Smart meters for industrial energy conservation and efficiency optimization in Pakistan: Scope, technology and applications

Renewable and Sustainable Energy Reviews 44 (2015) 933–943

Contents lists available at ScienceDirect

Renewable and Sustainable Energy Reviews journal homepage: www.elsevier.com/locate/rser

Smart meters for industrial energy conservation and efficiency optimization in Pakistan: Scope, technology and applications Waleed Aslam n, Muhammad Soban, Farwa Akhtar, Nauman A. Zaffar Department of Electrical Engineering, Syed Babar Ali School of Science and Engineering, Lahore University of Management Sciences (LUMS), D.H.A., Lahore 54792, Pakistan

art ic l e i nf o

a b s t r a c t

Article history: Received 27 September 2013 Received in revised form 15 November 2014 Accepted 3 January 2015 Available online 30 January 2015

The electrical grid in most of the developing countries has inefficiencies in different areas such as transmission and distribution, power quality, grid reliability and system protection. These inadequacies in grid operations and asset protection, along with revenue leakage constitute an overall troubled energy profile. The fixes usually proposed in this regard are directed at reducing distribution network losses and improving residential and commercial demand side management. Lessons from the residential and commercial implementation of smart meters can be extended to industrial consumers to help relieve grid congestion and achieve better efficiency goals. This paper reviews smart meter technology and applications across residential, commercial and industrial sectors. We point out the areas for power quality and energy efficiency improvement within industries and propose ways for achieving them through smart meters, specifically in the context of Pakistan. We have incorporated empirical evidence from experimental setup at our university grid as proof of concept. We have also elaborated on the implementation methodology to avoid the possible pitfalls in the proposed solution. & 2015 Elsevier Ltd. All rights reserved.

Keywords: Smart meters Energy efficiency Industries Pakistan

Contents 1. 2. 3.

Introduction . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Technology description . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Residential, commercial and network implementations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3.1. Consumer-led. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4. Proposed industrial implementations . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.1. Energy auditing . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.2. Load diagrams . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.3. Demand response . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.4. Performance characteristic line. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 4.5. Fault analysis . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5. Implementation methodology. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6. Future directions . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7. Conclusion . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . References . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

933 934 936 937 938 938 938 940 940 941 941 942 942 942 942 942

1. Introduction

n

Corresponding author. E-mail address: [email protected] (W. Aslam).

http://dx.doi.org/10.1016/j.rser.2015.01.004 1364-0321/& 2015 Elsevier Ltd. All rights reserved.

Pakistan’s energy demand has sharply risen over the past few years and still continues to rise. The total electricity demand is expected to reach 49,078 MW by 2025 [1]. Over the years

934

W. Aslam et al. / Renewable and Sustainable Energy Reviews 44 (2015) 933–943

generation capacity has not increased in the same proportion as the demand which has created serious energy deficit. According to Pakistan Electric Power Company (PEPCO), the average power shortfall in Pakistan is more than 5000 MW [2] and is likely to become worse in the coming years (Fig. 1) [3]. Further the operational inefficiencies within the grid are quite huge; energy theft is prevalent on most parts of the electrical network and line losses add up to 19.7% of total generation in the country [4]. As a result, the country is facing frequent daily blackouts which have serious economic implications. The oil imports have increased to support electricity generation and resultant oil bills have sky rocketed over the past few years, adding further burden on the economy. It has been estimated that load shedding has cut about 3–4% of Pakistan’s GDP, costing about $13.5 billion per year to the country’s economy [5,6]. Converting 5000 MW average daily shortfall at the current rate of generation, this amount would have been enough to cover the cost of energy shortfall for approximately 11 years (Appendix). The solutions that are often suggested with regards to energy crisis include increase in generation capacity and demand side management. In Pakistan’s case, a prudent approach would be to look at energy conservation opportunities and incorporate more renewable energy sources into the energy mix. Refs. [7–12] review the conventional and renewable energy scenario in Pakistan and give suggestions for the efficient use of renewable technologies. Khan and Pervaiz have specifically identified the technological shortcomings of Solar PV in Pakistan and point out the areas for improvement [11]. Similarly Mirza et al. have evaluated wind energy development in the country and have given a policy perspective on development [12]. Distribution network losses are another area for efficiency improvement. As for the distribution network, [13,14] discuss the methods for identification and prevention of electricity theft using smart meters. The demand side management (DSM) solutions proposed to overcome energy deficit are usually designed for the residential and commercial consumers. The most common form of DSM that is presently practiced in Pakistan is the worst kind i.e. complete load-shedding (blackout) although there exist other forms of DSM schemes which are more flexible in terms of load constraint. Architecture for DSM in buildings has been discussed in detail in [15]. Normally the industrial demand is not incorporated into the demand management plans even though industrial demand constitutes significant portion of the total load. There are only a few studies that are specifically tailored for industries. Ref. [16,17] elaborate on the energy efficiency improvement opportunities

Demand

within industries. This paper builds upon the previous work and identifies constraints and opportunities for smart meter use specifically for industries in Pakistan. It then lays out a detailed plan for the implementation of proposed solutions. This study can be extended within reasonable limits to any developing country. To our knowledge, this work is a first of its kind in addressing energy efficiency improvement using smart meters for industries in Pakistan. Rest of the paper is organized as follows: Section 2 gives the technology description of smart meters; Section 3 elaborates the use of smart meters in residential, commercial and distribution network settings through various technical development pathways; Section 4 identifies the industrial efficiency solutions using smart meters; Section 5 details the implementation methodology; Section 6 gives out the future direction for such efficiency measures and Section 7 concludes the study.

2. Technology description Smart meters are meters that record electrical energy consumption and periodically send the data back to a central server for monitoring, analysis and control. Fig. 2 shows the features of a typical smart meter, which are further detailed as follows:


Measurement of electrical parameters: Basic smart meters mea-




Supply

140000 120000

Supply (MW)

100000 80000 60000 40000




20000 0 2005

2010

2015

2020

2025

Years Fig. 1. Demand and supply projections.

2030

2035

sure only the energy values whereas the advanced versions are capable of measuring a range of electrical parameters such as voltage, current, power, frequency and power factor [18]. These electrical parameters are important metrics in load management, load profiling and fault analysis. Most of the smart meters already rolled out in different smart metering programs in Pakistan feature only the most basic specifications. However, recent implementations have begun to use the advanced features [19]. This choice is driven by both the benefits that these additional features offer and the availability of low cost smart meters that offer these functionalities. Periodic data recording: Smart meters record different electrical parameters at particular time intervals. The granularity can usually vary from a few minutes to a few hours. This periodic recording of data provides a more detailed insight of the event occurrence and helps develop a greater understanding of load use patterns. Presently Pakistan’s entire electrical grid features legacy meters whose readings are manually recorded every month. Due to low granularity, the information that can be extracted out of these readings is quite limited. Further, in many cases the readings can’t be trusted because of the prevalent bribery amongst the linemen who record the meter readings. Therefore the load forecasts are often inaccurate and the resulting demand management practices are quite basic. Smart meters ensure highly granular and credible meter readings, and can therefore help devise effective demand management strategies [20]. In addition, smart meters generate alerts and notifications for a series of alarm conditions on different monitored parameters. These alerts are useful in identifying non-technical power losses in the electrical grid. In Pakistan, power losses and energy theft form a significant chunk of the total generation and their identification is often difficult, particularly in congested inner city areas [21]. Smart meter alerts offer a feasible solution for identifying and eliminating these system inefficiencies. Communication: Smart meters usually send data that they have recorded to a central server, where it can be accessed by the utility and the user. In some cases, the data for utility is sent directly to utility communication nodes [22]. Communication infrastructure allows different possible modes of data exchange between the server and the meters. The communication modes

W. Aslam et al. / Renewable and Sustainable Energy Reviews 44 (2015) 933–943

935

Fig. 2. Functional features of smart meter.

Table 1 Smart meter communication technologies. Communication technology GSM/GPRS

Comments


Feasible for distantly located individual



Zigbee

consumers, distribution transformers or if maintenance of communication infrastructure is not desired Wide coverage area throughout Pakistan Extremely low cost of subscription


Feasible for Home Area Network (HAN)/building automation


Suitable for harsh environments
Highly secure WLAN/Wi-Fi


Feasible for HAN and NAN (Neighborhood area network)


Performance degradation in dense mesh environments Power line communication


No major additional infrastructure required
Noise and data Attenuation
Interoperability Issues

WiMAX


Long range and Interoperable
Suitable for distribution area network (DAN)
Expensive

DASH 7


Longer range than higher carrier frequency counterparts


Eliminates the need for mesh networks in HAN Serial communication (Ethernet/RS-232)


Reliable
Low cost

can either be wired or wireless. Further, there can be near field or far field implementations of both wired and wireless technologies [23]. Sometimes there are primary and secondary modes of data exchange. Such an arrangement provides robustness and reliability in communication [24]. Importantly, the communication that the smart meters allow should be two way in order to facilitate on-demand data retrieval and remote load control as and when needed. Overall, the data communication allows near real time monitoring and control of different electrical loads. The following table shows various communication technologies that are used in smart meters [23,25] (Table 1). Both wired and wireless bi-directional communication technologies are available in Pakistan. Amongst the wired alternatives, the most common ones are Ethernet and RS 232/485 serial communication whereas the wireless technologies include zigbee, GSM and RF based communication methods [20]. Infra-red and optical communication are also available on some meters as secondary communication modes. Some metering vendors even allow setting up of a wireless hub that can connect multiple meters and can act as a gateway to central server. Of these technologies, GSM, serial communication and ethernet send data directly to the server. These technologies require little infrastructural changes in order to operate and are therefore in much greater use. On the other hand, communication modes such as zigbee require additional infrastructure to be installed on the grid before data could be received at the server and have therefore lesser market penetration. GSM/GPRS based communication for smart meters leverages the existing communication infrastructure in the country so overall costs incurred are not very high. Further, different cellular networks in the Pakistan have very low subscription charges for GSM based services [26]. Therefore in Pakistan’s context, GSM/

936

W. Aslam et al. / Renewable and Sustainable Energy Reviews 44 (2015) 933–943

GPRS communication provides great balance between the requirements, performance and cost as compared to other communication modes.

Gateway - Prepaid metering - Data Aggregation


Data storage: Smart meters have the capability of storing the










data within them [27]. This feature allows for preservation of data in case of communication failures, so the data is never lost even if it is not communicated to the central server and can be retrieved from the meter. Data storage is typically for a couple of months but can vary depending upon the granularity of data. Data storage presents unique contingency against natural failures and particularly meter tampering that is directed at disrupting the communication infrastructure. Even though the communication technologies presently used are quite reliable and foolproof, the presence of such a feature adds to reliability of the system and is particularly important for Pakistan since the country has a very strong electricity theft mafia [28]. Data collection: Data from smart meters is usually available through a data collection software or web hosting solution. The solutions are generally vendor specific and can communicate with a particular meter type. Such solutions allow monitoring, reporting, aggregation and data analysis. Refs. [29,30] elaborate on the mechanism of data collection and associated services. In Pakistan, some international and local vendors such as Schneider Electric and MicroTech Industries (MTI) offer data collection tools alongside their smart metering solutions [31,32]. Currently these solutions offer limited to advanced functionalities for residential and commercial users, respectively. However, large enterprises can have access to application programming interface (API) and can use them to tailor customized solutions. Load control: Smart meters allow power flow control through demand response or load control. During peak demand hours, electricity demand to non-critical loads can be cut off in order to allow critical loads to continue. This enables the generation capacity to meet the peak demand [33]. The implementations can range from fairly basic mathematical programs to highly complexity algorithms designed to meet peak load and cost reduction goals [34]. Currently the smart meters available in Pakistan don’t act as power controller for individual loads; rather they allow the entire set of devices to which they are connected to either be ON or OFF [32]. At best, they can set a cap on the peak power flow through them (load thresholding). Therefore currently in Pakistan, load thresholding necessitates user agent (human or computer) to implement highly effective demand response schemes. Two way power flow monitoring: Smart meters can monitor bidirectional power flow. Consumers who house an electricity generation source and provide excess generation to utilities need to keep track of their generation and consumption portfolio. Such consumers can use smart meters to achieve this goal. Many smart meters with bi-directional power flow measurement are available in Pakistan, mostly for industrial consumers [31,32].

3. Residential, commercial and network implementations Marvin et al. have identified four smart metering development pathways for residential, commercial and distribution network settings: Monitoring, Gatekeeping, Producer Led and Consumer Led pathways (Fig. 3) [35].

Consumer-Led Monitoring - Automated Data Records

Development Pathways

- Feedback Driven with real time pricing & incentives - User Agents for consumer utility

Producer-Led - Real time system Monitoring - Fault & Theft Detection - Utility Enforced Load Control

Fig. 3. Smart meter implementation models.

 Monitoring Electromechanical meters had traditionally been used to record the total energy consumptions which were then reported manually by the meter readers. Smart meters have automated the meter reading and reporting process. Consumption information can be relayed to the utility, which is then passed onto the customer monthly along with the electricity bills. The only difference from the older metering system might be the day-today energy consumption readings which themselves are relayed to consumer on monthly basis. This provides the users with a coarse feedback on their energy consumption allowing them to make appropriate behavioral changes that impact their consumption and bills positively. In other words, this development pathway is a mere extension of the older metering system and fails to provide any added functionality. Early smart meter installations in Pakistan offered such limited functionalities and can therefore be categorized under this model. However, more recent implementations have begun to move away from this pathway because of its functional limitations.  Gatekeeping The smart meters can act as prepaid meters that provide electricity services against a token or card. They are normally configured to warn the user before the card or token has reached its pre-set limit. They even allow credit to be loaned after the limit has been reached to ensure the continuity of supply to the consumer. The prepaid metering available to the consumers tends to make their usage more efficient [36]. However the customers still end up paying more with prepayment meters because of higher meter and servicing costs [37]. Smart meters also aggregate user data to address privacy concerns arising from leakage of consumption information to sinister elements. The data can be again disaggregated at the utility or consumer end depending upon system architecture. Presently no smart meter implementation in Pakistan can be considered to fall under this model. The government, however, plans to introduce prepaid smart meters to streamline power consumption and discourage defaults [38].  Producer-led This model empowers the utility/ISO (or any other licensed third party) to configure the technical aspects of the grid system. The utility/ISO invokes smart meter communication infrastructure to carry out real-time monitoring and demand management. Monitored data serves to keep track of present electricity consumptions and the system health. One of the uses of this data is the accurate assessment of voltage and current harmonics which is essential for realization of distributed generation within the grid [39,40]. Apart from the

W. Aslam et al. / Renewable and Sustainable Energy Reviews 44 (2015) 933–943

monitored data, intelligent meters have also helped alleviate micro generator registration issues by using pattern recognition techniques. Different generation sources (e.g. wind) now have plug and play functionality, and can be added into the grid by micro generator suppliers with the meter taking care of generation ‘source’ and ‘type’ registration [41].

On the distribution network, smart meters along with SCADA systems carry out fault detection and poll out the regions where the fault has been sensed to ensure a self-healing network [42]. Additionally, geographical mapping of the electrical network can be obtained using smart meters as demonstrated in [43]. The mapping is quite useful because geographic information systems (GIS) with real-time data offer better outage and theft monitoring. In this regard, smart meters have been known to reduce the outage duration by 4–6 min [18]. Real time data from various smart meters in the distribution hierarchy should always reconcile within reasonable limits; usually a major irregularity indicates an outage or theft. Fig. 4 shows one instance of electricity theft at a local grid in Pakistan. The data shown was collected using a combination of smart meters installed at a distribution transformer and downstream consumers. It can clearly be seen that the slave meters’ cumulative consumption (red) is less than the overall electricity supplied to the transformer (blue) even when the line losses are taken into account. This approach can be used to identify high loss areas for deeper investigation. Depuru et al. have discussed a method that not only detects electricity theft but also rebuffs the appliances of people responsible for theft [14]. Long term load balancing of different phases at feeders can also be ensured via smart meters. Cumulative loads on different phases are monitored over time to check if the network is balanced. In case the network is unbalanced, a remapping and reconfiguration of the network phases can be done using GIS and smart meters to come up with the most stable and balanced network. However the physical changes on feeder and transformers still have to be made manually. Generally in order to lessen the gap between demand and supply in short term such as in the event of loss of generation or connection of a large load, smart meters allow control of different loads on the grid. The utility/ISO controls appliances running at the user end through smart meters to manage demand. It is also ensured by the producer that the system stability is not

937

sacrificed. One such scheme of system-wide frequency drop prevention is discussed in [44]. Currently in Pakistan, greatest smart metering roll out is being carried out under this model. USAID power distribution program (PDP) began in 2010 and aims to improve power distribution operations [45]. The program works with nine major public owned distribution companies (DISCOs) and has been successful until now. Private entities have also undertaken their own initiatives towards this end. K-Electric, which is the largest private utility in Pakistan, recently signed an agreement with Info Tech Pakistan Limited (IPL) to implement its Smart Grid technology platform [46]. The focus on producer-driven pathway can be explained by great potential savings that might result in response to relatively small infrastructural changes. 3.1. Consumer-led This model has similar technical components to the producer led pathway but the information flows and configuration capabilities lie at the consumer end. Whereas the producer led model cared little about consumer preferences in fulfilling the goal of demand management, this model places user high on the priority list whilst achieving the objective. Under this model, the user is constantly being updated about his consumption through in-home displays (IHDs), web portals and smart phone applications. OPower and the now-defunct Google Power Meter and Microsoft Hohm are some examples of the user centric energy management and viewing applications. The real-time feedback from such applications tends to make the user more energy aware and proactive about his consumption. It has been estimated that active use of IHDs can lead to about 7% energy savings [47]. Multiple energy competitions and games are also offered to the user as part of this model to induce long term energy saving behavior. Ref. [48] evaluates an energy saving competition known as Kukui Cup at University of Hawaii and lays down the lessons for user stickiness to energy conservation programs. Along with the provision of energy consumption information, the user led model normally incorporates carbon emissions and savings numbers as well. This is different from the previous models which implicitly dealt with carbon footprint. In addition to this, the given model incorporates user comfort and preferences more explicitly, when carrying out demand side management. The energy management system can be configured

Energy Consumption (kWh)

450000 400000 350000 300000 250000 200000 150000 100000 50000 0 Jun-08 Cust - 01 Cust - 05 Cust - 09 Cust - 13 Cust - 17 Cust - 21 Cust - 25 Cust - 29 Difference

Jul-08

Aug-08

Sep-08 Cust - 02 Cust - 06 Cust - 10 Cust - 14 Cust - 18 Cust - 22 Cust - 26 Cust - 30

Oct-08

Nov-08

Dec-08

Jan-09

Feb-09

Cust - 03 Cust - 07 Cust - 11 Cust - 15 Cust - 19 Cust - 23 Cust - 27 Cumulative consumption

Mar-09

Apr-09

May-09

Cust - 04 Cust - 08 Cust - 12 Cust - 16 Cust - 20 Cust - 24 Cust - 28 Total Generated

Fig. 4. Distribution transformer (blue) and aggregated downstream meter data (red) in kW h. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

938

W. Aslam et al. / Renewable and Sustainable Energy Reviews 44 (2015) 933–943

by the user to maximize utility and reduce energy bills. This optimization is carried out in response to utility specified pricing or load control signals. Different user centric implementations of DSM have been given in [49–51]. Primary focus and patronage within Pakistan is on the producer-led pathway. On the other hand, there are only a few instances of the consumer-led model. The AMR energy meters installation by Gujranwala Electric Power Company (GEPCO) is an example of the consumer-led framework, though a fairly basic one [52]. The dearth of support for consumer-led model in Pakistan is often explained by lack of technology penetration and awareness amongst people. However, the growing popularity of smart phones in Pakistan coupled with the country’s broad base of cell phone users opens new possibilities for user-centric energy applications [53]. And in coming years, this model might be an important component for the smart grid in Pakistan.

4. Proposed industrial implementations This section proposes various ways in which smart meter based solutions can be implemented in an industry in Pakistan and in general around the world. The proposed solutions are meant for energy intensive industries (such as Textile, Leather, and Glass) but can also be implemented in small and medium enterprises (SMEs) (Fig. 5). 4.1. Energy auditing Energy auditing is the process of assessing the energy efficiency of a system and identifying ways in which it can be improved [54]. Industrial energy audits play an important role in achieving cost optimization and peak load reduction goals for industry and utility, respectively. These audits can become much more effective if greater details about consumption breakdown and user behavior are known. Smart meters offer varying levels of granular data that can precisely reveal the energy consumption trends, hence allowing effective audits. In order to simulate an industrial energy audit, smart meters were installed in the faculty apartments at Lahore University of Management Sciences (LUMS). Data from the meters was collected

using a custom software built over the vendor provided Application Programming Interface (API). The data was then analyzed using various techniques. Fig. 6 illustrates and compares the total daily electricity consumption for different apartments. Here each colored bar represents a separate apartment. Net consumption breakdown on the basis of weekdays is given in Fig. 7. Through such analyses, one can get a feel of where and when most of the energy is being utilized. Smart meters can be installed in industries where they can identify the most energy intensive machines and processes, and compare them against industry benchmarks to develop techniques for energy conservation. In addition to LUMS smart meters, electricity consumption data from a local software house was acquired and analyzed (Fig. 8). The data showed interesting insight into the behavior of people working at the place. It can be appreciated that the electricity consumption is minimal on weekends since these are holidays. However, a gradual increase in electricity consumption from Mondays to Fridays can be noticed. This increase can be interpreted as the pressure of deadlines on the employees just as the week progresses. On Monday, the employees have generally caught up with previous week deadlines and are more relaxed. But as the end of the week approaches and the pressure to meet the deadlines increases, the employees put in long hours and energy consumption increases proportionately. Such kind of behavioral analysis is very useful for industries, since it can give the industries a peephole to look into the employee psyche and take measures to increase the employee productivity as well as satisfaction. Pakistan’s electricity grid has a unique set of problems at consumer and utility ends. The electrical network in the country suffers widespread meter tampering and defaults, particularly in the industrial sector [21]. Utility operations, on the other hand, are marred by extensive bribery and lack of transparency. Smart meters offer tangible solutions to such problems. Data from smart meters is sent periodically to a central server and can’t be easily manipulated by rogue elements within the utility. As for consumer tampering, smart meters generate alerts as soon as they detect any anomaly. The meters can also help disconnect these users using load control feature. Further, prepaid energy meters would solve the problem of defaulted payments in the country as power would be supplied against advance payments [38]. Yet another contribution of smart meters in Pakistan’s context is the provision of detailed analytics [32]. Previously the luxury of such analysis could only be afforded by large scale industries, because of the associated high costs. Even within those industries the use of analytical methods for energy efficiency was quite limited. But the emergence of locally manufactured smart meters and improvements in information and communication technologies (ICT) mean that smart metering solutions are much more accessible. Therefore small and medium scale industries are now potentially able to use advanced analytical practices in their planning and operations. Smart metering solutions within industries tend to be highly modular and scalable [55]. Their expansion is simple and economical as only new smart meters need to be added to the existing infrastructure. It is occasionally that other infrastructural changes have to be made. Many industries in Pakistan are still relatively small, and for these industries adopting smart metering solutions would be quite prudent. Not only should these industries be able to afford the solutions, but should be able to leverage the benefits as they continue to expand. 4.2. Load diagrams

Fig. 5. Problems for industries in Pakistan and benefits of proposed smart metering solutions.

A load profile is a graph of power consumption versus real-time whereas a load duration diagram represents cumulative power

W. Aslam et al. / Renewable and Sustainable Energy Reviews 44 (2015) 933–943

939

Fig. 6. Daily electricity consumption—LUMS faculty apartments.

1100

Energy Consumed (KW h)

6

Week Number

5 4 3 2 1 0

10

20

30

40

50

Energy Mon

Tue

60

70

80

90

100

1000 900 800 700 600 500 400 1

3

5

7

Percentage Wed

Thu

Fri

Sat

9

11

13

15

17

19

21

23

25

27

29

Day/ Feburary 2011 Sun

Fig. 7. Consumption breakdown per weekday for six weeks—LUMS faculty apartments. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

consumption against operational duration [56]. Depending upon how the smart meters are set up to collect data, different kinds of load diagrams can be made. Within the industries, these diagrams can correspond to individual machines or entire processes and each of these diagrams can highlight a number of important factors as shown in Table 2 [57]. The industrial load management programs in Pakistan are based on manual judgment and experience. In many instances, these programs fail to understand and incorporate the present energy profile of the country which has further worsened the energy shortfall. On the other hand, load diagrams made from

Fig. 8. Software house consumption data.

Table 2 Classification of load diagrams and their benefits. Machine specific diagrams

Process specific diagrams

Identification of load start-up/ shutdown increments and their effects on the overall demand Load cycle analysis and comparison with industrial benchmarks

Reveal magnitude, duration and time for peak demand periods

Easy and accurate assessment of problem areas such as short cycling compressors

Help understand system interactions e.g. increased electric heating demand when ventilation dampers are opened Establish production, occupancy and weather effects on electrical demand

W. Aslam et al. / Renewable and Sustainable Energy Reviews 44 (2015) 933–943

Load (kW)

940

350

350

300

300

250

250

200

200

150

150

100

100

50

50

0

0

Variable Load

Fixed

22:45

21:00

19:15

17:30

15:45

14:00

12:15

10:30

8:45

7:00

5:15

1:45

3:30

0:00

23:00 22:00 21:00 20:00 19:00 18:00 17:00 16:00 15:00 14:00 13:00 12:00 11:00 10:00 9:00 8:00 7:00 6:00 5:00 4:00 3:00 2:00 1:00 0:00

Load Duration (hours)

Time (hours) Fig. 9. Load profile and load duration diagrams [56].

smart meter data present a low hanging fruit for the design of effective demand management programs and load limits [58]. With lot greater and accurate information about the industrial demand, the utilities would be able to develop better understanding of demand patterns and be precisely able to assess the effects of their policies. Within the industries in Pakistan, there is little knowledge about energy consumption of different machines and processes over short periods of time. The instantaneous power draw is often very important since it can indicate the problem areas within a process or machine. By using smart meters these instantaneous power draws can be easily captured in the form of load profiles [59]. This measure would not only increase the efficiency of the system but would also significantly reduce the operational and maintenance costs (Fig. 9). 4.3. Demand response Demand response (DR) is broadly understood as the modifications in electrical energy consumption pattern of a consumer in view of some incentive, pricing effect or even a signal from utility [34]. Energy efficiency and demand response are different in nature and objectives as explained by Goli et al. Energy efficiency programs aim to decrease the consumer energy usage on a permanent basis by the installation of energy efficient technologies whereas demand response is a series of steps that reduce electric loads in case of grid failure, peak demand or any other emergency [17]. Demand response in industries is well established in US and parts of Europe [60]. Smart Meter’s role in Demand Response is illustrated by a typical three-step DR event [16]:  Overload: The system’s peak load goes beyond a certain predefined threshold.  Notification: The smart meter triggers a notification at the server indicating the overload.  Client response: Depending upon the industry, the server can take various actions – turn off the non-critical load, execute a back-up generator and so on. The response to the notification from the smart meter can either be manual or automated. In manual client action, there is a person at the facility end managing the loads (switching or clipping of non-critical loads). The drawback of this simple method is the element of human-in-the-loop which may result in a slower and sometimes inaccurate response. The second method i.e.

automated demand response initiates pre-programmed steps of operations in response to a notification. Samad et al. look at how an automated demand response is much faster and reliable. They have also discussed various architectural implementations of Open ADR protocol [16]. As seen in the previous sub-section, smart meter analytics play an important role in the design of DR programs. In addition smart meters can act as the platform on which these DR programs are implemented [61]. Complex smart grid implementations might require dedicated home or distribution level controllers but smart meters can serve as good substitutes. This can be particularly important for Pakistan since it might be a long way before smart grid can be truly realized in the country. This lack of development might be more of a problem in the residential sector where more wholesale infrastructural changes are required but this should not deter industries from leveraging the benefits of smart grid paradigm. For now, industries can use smart meters as the infrastructure for implementing demand response programs. This would result in substantial peak load reduction and would eliminate the need for costly peaking generators within industries. Further the utilization of captive generation would be enhanced by smart meters since they offer load control through thresholding and demand response opportunity for even the residential and commercial areas within industries. Using smart meters in conjunction with variable speed drives (VSDs) and variable frequency drives (VFDs) offer numerous exciting demand response opportunities. Smart meter can act as both a sensor and controller for the drives, sending out signals to the automation system for changing drive frequency or speed [62]. Lee et al. discuss how intelligently using smart meters with VFDs offers an effective way of increasing energy efficiency [63]. Another study done by Hasanbeigi et al. at Lawrence Berkeley National Laboratory has identified a number of potential areas where using VFDs for non-critical load, also sometimes called elastic load, can lead to energy saving [64,65]. Some of the possible machines suitable for VFDs include washer pump motors in humidification plant, supply fans and return fans in ventilation systems. The recent focus on renewable generation in Pakistan has pressed the need for having variable frequency drives in industries. By using smart meters, these adjustable speed drives can be easily incorporated within the industries. 4.4. Performance characteristic line Electrical energy cost is usually an important part of the overall cost of an industrial process and the product itself. For energy

intensive industries such as cement or glass, the need for measuring the exact electrical cost becomes more pronounced. Unfortunately in Pakistan, an overwhelming majority of industrial owners don’t employ precise measurement methods for calculation of productivity and electric costs in industrial processes. They instead resort to manual estimation of these parameters which are often inaccurate. Smart meters present an easy way for carrying out energy auditing and constructing load diagrams which form the basis of this performance analysis and cost allocation [64,65]. These analysis tools enable a systematic approach for handling business processes and can lead to significant cost and energy savings. The exact methodology of analysis can be categorized under the performance characteristic line (Fig. 10). It is the mathematical equation that depicts trend of energy usage against product output [57,67].The performance characteristic line can reveal a number of important factors: (1) Specific energy consumption (SEC): It is energy usage per unit product output and is represented by the gradient of the performance characteristic line. SEC is useful because it provides an exact measure of the share of electrical cost for product output. Further, plotting specific energy consumption (SEC) against time (month or a year) can indicate the trend in efficiency of a particular machine/process [57,66]. (2) Base load: Base load represents the fixed energy requirements of an industry and usually plays an important role in determining the overall industrial cost. On the diagram, it is the point of intersection of vertical axis (energy usage) and the line. (3) Target growth prediction: By extrapolating the line, the effects of meeting a certain target output can be seen. Even though this regression analysis might require deep understanding of industrial operations still it is quite helpful for plant managers in devising new policies. In addition, comparing the actual and predicted electrical usage at any point would help gauge the effectiveness of previously implemented measures and/or might point out the shortcomings in prediction models (Fig. 11). 4.5. Fault analysis The electrical grid in Pakistan is quite unstable and faults are a regular occurrence. There is no infrastructure for the identification of faults and finding the exact fault location usually takes long hours. This results in loss of power to industries which disrupts their normal operations thus forcing them to use costly backup generators. The data and notifications from smart meters can accurately point the fault location and have the potential to greatly reduce the outage times. So just by using smart meters, far more

Specific Energy Consumption (MWh/kg)

W. Aslam et al. / Renewable and Sustainable Energy Reviews 44 (2015) 933–943

0 200

250

300

Base Load

Performance

Line

Fig. 10. Performance characteristic line [66].

350

400

EnergyUsage(MWh)

3/1/2008

5/1/2008

7/1/2008

9/1/2008

11/1/2008


Determination of system goals: The objectives of implementa-

100

Weight (1000 kg)

0.5

The use of smart meters in industries would require three different classes of transformation i.e. change in infrastructure, addition of monitoring and analysis layer and modifications in business processes. A blind implementation of these changes can be extremely inefficient and costly. Therefore it is essential to first understand different layers of smart metering infrastructure (applications layer, communications layer and power layer) and identify the business, financial and technical requirements that are to be satisfied by this infrastructure. Though the final use cases of the infrastructure are of prime significance, it is also important to consider the transitioning goals and constraints. Kazman et al. have discussed an approach for detailed architecture evaluation which is extremely relevant given the absence of a well-defined architecture for the use of smart meters within industries [68]. A possible sequence of steps for the deployment of meters is listed below:

200

150

1

5. Implementation methodology

300

100

1.5

regular supply of power to industries and residential consumers can be ensured. The voltage/current fluctuations on the country’s grid are sometimes responsible for damages to industrial assets and equipment that might cost millions. The use of smart meters on power lines within industry would enable quality monitoring of the delivered power. Early detection of faults or variability in power quality by intelligent decision support systems can help safeguard industrial assets. Lee et all have mentioned the possibility of using smart meters to monitor sag/swells, transient or over current waveforms, duration of voltage dip and harmonic patterns of the grid to detect problems in the power line [63].




50

2

Fig. 11. Specific energy consumption—actual (blue) vs predicted (yellow) [66]. (For interpretation of the references to color in this figure legend, the reader is referred to the web version of this article.)

600

0

2.5

1/1/2008

700

400

3

0

800

500

941




tion should be decided. A set of possible goals might include asset protection and cost optimization, self-healing capability for the system and ease of electrical consumption and generation analysis. The goals should be mathematically represented as a function of interest variables. Identification of system architecture and use cases: Organization of the system and protocols of interaction between different system components, which have the capability to address different objectives, should be defined. The most common working configurations of smart meter implementation should also be known. Developing the system constraints and identification of risks: Different constraints that the system would be subjected to should be found out. The constraint matrix should include not only final operational and business constraints but also various

942




W. Aslam et al. / Renewable and Sustainable Energy Reviews 44 (2015) 933–943

transitioning elements that would eventually affect system goals and costs. Different risks should also be addressed and be incorporated in the overall system model. Assessment of solution: The system model that has been formed should then be objectively assessed to determine optimality of the solution.

in industries and provides plug and play functionality with limited concurrent meters. With an objective and systematic implementation, the approaches presented can use smart meters to mitigate grid inefficiencies in Pakistan. In future, we plan to establish a testbed in collaboration with major industries to fully assess the benefits and problems associated with smart meter implementation and create working prototypes for specific industries.

6. Future directions

Acknowledgements

This paper lays out the general framework for smart meter implementation within industries in Pakistan. However there are a few areas that the paper doesn’t directly touch upon:

The authors would like to thank Mr. Shayan Saeed for his contribution in test-bed simulations. The authors also acknowledge financial and technical support provided by Cleaner Production Institute (CPI) and Microtech Industries (MTI )throughout the project.


The degree of acceptance of smart meters from industries has







not been considered. Until now, the lack of awareness about energy conservation amongst industries has been a huge hurdle in achieving the energy efficiency goals in Pakistan. In addition to this, spending significant upfront costs might deter industries from deploying smart meters. Government should therefore take lead in introducing energy awareness programs and incentivizing smart metering usage within industries with programs such as [69]. Different universities can also play an important role towards this end by offering courses and conducting detailed studies in related subject areas. IT sector can be taken on board as well by encouraging them to tailor low cost smart metering solutions for industries. The extent of change in business processes in industries as a result of smart metering usage has not been taken into account. ‘Business process’ is a very loosely defined term within Pakistan’s context and may have a very subjective interpretation across different industries. A significant part of smart metering success depends upon making the business processes more systematic. Therefore it is important that the industries are not only ready but are also able to make the right changes in their business processes. This can be ensured by increasing awareness about business processes management and re-engineering amongst industry personnel. Measures such as the ones mentioned in [20,45] can serve as a good starting point. Security and privacy concerns regarding data have not been addressed. In that regard, standard operational and network security procedures e.g. data aggregation can be used to ensure a robust and failsafe network [70,71].

It is important to note that smart meters, though very crucial for industrial energy efficiency, are only a part of the true smart grid paradigm. Therefore it is imperative that the smart meters are implemented considering the future smart grid model.

7. Conclusion Pakistan currently faces a dire energy crisis. Present energy situation in the country necessitates use of technology for cost reduction and efficiency improvement. This paper discusses methodology for enabling such improvements through the use of smart meters in industrial sector. It builds upon the residential and commercial models of smart meter implementation and proposes ways smart meters could be used within industries. It identifies fine-grained analyses/energy audits, asset protection and peak demand reduction as some of the resulting benefits. The work presented in this paper demonstrates various proof of concepts through the use of data from university grid smart meters which mimic an industrial setup. It further lays out blueprint for flexible, scalable and secure energy audit infrastructure that can be utilized

Appendix Daily energy shortfall ¼ 5000 MW Average daily shortfall duration¼8 h Table A1: [72] Average cost per unit for current generation mix (calculated from above data) ¼Rs 8.835¼ $ 0.083 Total cost ¼$ 13.5 billion 13:5109 Time period ¼ ¼ 11:14 years 3 500010 80:083365

Table A1 Generation mix and energy production costs in Pakistan. Source

% Generation share

Price per unit (Rs/kW h)

Hydel High speed diesel Residual fuel oil Gas Nuclear Import (Iran) Mixed Wind

15.87 1.92 42.45 31.29 6.98 0.33 1.15 0.01

0.2246 19.26 16 4.3691 1.2145 9.02 13.5 9.1213

References [1] Munir S. Tackling the energy crisis. The Nation; 2013. [2] Annual report. National Electric Power Regulatory Authority (NEPRA); 2011–2012. [3] Ahmad, I., Renewables and electric power deficit. IEEEP; 2012. [4] World Bank Indicators. Available from: 〈http://www.tradingeconomics.com/ pakistan/electric-power-transmission-and-distribution-losses-kwh-wb-data. html〉 [2010 16.7.2013]. [5] Lights Out. The Economist; 2011. [6] Shah S. Priority in Pakistan: turn on lights. Wall Street J 2013. [7] Sheikh, M A. Energy and renewable energy scenario of Pakistan. Renewable Sustainable Energy Rev 2009. [8] Sahir, M H, Qureshi AH. Assessment of new and renewable energy resources potential and identification of barriers to their significant utilization in Pakistan. Renewable Sustainable Energy Rev 2006. [9] Chaudhry MA, Raza R, Hayat SA. Renewable energy technologies in Pakistan: prospects and challenges. Renewable Sustainable Energy Rev 2008. [10] Amer M, Daim TU. Selection of renewable energy technologies for a developing county: a case of Pakistan. Energy for Sustainable Development 2011. [11] A.Khan H, Pervaiz S. Technological review on solar PV in Pakistan: scope, practices and recommendations for optimized system design. Renewable Sustainable Energy Rev 2013. [12] Mirza K, U., et al. Wind energy development in Pakistan. Renewable Sustainable Energy Rev 2006. [13] Deb, S., Bhowmik, P.K., Paul, A. Remote detection of illegal electricity usage employing smart energy meter—a current based technique. In: IEEE PES innovative smart grid technologies, Kolam, Kerala; 2011.

W. Aslam et al. / Renewable and Sustainable Energy Reviews 44 (2015) 933–943

[14] Depuru SSSR, Wang L, Devabhaktuni Vijay. Electricity theft: overview, issues, prevention and a smart meter based approach to control theft. Energy Policy 2010. [15] Costanzo GT, et al. A system architecture for autonomous demand side load management in smart buildings. IEEE Trans Smart Grid 2012. [16] Samad T, Kiliccote S. Smart grid technologies and applications for the industrial sector. Comput Chem Eng 2012. [17] Goli, S., McKane, A., Olsen, D., Demand response opportunities in industrial refrigerated warehouses in California, E-Scholarship; 2011. [18] Depuru SSSR, Wang L, Devabhaktuni V. Smart meters for power grid: challenges, issues, advantages and status. Renewable Sustainable Energy Rev 2011. [19] Electric meters to be replaced, in Express Tribune; 2011. [20] USAID to roll-out largest automated meter reading project in Pakistan; 2013. Available from: 〈http://www.pakistantoday.com.pk/2013/01/18/business/usaidto-roll-out-largest-automated-meter-reading-project-in-pakistan/〉 [cited 2014]. [21] Haq, S. In anti-theft drive, LESCO plans to install smart meters; 2013. Available from: http://tribune.com.pk/story/613246/in-anti-theft-drive-lesco-plans-toinstall-smart-meters/ [cited 2014]. [22] Zubair Md. Fadlullah, M.M.F., Nei, Kato, Akira, Takeuchi, Noboru, Iwasaki, Yousuke, Nozaki. Towards intelligent machine-to-machine communications in smart grid. In: IEEE communications magazine; 2011. [23] Usman A, Shami SH. Evolution of communication technologies for smart grid applications. Renewable Sustainable Energy Rev 2012. [24] Wenye Wang YX, Khanna Mohit. A survey on the communication architectures in smart grid. Comput Networks 2011:3604–29. [25] Kashif Laeeq, W.L. A comparative study among possible wireless technologies for smart grid communication networks. In: First international conference on modern communication & computing technologies; 2014. [26] Pakistan has the lowest SMS cost in the world; 2011. Available from: 〈http:// www.telecomrecorder.com/uncategorized/pakistan-has-the-lowest-sms-cos t-in-the-world-research-report/〉 [cited 2014]. [27] Mar´ıa Arenas-Mart´ınez, S.H.-L., Sanchez, Abel, Williams, John R., Roth, Paul, Hofmann, Paul, Zeier, Alexander. A Comparative study of data storage and processing architectures for the smart grid. In: First IEEE international conference on smart grid communications. Gaithersburg, MD; 2010. [28] Islamabad Chamber of Commerce and Industry lauds Prime Minister’s move against power theft; 2013. Available from: 〈http://asianetpakistan.com/officialnews/chambers-of-commerce/148364/islamabad-chamber-of-commerce-and-in dustry-lauds-prime-ministers-move-against-power-theft/〉 [cited 2014]. [29] Achieving high performance with smart meter data management systems. Accenture; 2011. [30] Popa, M., Data collecting from smart meters in an advanced metering infrastructure. In: 15th IEEE international conference on intelligent energy systems. Poprad; 2011. [31] Power & Energy Monitoring System; 2014. Available from: 〈http://www. schneider-electric.com/products/ww/en/4100-power-energy-monitoring-sys tem/?APPLICATION=1009〉 [cited 2014]. [32] Smart Meters; 2014. Available from: 〈http://www.mtilimited.com/meters/ smart-meters.php〉 [cited 2014]. [33] Lipski, M. Demand response—technology for the smart grid; 2011. [34] Balijepalli, V.S.K.M., et al. Review of demand response under smart grid paradigm. In: Innovative Smart Grid Technologies (ISGT). Kollam, Kerala; 2011. [35] Marvin S, Chappells H, Guy S. Pathways of smart metering development: shaping environmental innovation. Comput Environ Urban Syst 1999. [36] Smart meters as prepaid meters. Intelligent utility; 2013. Available from: 〈http:// www.intelligentutility.com/article/13/07/smart-meters-prepaid-meters〉. [37] Owen, G., Ward, J., Smart meters in Great Britain: the next steps? Sustainability first; 2007. [38] Wapda to introduce pre-paid meters to streamline recovery. Dawn; 2013. [39] Capua, C.d.., Romeo, E., A Smart THD meter performing an original uncertainty evaluation procedure. IEEE Trans Instrum Meas, 2007. [40] Lee, P.K., Lai, L.L. A practical approach of smart metering in remote monitoring of renewable energy applications. In: Power and energy society general meeting. Calgary, AB; 2009. [41] Boait, P. Smart metering of renewable micro generators by output pattern recognition. In: 20th International conference on electricity distribution. Prague; 2009. [42] Mak, S.T., Farah, N. Synchronizing SCADA and smart meters operation for advanced smart distribution grid applications. In: Innovative Smart Grid Technologies (ISGT). Washington, DC; 2012.

943

[43] Short TA. Advanced metering for phase identification, transformer identification, and secondary modeling. IEEE Trans Smart Grid 2013. [44] Samarakoon, K., Ekanayake, J. Demand side primary frequency response support through smart meter control. In: Proceedings of the 44th International Universities Power Engineering Conference (UPEC). Glasgow; 2009. [45] USAID Power Distribution Program; 2014. Available from: 〈http://www.pdip. pk/about-pdip/〉. [46] KE and InfoTech Pakistan sign agreement for implementation of smart grid technology platform; 2014. [47] Faruqui A, Sergici S, Sharif A. The impact of informational feedback on energy consumption—a survey of the experimental evidence. Energy 2009. [48] Brewer, R.S., et al. Energy feedback for smart grid consumers: lessons learned from the Kukui Cup, in ENERGY 2013. In: The third international conference on smart grids, green communications and IT energy-aware technologies. Lisbon, Portugal; 2013. [49] Molderink, A., et al. Domestic energy management methodology for optimizing efficiency in Smart Grids. In: IEEE Bucharest power tech conference. Bucharest, Romania; 2009. [50] Mohsenian-Rad, A.-H., et al. Optimal and autonomous incentive-based energy consumption scheduling algorithm for smart grid. In: Innovative Smart Grid Technologies (ISGT). Gaithersburg, MD; 2010. [51] Barker, S., et al. SmartCap: flattening peak electricity demand in smart homes. In: IEEE international conference on Pervasive Computing and Communications (PerCom). Lugano; 2012. [52] MicroTech successfully installs its AMR Energy Meters at GEPCO. Available from: 〈http://www.mtilimited.com/newsroom/gepco-meters.php〉. [53] Pakistan telecommunication authority annual report; 2013. [54] Canbay CS, Hepbasli A, Gokcen G. Evaluating performance indices of a shopping centre and implementing HVAC control principles to minimize energy usage. Energy Build 2004. [55] Gungor Vehbi C, G.P.H. Industrial wireless sensor networks: challenges, design principles, and technical approaches. IEEE Trans Ind Electron 2009;56:10. [56] Morvay ZK, Gvozdenac DD. Applied industrial energy and environmental management. UK: John Wiley & Sons Ltd.; 2008. [57] Energy savings toolbox—an energy audit manual and tool, in Canadian Industry Program for Energy Conservation (CIPEC); 2009. [58] Carrie Armel K, Gireesh Shrimali AG, Albert Adrian. Is disaggregation the holy grail of energy efficiency? The case of electricity Energy Policy 2013. [59] Hsueh-Hsien Chang, H.-T.Y., Lin, Ching-Lung. Load identification in neural networks for a non-intrusive monitoring of industrial electrical loads. In: Computer supported cooperative work in design IV. Melbourne, Australia; 2008. [60] Torriti J, Hassan MG, Leach M. Demand response experience in Europe: policies, programmes and implementation. Energy 2009. [61] Rob van Gerwen, S.J., Rob Wilhite smart metering; 2006. [62] 4 FM quick reads on smart grid. Available from: 〈http://www.facilitiesnet.com/ powercommunication/tip/The-Role-of-Smart-Meters-in-the-Smart-Grid–22682#〉. [63] Lee, P.K., Lai, L.L. A practical approach of smart metering integration in micro-grid. In: IEEE power and energy society general meeting. Minneapolis, MN; 2010. [64] Hasanbeigi A, Price L. A review of energy use and energy efficiency technologies for the textile industry. Renewable Sustainable Energy Rev 2012. [65] Hasanbeigi A, Price L. Industrial energy audit guidebook: guidelines for conducting an energy audit in industrial facilities. Lawrence Berkeley National Laboratory 2010. [66] Cruycke, B. Energy monitoring: a must for textile companies;2013. Available from: 〈http://www.visionbms.com/textiles/en/testimonials/EnergyMaster% 20Textiles_E.pdf〉 [cited 2013]. [67] Kent R. What’s your process energy fingerprint? Plast Technol 2009. [68] Kazman, R., et al. Architecture evaluation without an architecture: experience with the smart grid. In: Proceedings of the 33rd international conference on software engineering; 2011. [69] Energy Efficiency Initiatives; 2014. Available from: 〈http://www.smeda.org/index. php?option=com_content&view=article&id=18&Itemid=120〉 [cited 2014]. [70] Ross Anderson, S.F. Who controls the off switch? In: First IEEE international conference on smart grid communications. Gaithersburg, MD; 2010. [71] Massoud Amin, S., A.M.G., Smart grid—safe, secure, self-healing. In: IEEE power and energy magazine; 2012. [72] Ghumman, M. Pepco buying 43pc electricity from furnace oil fired plants at Rs16/unit; 2012. Available from: 〈http://www.brecorder.com/top-news/ 1-front-top-news/49898-pepco-buying-43pc-electricity-from-furnace-oil-fir ed-plants-at-rs16unit-.html〉 [cited 2013].