.
- 在所有方法中,湿度含量以根平均平方预测误差低于1%的方法确定.
- SHAP分析为MLP模型提供了透明度,确定了影响预测的关键参数.
结论:
- 可解释的ANN是药品制造中的软传感传统分析方法的可行和经济高效的替代方案.
- 开发的ANN软传感器提高了连续制造过程中的流程理解和可靠性.
- 这种方法为监测关键质量属性提供了一个直角且透明的方法.
相关概念视频
01:10Moisture Content and Bulking of Aggregate
216
The moisture content of aggregates is a crucial factor in construction, particularly in concrete mixing, as it influences the total water required in the mix. Moisture content represents the water coated on the exterior surface of the aggregate existing in a saturated and surface-dry condition. The total water content of a moist aggregate is the sum of its moisture content and water absorption.
When aggregates are exposed to rain or sit in stockpiles, they absorb moisture, which must be...
When aggregates are exposed to rain or sit in stockpiles, they absorb moisture, which must be...
216
02:35Key Elements for Plant Nutrition
22.0K
Like all living organisms, plants require organic and inorganic nutrients to survive, reproduce, grow and maintain homeostasis. To identify nutrients that are essential for plant functioning, researchers have leveraged a technique called hydroponics. In hydroponic culture systems, plants are grown—without soil—in water-based solutions containing nutrients. At least 17 nutrients have been identified as essential elements required by plants. Plants acquire these elements from the...
22.0K
01:07End Point Prediction: Gran Plot
614
A Gran plot is used to predict the equivalence volume or endpoint of a potentiometric or acid-base titration without reaching the endpoint. Typically, titration data is collected as a function of the titrant's volume up to a point less than the equivalence volume and then transformed into a linear format. The straight line is extended to the x-axis, indicating the necessary titrant volume to achieve the equivalence point.
For potentiometric titration, the Gran plot is created by plotting...
For potentiometric titration, the Gran plot is created by plotting...
614