Data Analytics for Performance Evaluation Under Uncertainties Applied to an Industrial Refrigeration Plant

Abstract

Artificial intelligence has bounced into industrial applications contributing several advantages to the field and have led to the possibility to open new ways to solve many actual problems. In this paper, a data-driven performance evaluation methodology is presented and applied to an industrial refrigeration system. The strategy takes advantage of the Multivariate Kernel Density Estimation technique and Self-Organizing Maps to develop a robust method, which is able to determine a near-optimal performance map, taking into account the system uncertainties and the multiple signals involved in the process. A normality model is used to detect and filter non-representative operating samples to subsequently develop a reliable performance map. The performance map allows comparing the plant assessment under the same operating conditions and permits to identify the potential system improvement capabilities. To ensure that the resulting evaluation is trustworthy, a robustness strategy is developed to identify either possible new operation conditions or abnormal situations in order to avoid uncertain assessments. Furthermore, the proposed approach is tested with real industrial plant data to validate the suitability of the method.

Improving the Energy Efficiency of Industrial Refrigeration Systems by Means of Data-Driven Load Management

“…The reference day is the most similar day found in one year of historical operation taking into account some key parameters or variables: the suction pressure, discharge pressure, and cooling capacity of the available historical dataset, which are the main variables that affect the compressor performance, and can be used as a reference to compare similar days in operational and load demands. Further information of the importance of those variables is explained in our previous work [34].…”

A Data-Driven-Based Industrial Refrigeration Optimization Method Considering Demand Forecasting

“…Furthermore, as can be seen in the exhaustive study made by Domingues et al [23], the selected algorithm presents high robustness against noise, high dimensionality and stability. The details of the novelty model are explained in more detail in our previous work [17].…”

“…With the system operation discretized and the outliers filtered, the best historical setpoints of each neuron of the grid are selected, in B4, to obtain the best PLR recommendations that the system has ever achieved. All the phases described below are part of our previous work [17], where they are comprehensively described.…”

“…To address such limitations, a near real-time data-driven methodology for the PLR problem is proposed. The method provides robust setpoint recommendations, considering the variables that affect the operation and the uncertainties associated with the data, employing a discretization technique and a novelty model as in our previous work [17]. Therefore, the exposed limitations of the data-driven techniques are addressed using a near real-time setpoint proposal, which reduces the modelling uncertainties, as well as a proliferation strategy [18], in order to create new optimal PLR configurations under each operational condition.…”