Energy monitoring of process systems: time-series segmentation-based targeting models
Energy monitoring systems calculate actual energy use, estimate energy needs at normal operation, track energy metrics, and highlight issues related to energy efficiency of process plants. Analysis of key energy indicators (KEIs) allows the comparison of process efficiency at different operating regimes. Based on the extracted knowledge realistic targets of KEIs can be determined. The performance of data-driven targeting models depends on how effective the operating regimes are characterized. Till now this modeling task is performed manually based on heuristic and subjective evaluation of the operation. A goal-oriented time-series segmentation technique has been developed to automate the selection of proper data used for the identification of targeting models. With the proposed novel segmentation algorithm targeting-models for different operating regions can be automatically determined. The concept of the resulted energy monitoring system is demonstrated at Heavy Naphtha Hydrotreater and CCR Reforming Units of MOL Hungarian Oil and Gas Company.
Statistical Process Control based Energy Monitoring of Chemical Processes
Advanced chemical process systems expected to maximize productivity and minimize cost and emission. Cost reduction needs Energy Monitoring and Targeting Systems that calculate actual energy usage, estimate energy needs at normal operation and highlight issues related to energy efficiency. Monitoring is based on continuous comparison of actual and estimated energy consumption. We developed Partial Least Squares (PLS) regression based targeting models that not only predict the expected value of energy consumption, but also visualize the operating regimes of the process. Soft-sensors working with PLS regression are widely used in chemical industry. The development of PLS models could be problematic because previous feature selection is needed. Since complex set of process variables determines Key Energy Indicators (KEIs) we applied Self-Organizing Map (SOM) models of that support visualization and feature selection of the process variables. Local linear target-models of different operating regions can be automatically determined based on the Voronoi diagram of the codebook of the SOM. We used Statistical Process Control (SPC) techniques to monitor the difference between the targeted and the measured energy consumption. We applied the concept of the resulted energy monitoring system at Heavy Naphtha Hydrotreater and CCR Reforming Units of MOL Hungarian Oil and Gas Company.
Feature Selection Based Root Cause Analysis for Energy Monitoring and Targeting
Energy Monitoring (EM) systems are based on monitoring the difference between targeted and measured energy consumption. Data-driven dynamic targeting models can be used to estimate values of key energy indicators (KEI). In some cases it is difficult to determine which process variables influence the KEIs. We developed an automated root cause analysis (RCA) technique to find the most important driving factors of energy efficiency. The proposed concept is based on the application of feature selection algorithms. We applied Orthogonal Least Squares (OLS) and Random Forest Regression (RFR) to find the proper set of input variables of the targeting models. The concept of the resulted energy monitoring system is applied at the Duna Refinery of MOL Hungarian Oil and Gas Company.