A methodology for electricity monitoring and targeting (M&T) in an Irish precision engineering SME

Abstract

Energy management in small to medium enterprises (SMEs) remains undeveloped due to competing priorities and a lack of specialist knowledge. However considerable savings can be demonstrated where companies take the time to investigate their energy use. Savings of over 20% can be achieved through changes to operational and behavioural practice. Additional benefits, such as improved production tracking and improved maintenance, can been seen which add to the value in undertaking an energy monitoring and targeting (M&T) plan. The method described involves the application of a series of ‘virtuous’ cycles of engagement and energy savings that can be applied from the highest factory level right down to a specific machine. The analysis of overall energy use from utility bills and the visualisation of typical machine power profiles aid in understanding the drivers of energy consumption and in engaging management in energy efficiency. The monitoring of specific machines in production highlights the significant consumption of electricity during non-productive times. The development of energy performance indicators is described for product variations which can be useful in tendering for business and selecting optimum production pathways. The approach is illustrated with data from a case study of a precision engineering SME based in Limerick, Ireland.

Keywords:

• Energy efficiency in industry
• manufacturing
• SMEs
• energy performance indicators

Introduction

According to the European Energy Agency (EEA) (2016), industry remained the largest electricity-consuming sector in the EU-28 in 2014, accounting for 37% of all electricity consumption. According to the World Energy Outlook (WEO) report (IEA 2013), the industry sector is responsible for 29% of total final consumption of all fuels. The report states that the industry sector is very complex, and a detailed understanding of the various processes or product types is necessary to monitor energy efficiency.

Even with the energy efficiency initiatives that are underway it is estimated that energy consumption in industry will rise by another 20% between 2012 and 2020 (IEA 2013). Previous research (IEA 2012) has estimated that only approximately 40% of the potential for improving the efficiency of energy use in industry is exploited. The barriers to energy efficiency include the lack of visibility, low awareness, limited know-how and fragmentation of energy consumption (Lunt, Ball, and Levers 2014; O’Malley and Scott 2003; Sorrell, Mallett, and Nye 2011; Thollander and Palm 2012).

Sustainable manufacturing

The smart sustainable factory of the future (EU 2010) will be one where there is full integration between the production activity and the associated energy used and where the operation of the factory can be optimised around its energy and ecological impact. The Specific Energy Consumption (SEC) was defined by the Asian Pacific Energy Research Centre (APEC) (2000) as the energy consumption per unit of physical production which is commonly used as the measure of energy efficiency in industry (Phylipsen et al. 2002; Irrek and Thomas 2006). A study (IEA 2010) concluded that manufacturing industry can improve its energy efficiency by up to 26%, while reducing the sector’s CO₂ emissions by up to 32%, based on proven technology. In Denmark, Rasmussen, Nielsen, and Reinaud (2014) showed that enterprises stand to gain by saving up to 30% of their annual energy use, and increasing their productivity, through better energy management. According to Granade et al. (2009) energy management and behavioural changes can achieve up to half of the remaining energy efficiency potential in industry.

Manufacturers who participated in energy efficiency programmes also experienced significant additional cost savings (RGGI (Regional Greenhouse Gas Initiative) 2012). These Non-Energy Benefit (NEBs) include; improved production, cleaner environments, improved morale and more reliable operation. The search for energy efficiency has been shown by Porter and Linde (1995) to also lead to productivity improvements. The uptake of industrial energy efficiency projects has been shown by Lilly and Pearson (1999) to result in lower maintenance costs and replacement costs of related components. Research by the IEA (2015) has shown that if NEBs are included, the true value of the energy efficiency projects might be up to 2.5 times higher than if looking at the energy efficiency improvements alone.

The Energy Savings Opportunity Scheme (ESOS) is an initiative of the European Union to implement Article 8 of the Energy Efficiency Directive (2012/27/EU) (EU 2012). The goal of the initiative is to enforce energy audits on industry in order to encourage the wider adoption of energy savings measures. However the initiative was only applied to large undertakings (Companies employing more than 250 people or with a turnover greater than 50 million euros) and thus a key opportunity to promote energy conservation amongst manufacturing SMEs was missed.

Barriers to energy efficiency

O’Malley and Scott (2003) identified the major barriers to the adoption of energy efficiency in SMEs as; access to capital, hidden costs, imperfect information and values & organisational culture. Seow and Rahimifard (2011) stated that the detailed breakdown of energy consumption within various processes in companies is not well understood. In a significant review, Sorrell, Mallett, and Nye (2011) define the energy efficiency gap as the gap between what appears to be an attainable cost-effective level of energy efficiency and the level of energy efficiency actually observed in practice. They identified access to capital and hidden costs as the most important barriers to energy efficiency. They also highlighted that senior management are often unaware of the opportunities available. The main reasons given by Thollander and Palm (2012) for not managing energy in SMEs are; lack of time, lack of resources, lack of knowledge and a primary focus on production. Wijnants and Wellens (2013) reported that suitable energy efficiency measures in SMEs are known but not implemented. Lunt, Ball, and Levers (2014) identified the key barriers as; a lack of accountability, no clear ownership and no sense of urgency. The United States Department of the Environment, in a report to congress, identified the main barriers to energy efficiency in industry as; failure to capture the value of energy savings, lack of knowledge of incentives and risks, lack of disaggregated energy consumption data and lack of in-house technical expertise (USDoE 2015). Fawkes, Oung, and Thorpe (2016) identified the key internal barriers within firms to energy efficiency measures as; lack of knowledge, lack of finance and the improved efficiency not being regarded as strategically important.

Energy usage in production machines

Gutowski, Dahmus, and Thiriez (2006) showed that the most important characteristic of a process in relation to its energy consumption is its rate of production as the specific electrical energy requirement is often dominated by the support features of the equipment (33%) rather than the actual physical mechanism. In another case study by Rahimifard, Seow, and Childs (2014) the effective energy used to directly make the product in a metal processing factory was analysed as 48% of the total energy consumption and this was referred to as an efficient production process. According to Fysikopoulos et al. (2012) a common characteristic of almost all manufacturing processes is that even when the machine is idle, it is consuming more than 50 per cent of its maximum power. A study by Vikhorev, Greenough, and Brown (2013) showed that one of the main energy losses in the factory relates to production machine idling. For the machining line they monitored, idling accounted for 23% of the lines annual energy consumption. The report states that the idling energy losses are usually caused by inefficient operation by line personnel. With knowledge of the direct and indirect energy flows and their relationship to production activities it is possible to identify the auxiliary (non-value added) energy within production where there may be significant potential for energy reduction (Mustafaraj et al. 2015).

Methodology

The proposed approach extends the methodology described by Rivas, Hardiman, and Cosgrove (2015) to take into account the lack of existing data (energy and production) available in a typical SME.

Virtuous cycles of engagement and energy saving

As described by Moen (2009), Edward Deming modified the Shewhart cycle in 1993 and called it the PDSA (Plan, Do, Study, Act) Cycle. He described it as a flow diagram for learning, and for improvement of a product or of a process. Based on the Deming PDSA Cycle, Ishikawa (1985) redefined it as the PDCA (Plan Do Check Act) Cycle and included targets and methods for reaching the goals in the planning step. He stressed how good control meant allowing standards to be revised constantly to reflect the voices of consumers. Pakbin (2014) states that to be effective in practice, an energy management system needs constant attention and outlines how the central implementation part of the ISO 50001 Energy Management certification is derived from the PDCA cycle from the ISO 9001 (quality management) and the ISO 14001 (environmental management) standards, applied as follows;

Plan: Understand the situation and create goals and plans for improvement on the basis of energy output performance indicators (ENPIs).

Do: Put plans into action.

Check: Measure results, monitor, document and evaluate.

Act: Strengthen positive experiences and continue development across the entire organisational system and over various production teams.

Coghlan and Brannick (2001) proposed the concepts of diagnosing, planning, taking and evaluating actions in a continuous series of repeating cycles to effect change in industry. The proposed methodology extends these approaches to the engagement of ‘owners’ and ‘operators’ of a process and to enacting targeted energy efficiency measures that can be shown to deliver verified and persistent savings. The cycle is described in Figure 1. Repeated iterations of the cycle are needed as the monitoring and targeting (M&T) is bounded both at the hierarchical level of the analysis and in a temporal sense. Thus the energy efficiency measures and therefore the success KPIs need to be revised at each iteration.

Figure 1. Virtuous cycles of engagement and energy saving.

Shahin and Mahbod (2007) identified a clear gap in the absence of a hierarchy that links manufacturing and energy to provide a view of overall plant performance. Vijayaraghavan and Dornfeld (2010) described the energy consumption in metalworking and machining-based manufacturing systems and highlighted how the temporal profiles of the different levels ranging from the entire enterprise to the machine tool could be analysed and used to reduce the embedded energy of the manufactured parts. Pechmann and Scholer (2011) proposed an approach to link the factory level with defined products and down to specific machines. However, their key performance indicator (KPI) ignores the provision of technical services (compressed air, chilled water) which may be a significant component. As shown in Figure 2, this methodology proposes a hierarchical structure for the levels that need to be addressed in developing and energy efficient approach in industry.

Figure 2. Hierarchical levels of energy management in industry.

The virtuous cycles of engagement and energy savings need to be applied at the appropriate levels, as shown in Table 1. Also shown are some of the appropriate tools for each iteration of the cycle. By definition, a different set of KPIs needs to be defined at each level however they should be interlinked and where available, the lowest level KPIs should aggregate to validate the higher level ones. The methodology does not extend below the machine to the component level as the monitoring and control facilities would generally not be available. As described by Salonitis (2015), the main challenge when measuring machine tool consumption lies in the fact that machine tools are composed of a very large number of subsystems that are not possible to isolate and measure individually, thus the boundaries of the system under investigation have to be defined accurately.

Table 1. Hierarchical cycles of energy management.

	Level	Key performance indicators	Energy tools
Cycle 1	Factory	Total energy consumption	Utility energy analysis, deskstudy, electricity heatmap, waste energy cost study
Cycle 2	Function	Significant energy users (SEU)	Facilities analysis, walkthrough audit, aggregated time-of-use costs
Cycle 3	Value Stream	Specific energy consumption (SEC), Co-efficient of performance (CoP)	Process mapping, specific data logging, power profile analysis
Cycle 4	WorkCell	Energy performance Co-efficient (EnPC)	Monitoring of activity and energy consumption in realtime

Overall electricity consumption

As a rule, manufacturing SMEs have limited metering of electricity, often just the main incoming utility meter. For Cycle_1 analysis, limited electricity data may be drawn from utility bills which typically aggregate the daily consumption and which may also give a break-down between day-rate and night-rate units. Graphically displaying the daily electricity consumption for the previous six months or year, as shown in Figure 3, will give a picture of the monthly/weekly patterns of behaviour of the business. The graph clearly shows the summer and Christmas shut-down periods and the lower (but still significant) electricity consumption each weekend.

Figure 3. Annual electricity profile of manufacturing facility.

It is important for further analysis to establish estimates of the actual cost of electricity (and savings) in financial terms to the business and how it varies at different times of the day/week. This can be done through a historical analysis of the utility bills of the company over a suitable period of consistent tariff charges and structures (e.g. one to three months) and rolling up all the unit costs, distribution charges, transmission charges, levies, carbon taxes and other charges vs. the number of kilo-watt hour (kWhr) units consumed.(1)

where the data is available this can be disaggregated in to Electricity_UnitCost_Daytime and Electricity_UnitCost_Nighttime or into what-ever alternative tariff structure exists. The analysis will need to take into account any significant seasonal trends or major production shifts over the time period. If necessary a shorter or longer period may need to be analysed. For ongoing energy management in the business a monthly rolling analysis of the Electricity_UnitCost should be established as an Energy Performance Indicator (EnPI). In addition to consumption data, the utility bill also provides carbon emissions measures for the specific supplier, e.g. (Energia 2015). This data can be combined with the unit consumption data to provide the company with a metric for Electricty_UnitCarbon.

Power profile

Where it is possible for the company to access 15 min interval data from their utility providers online site, additional analysis methods become available. Combining online (and historical analysis) of electricity consumption with empirical knowledge of the production profile and history identifies high-level patterns and opportunities for further investigation. Where time-stamped data of overall electricity consumption is not available then the installation of a data-logger is necessary for sufficient weeks to establish the power profile.

One Cycle_1 approach is to aggregate energy consumption data against weekly working hour schedules, e.g. shift1, shift2, night-time, weekend. Whilst there may be variations from time-to-time an initial analysis against the scheduled productive and non-productive times does provide an estimate of the potential waste of electricity through idle running of equipment. This estimate is a useful means to engage the attention of the business owner/factory manager. For example, Figure 4 shows the aggregated weekly electrical consumption profile for a manufacturing plant and indicates the scale of weekend electricity consumption (20%) compared to known weekend production volumes (5%), thus highlighting a significant opportunity for savings.

Figure 4. Aggregated weekly power profile (kWhrs) for a manufacturing plant.

With the support of the owner established, a Cycle_2 approach using temporary logging may be suitable to investigate the potential savings estimated. The ‘owner’ of the process may now be the Facilities Manager or Operations Manager who has responsibility for that function in the factory. One useful approach, if appropriate, is to carry out data-logging through a shut-down period such as Christmas, holiday weekend, etc. in order to establish a baseline of non-productive power consumption. For example, in the manufacturing plant illustrated in Figure 5 the electricity consumption over the Christmas shut-down accounted for a waste of €1000 over six days. Whilst some of this electricity consumption may have been necessary for security, IT and facilities purposes, the monitoring showed that the majority of this was associated with machines in idle mode. Assuming that the same idle-power consumption continues through-out all non-productive times in the year (nights/weekends), the waste accounts for up to 30% of the factories’ total electricity consumption.

Figure 5. Base-load through Christmas shut-down.

For ongoing energy management a monthly rolling analysis of the Electricity_Non-ProductiveConsumption should be established as an Energy Performance Indicator (EnPI).

Another Cycle_1 approach is to use detailed interval data to create a visualisation of the electricity consumption data using conditional formatting in Microsoft Excel. Figure 6 shows an extract from an electricity map spanning 5 weeks from the 1 June 2015 for a manufacturing plant. The rows represent calendar days and the columns are aggregated to 30 min data. Normally the full 24 h period is included, but in the diagram below an excerpt from 4:00 am to 5:00 pm is shown for clarity. The non-productive weekends (green rows all the way across) can be clearly seen as can the low levels of electricity use before 8:00 am (green area on the left). The first week is notable as June 1st was a bank holiday and there were four productive days that week with a noticeable difference in pattern – less electricity consumed on Tuesday and Wednesday and greater than average later in the week. This can be explained by a variation in client orders due to the holiday weekend. In conjunction with interview of the production personnel the profile peaks can be investigated to understand what activity in the factory drives the major periods of consumption. In addition, the availability of real-time production tracking data, machine-level data or other potential proxy data at the times of peak consumption should be identified.

Figure 6. ‘Heat Map’ of electricity consumption.

Production process analysis

Based on the information from the Cycle_1 approach the significant production process steps and/or technical services can be identified. With clear savings estimated and measurable metrics available, the owner is generally sufficiently engaged to promote the next level of energy efficiency analysis. For the significant production processes, both the throughput (i.e. batch size) and the cycle time for each unit of manufacturing (i.e. cycle time/batch) should then be analysed. Name-plate data and/or maintenance records should provide data on the rated power consumption of the machine. Where that is insufficient, temporary non-invasive metering and logging of representative power consumption should be carried out. From this information it is possible to analyse the auxiliary energy or waste that occurs during productive time (Seow and Rahimifard 2011).(2)

Therefore, the auxiliary electricity consumption for a specific Machine is;(3)

This can be compared historically against past performance, set against a target to be achieved and used as a measure of comparison between machines carrying out the same function. Monitored over time, this metric also gives a clear measure of the machine loading or utilisation rate. Using the Electricity_UnitCost developed above, the potential savings that can be accrued over a time period (shift, week, month, etc.) by increasing the capacity loading or by turning off the machine during idle times can be calculated.

In addition, selecting a product type as the basis for analysis can provide specific EnPIs which identify the embedded energy cost per production part. Figure 7 shows a comparison in machine electricity consumption for the same part profile completed in steel vs. one completed in aluminium. This is particularly useful for manufacturing SMEs where the knowledge of specific cost, for example, per kg of steel, would be useful in tendering for work and in selecting optimum production pathways.

Figure 7. Material profile variation on a machine.

Empirical case study

The use-case was a precision-engineering facility based in Ireland (Manufacturing_SME1). The company employees over 40 skilled personnel and has been operating for 20 years. They primarily process parts in steel, aluminium and plastic for applications in the medical devices industry. The principle objective for the company was to quantify and determine the individual cost per part manufactured and the carbon impact of this energy usage. In addition they wished to increase their understanding of the factors that affect the sites energy consumption with a view towards reducing energy consumption and costs in production.

Overall energy patterns

Manufacturing_SME1 operates 5 days a week with a day shift only. An analysis of their electricity bills from April 2014 to Jan 2015 was carried out, showing electricity consumption from the grid of 260,000 kWh. As the company operates mainly from 8:00 until 16:30, this is the period that has high energy use, however significant energy use and cost are evident for non-productive times. The number of units consumed during the day and night periods were analysed together with the various costs itemised in the electricity bills to establish the actual monetary cost of a unit of electricity during the day and night periods. This provided the following metrics (Table 2);

Table 2. Energy performance indicators.

Electricity_UnitCost_Daytime	€0.2195 per kWh
Electricity_UnitCost_Nighttime	€0.1652 per kWh

Analysis of the electricity utility bills day and night values (two values per day) was carried out to identify the amount of energy used in productive and non-productive times. A baseline was established taking night time values and making the assumption that a similar amount of energy was being consumed during the evening after production from 16:30 at the day rate until 23:00 when the night rate begins, it can be estimated that at least 40% of the total energy usage is related to non-productive times giving a monetary value for potential savings of up to €23,000 per annum.

A significant amount of this energy is due to machines being left on with fans, heaters, components etc. running. From a shutdown periods over the Christmas holidays, the total electricity usage was as low as 125 kWh for a 24 h period. This compares to a minimum of 200 kWh for a 24 h period at weekends and bank holidays throughout the year. Thus, just stepping equipment at weekends back to the level achieved during the shutdown could provide savings of 8000 kWh per annum.

Presentation of the Cycle_1 analysis to the Facilities and Maintenance Manager quantified the cost of the losses associated with keeping the machines and technical services in idle mode during non-productive times. Savings were immediately achieved through changes to the factory end-of-shift procedures. Savings were also achieved through changes to the factory’s max-import capacity thereby reducing utility standing charges and through repairs to the power-factor equipment which eliminated utility penalty charges.

The follow-up Cycle_2 analysis included a walk-through audit of the facility which highlighted the compressed air as a significant energy user (SEU) in production and the presence of leaks. Temporary monitoring of the compressor at non-productive times showed the cost of supplying the leaks to be approximately 34% of the total energy for compressed air and estimated as a loss of €5630 per year. A repair and maintenance programme was initiated which eliminated most of these losses.

Production process analysis

Cycle _3 analysis was then directed towards the production machines which were estimated to be the next most significant energy users. After initial analysis and spot measurements, a number of production machines were selected for further monitoring using 3-phase dataloggers.

The supply to a Hi-Turner machine was monitored over a two-week period in March 2015 as shown in Figure 8. The production data for the parts produced were compared with the energy consumed. It can be seen that the machine was only productive for 3 days of this period and while it was in idle mode it consumed approximately 33% of its full load usage.

Figure 8. Hi-turner current profile.

From the production figures, 25 parts (Product A) were manufactured each day. Each part cost approximately €0.23 in electricity to manufacture on this machine with an associated carbon emission impact of 0.56 kg. It is calculated that the machine electricity cost approximately €5.80 for each production shift. While the machine was in the idle or waiting state, it cost €5.89 per day or €0.25 per hour in comparison with a productive cost of €0.64 per hour. Further investigation is necessary to verify that the machine can be fully isolated when not required without affecting production operations.

An Okuma Lathe was monitored for a two-week period with the results shown in Figure 9.

Figure 9. Okuma Current Profile 13 March 2015–27 March 2015.

It can be seen that this machine was in operation for approximately 9 days from the measurement period of the 13th of March until the 27th March. During this period a significant level of electricity was consumed when the machine is in an idle or waiting state. In particular, there was an unusually high idle consumption overnight on the 27th March which does not correlate with production activity and may indicate that some machine maintenance is required. It was calculated that the electrical costs of running this machine vary from approximately €6.13 to €7.81 for each shift of actual production.

For example; production figures for the 16th March shows that 16 parts (Product B) were manufactured. Thus each part cost approximately €0.53 to manufacture on this machine with CO₂ emissions of approximately 1.26 kg per part. Further investigation is necessary to verify that the machine can be fully isolated when not required without affecting production operations.

A Cincon Lathe (M32) was monitored for a two-week period with the results shown in Figure 10. Although the machine is not electrically isolated after production it is not consuming large amounts of energy (approx. €0.03 per hour).

Figure 10. Cincon M32 current profile.

Energy usage during production costs €0.97 per hour with each shift costing ca. 21.1 kg in carbon emissions. Very limited savings can be achieved through isolation of the machine during non-productive times.

As well as the energy savings, the Facilities and Maintenance Manager for the factory has been particularly impressed by the useful information that can be gained through logging the electricity consumption on the machines, the potential for better understanding production operations and the ability to clearly identify the available capacity on machines. The company is investigating monitoring all its machines for energy and production purposes based on the opportunities identified in the case study.

Conclusions

Significant electricity and carbon emission savings can be achieved in precision engineering SMEs through behaviour and operational changes. The attention of SME business owners can be directed towards energy efficiency savings by presenting available electricity consumption data in graphical and visual forms that highlight the factory profile.

Approximate calculations of actual electricity unit costs and analysis of productive and non-productive times can be used to highlight and quantify potential electricity and carbon emissions savings that can be achieved through change in practice. In the use-case under study, potential savings of 40% amounting to €23,000 per annum were shown to be available.

A methodology was described to engage with the different levels of the organisation and different process ‘owners’ and ‘operators’ with the appropriate tools, KPIs and energy analysis. This approach has been shown to be successful and builds a foundation to align energy efficiency with the business priorities. Developing a clear link between the temporal profile and efficiency of the energy consumption by a specific production machine can provide useful measures for costing and organising jobs. Detailed machine level monitoring clearly identified waste due to idle running. An example was shown where 33% of total electricity consumption at a specific production machine added no value. Suitable metrics were discussed which if maintained on a continuous basis would provide simple tools in the better energy management of manufacturing in SMEs.

Notes on contributors

John Cosgrove is the Section Head of Electrical Engineering in the Department of Electrical and Electronic Engineering, Limerick Institute of Technology, Limerick, Ireland. His research interests are in energy efficiency in industry, industrial monitoring and control systems and Industry 4.0.

Frank Doyle is a lecturer of Electrical Engineering in the Department of Electrical and Electronic Engineering, Limerick Institute of Technology, Limerick, Ireland. His research interests are in energy efficiency in industry, industrial monitoring and control systems and cyber-physical systems.

John Littlewood, PhD, is the Head of the Ecological Built Environment Research & Enterprise (EBERE) Group, School of Art & Design, Cardiff Metropolitian University, Cardiff, Wales, UK. His research areas include building performance evaluation, building diagnostics and surveying and building services engineering.

Paul Wilgeroth is the Programme Directorin the School of Art & Design, Cardiff Metropolitian University, Cardiff, Wales, UK. His research areas include product design process, design for manufacture, assembly and disassembly and sustainability.

Disclosure statement

No potential conflict of interest was reported by the authors.

Funding

This research was supported by Enterprise Ireland (EI).

Acknowledgement

This research has been undertaken as part of a Professional Doctorate run by Cardiff Metropolitan University.