Search

The scientific community has shown considerable interest in Industry 4.0 due to its capacity to revolutionise the manufacturing sector through digitalisation and data-driven decision-making. However, the actual implementation of Industry 4.0 within complex industrial settings presents obstacles that are typically beyond the scope of mainstream research articles. In this paper, a comprehensive case-study detailing our collaborative partnership with a leading medical device manufacturer is presented. The study traces its evolution from a state of limited digitalisation to the development of a digital intelligence platform that leverages data and machine learning models to enhance operations across a wide range of critical machines and assets. The main business objective was to enhance the energy efficiency of the manufacturing process, thereby improving its sustainability measures while also saving costs. The project encompasses energy modelling and analytics, Fault Detection and Diagnostics (FDD), renewable energy integration and advanced visualisation tools. Together, these components enable informed decision making in the context of energy efficiency. [ABSTRACT FROM AUTHOR]

Using 10-minute wind turbine supervisory control and data acquisition (SCADA) system data to predict faults can be an attractive way of working toward a predictive maintenance strategy without needing to invest in extra hardware. Classification methods have been shown to be effective in this regard, but there have been some common issues in their application within the literature. To use these data-driven methods effectively, historical SCADA data must be accurately labelled with the periods when turbines were down due to faults, as well as with the reason for the fault. This can be manually achieved using maintenance logs, but can be highly tedious and time-consuming due to the often unstructured format in which this information is stored. Alarm systems can also help, but the sheer volume of alarms and false positives generated complicate efforts. Furthermore, a way to implement and evaluate the field deployed system beyond simple classification metrics is needed. In this work, we present a prescribed and reproducible framework for: (i) automatically identifying periods of faulty operation using rules applied to the turbine alarm system; (ii) using this information to perform classification which avoids some of the common pitfalls observed in literature; and (iii) generating alerts based on a sliding window metric to evaluate the performance of the system in a real-world scenario. The framework was applied to a dataset from an operating wind farm and the results show that the system can automatically and accurately label historical stoppages from the alarms data. For fault prediction, classification scores are quite low, with precision of 0.16 and recall of 0.49, but it is envisaged that this can be greatly improved with more training data. Nonetheless, the sliding window metric compensates for the low raw classification scores and shows that 71% of faults can be predicted with an average of 30 h notice, with false alarms being active for 122 h of the year. By adjusting some of the parameters of the fault prediction alerts, the duration of false alarms can be drastically reduced to 2 h, but this also reduces the number of predicted faults to 8%. [ABSTRACT FROM AUTHOR]

Highlights • A novel cloud computing application for measurement and verification 2.0 is presented. • IntelliMav offers a solution for near real-time energy savings verification. • Powerful machine learning regression techniques are automatically applied to energy data. • Parallel processing using powerful cloud computing infrastructure reduces resources required. • A real-world case study demonstrates the ease of use and benefits of the application. Abstract Energy conservation measures (ECMs) are implemented in all sectors with the objective of improving the efficiency with which energy is consumed. Measurement and verification (M&V) is required to verify the performance of every ECM to ensure its successful implementation and operation. The methodologies implemented to achieve this are currently evolving to a more dynamic state, known as measurement and verification 2.0, through the use of automated and advanced analytics. The primary barrier to the adoption of M&V 2.0 practices are the tools available to practitioners. This paper aims to populate the knowledge gap in the industrial buildings sector by presenting a novel cloud computing-based application, IntelliMaV, that applies advanced machine learning techniques on large datasets to automatically verify the performance of ECMs in near real-time. Additionally, a performance deviation detection system is incorporated, ensuring persistence of savings beyond the typical period of analysis in M&V. IntelliMaV allows M&V practitioners to quantify energy savings with minimum levels of uncertainty by applying powerful analytics to data readily available in industrial facilities. The use of a cloud computing-based architecture reduces the resources required on-site and decreases the time required to train the baseline energy model through the use of parallel processing. The robust nature of the application ensures it is applicable across the broad spectrum of ECMs in the industrial buildings sector. A case study carried out in a large biomedical manufacturing facility demonstrates the ease of use of the application and the benefits realised through its adoption. The energy savings from an ECM were calculated to be 2,353,225 kWh/yr with 25.5% uncertainty at a 90% confidence interval. [ABSTRACT FROM AUTHOR]

The foundations of all methodologies for the measurement and verification (M&V) of energy savings are based on the same five key principles: accuracy, completeness, conservatism, consistency and transparency. The most widely accepted methodologies tend to generalise M&V so as to ensure applicability across the spectrum of energy conservation measures (ECM’s). These do not provide a rigid calculation procedure to follow. This paper aims to bridge the gap between high-level methodologies and the practical application of modelling algorithms, with a particular focus on the industrial buildings sector. This is achieved with the development of a novel, machine learning supported methodology for M&V 2.0 which enables accurate and reliable quantification of savings. A novel and computationally efficient feature selection algorithm and powerful machine learning regression algorithms are employed to maximise the effectiveness of available data. The baseline period energy consumption is modelled using artificial neural networks, support vector machines, k-nearest neighbours and multiple ordinary least squares regression. Improved knowledge discovery and an expanded boundary of analysis allow more complex energy systems be analysed, thus increasing the applicability of M&V. A case study in a large biomedical manufacturing facility is used to demonstrate the methodology’s ability to accurately quantify the savings under real-world conditions. The ECM was found to result in 604,527 kWh of energy savings with 57% uncertainty at a confidence interval of 68%. 20 baseline energy models are developed using an exhaustive approach with the optimal model being used to quantify savings. The range of savings estimated with each model are presented and the acceptability of uncertainty is reviewed. The case study demonstrates the ability of the methodology to perform M&V to an acceptable standard in challenging circumstances. [ABSTRACT FROM AUTHOR]

Accurate energy modelling is a critical step in the measurement and verification (M&V) of energy savings, as a model for consumption in the baseline period is required. Machine learning (ML) algorithms offer an alternative approach to train these models with data-driven techniques. Industrial buildings offer the most challenging environment for the completion of M&V due to their complex energy systems. This paper investigates the novel use of ML algorithms for M&V of energy savings in industrial buildings. This approach enables the extension of the traditional project boundary also. The ML techniques applied consist of bi-variable and multi-variable ordinary least squares regression, decision trees, k-nearest neighbours, artificial neural networks and support vector machines. The prediction performances of the models are validated in the context of a biomedical manufacturing facility to find the optimal model parameters. Results show that models constructed using ML algorithms are more accurate than the conventional approach. A 51.09% reduction in error was achieved using the optimal model algorithm and parameters. The use of a higher measurement frequency reduced the spread of error across the six models. However, further analysis proved the use of more granular data did not always benefit model performance. Results of the sensitivity analysis showed the proposed ML approach to be beneficial in circumstances where missing baseline data limits the model training period length. [ABSTRACT FROM AUTHOR]