Abstract:The steam drum water level is a critical parameter that directly impacts the safety and efficiency of power plant operations. However, predicting the drum water level in boilers is challenging due to complex non-linear process dynamics originating from long-time delays and interrelations, as well as measurement noise. This paper investigates the application of Transformer-based models for predicting drum water levels in a steam boiler plant. Leveraging the capabilities of Transformer architectures, this study aims to develop an accurate and robust predictive framework to anticipate water level fluctuations and facilitate proactive control strategies. To this end, a prudent pipeline is proposed, including 1) data preprocess, 2) causal relation analysis, 3) delay inference, 4) variable augmentation, and 5) prediction. Through extensive experimentation and analysis, the effectiveness of Transformer-based approaches in steam drum water level prediction is evaluated, highlighting their potential to enhance operational stability and optimize plant performance.

What problem does this paper attempt to address?

The problem that this paper attempts to solve is the challenge of predicting the drum water level in steam boilers. Specifically, the drum water level is a key parameter that directly affects the safety and efficiency of power plant operation. However, due to the complex non - linear process dynamic characteristics, long - time delays, the inter - relationships among variables, and the existence of measurement noise, predicting the drum water level becomes very difficult. The paper proposes a model based on the Transformer architecture to predict the drum water level, aiming to develop an accurate and robust prediction framework to predict water level fluctuations and promote the implementation of active control strategies. ### Main problems 1. **Complex non - linear process dynamics**: The process dynamics in the boiler system are very complex, and traditional control methods are difficult to deal with these dynamic changes. 2. **Long - time delays**: The influence of some variables on the drum water level has a relatively long delay time, which makes prediction more difficult. 3. **Inter - relationships among variables**: There are complex causal relationships among multiple variables, and these relationships need to be accurately identified to improve prediction accuracy. 4. **Measurement noise**: There is noise in the actual measurement data, which affects the accuracy of prediction. ### Solutions The paper proposes a comprehensive framework, including the following steps: 1. **Data pre - processing**: including normalization, denoising, handling missing values and outliers. 2. **Causal relationship analysis**: Based on domain knowledge and Granger causality analysis, screen out the factors that directly affect the drum water level. 3. **Delay inference**: Determine the optimal delay time of each factor's influence on the drum water level. 4. **Variable enhancement**: Generate lag variables according to the inferred delay time. 5. **Prediction model**: Use the Transformer model for time - series prediction. Through this series of steps, the paper aims to improve the accuracy and stability of drum water level prediction, thereby optimizing the operating performance of power plants.

Transformer-based Drum-level Prediction in a Boiler Plant with Delayed Relations among Multivariates

A Fusion Transformer for Multivariable Time Series Forecasting: the Mooney Viscosity Prediction Case

PREDICTION OF PARAMETERS OF BOILER SUPERHEATER BASED ON TRANSFER LEARNING METHOD

Modeling and prediction of key parameters of circulating fluidized bed boiler based on Transformer

Enhanced Prediction of Transformers Vibrations under Complex Operating Conditions

Dynamic Real-Time Prediction of Reclaimed Water Volumes Using the Improved Transformer Model and Decomposition Integration Technology

Multi-Parameter Prediction for Steam Turbine Based on Real-Time Data Using Deep Learning Approaches

A Ladder Water Level Prediction Model for the Yangtze River Based on Transfer Learning and Transformer

Prediction Method for Power Transformer Running State Based on LSTM_DBN Network

A variational transformer for predicting turbopump bearing condition under diverse degradation processes

Steam turbine power prediction based on encode-decoder framework guided by the condenser vacuum degree

Cascade Generalized Predictive Control Strategy for Boiler Drum Level

Transformer based Multi-output Regression Learning for Wastewater Treatment

Evaluate Transformer model and Self-Attention mechanism in the Yangtze River basin runoff prediction

An Approach Based on Transfer Learning to Lifetime Degradation Rate Prediction of the Dry-Type Transformer

Transfer Learning with Transformer-Based Models for Mine Water Inrush Prediction: A Multivariate Analysis Using Sparse and Imbalanced Monitoring Data

Probing the limit of hydrologic predictability with the Transformer network

Enhancing hydrological modeling with transformers: a case study for 24-h streamflow prediction

Enhancing Nuclear Power Plant Operational Forecasting with Transformer Neural Networks: a Time-Series Data Approach

Transformer enhanced by Local Perception Self-Attention for dynamic soft sensor modeling of industrial processes.

A novel transformer-based multi-variable multi-step prediction method for chemical process fault prognosis