近日获知，马若鑫同学的论文Predictability Boundaries of Water-Quality Variables and Langelier Saturation Index-Based Scaling-Tendency States in Operation-Regulated Industrial Circulating Cooling Water被

Journal of Environmental Chemical Engineering接收。论文得到国家自然科学基金、陕西省杰出青年科学基金以及山东省重点研发计划的资助。摘要如下：

Industrial circulating cooling water (CCW) is strongly affected by evaporative concentration, make-up and blowdown, chemical dosing, and load fluctuations, leaving daily laboratory records noisy, incomplete, and operation-regulated. Using daily records from an industrial open CCW system, this study developed a process-informed, layered framework integrating synchronous reconstruction, one-day-ahead forecasting, and Langelier Saturation Index (LSI)-based scaling-tendency state identification. Static reconstruction indicated strong information redundancy among total hardness, calcium hardness, electrical conductivity, and chloride. A random forest model using hand-crafted temporal features forecast chloride reliably but performed poorly for operation-regulated or disturbance-sensitive variables. After multi-stage denoising, long short-term memory models achieved test R2 values of 0.8958, 0.8883, 0.8407, and 0.8045 for electrical conductivity, calcium hardness, chloride, and total hardness, respectively. Continuous LSI regression captured the central trajectory but missed peak-valley transitions. In contrast, distribution-based relative-state classification achieved an accuracy and macro-F1 of 0.72 and an F1-score of 0.80 for the relatively high-LSI state. By reframing LSI prediction as system-relative state identification, this study provides a process-informed and decision-oriented strategy for scaling-risk assessment under operational regulation.